--- Log opened Wed Apr 23 00:00:05 2014 | ||
--- Day changed Wed Apr 23 2014 | ||
fenn | unit tests | 00:00 |
---|---|---|
fenn | er, regression testing | 00:00 |
fenn | that's all i got | 00:00 |
xmj | customers yelling "Why does this not look the same across all my browsers / devices ???" | 00:01 |
xmj | "My macbook air displays differntly than my Galaxy S999!!!" | 00:02 |
@kanzure | often just me yelling at myself about that | 00:02 |
fenn | "because it's not an image?" | 00:02 |
@kanzure | just because it's not an image doesn't mean that i get to ignore rendering engines | 00:02 |
fenn | why not | 00:02 |
@kanzure | weinre is pretty helpful | 00:02 |
fenn | how many different engines are there anyway | 00:04 |
fenn | more than 3? | 00:04 |
@kanzure | gecko, webkit, presto, trident | 00:04 |
@kanzure | blink | 00:04 |
@kanzure | servo | 00:04 |
@kanzure | each have their own versions, bundled into applications that also have lots of versioning, and then those applications have different distributions for each device | 00:05 |
fenn | is it the goal of the engine developers to be standards compliant? | 00:07 |
fenn | i.e. do what the spec says as best they can | 00:07 |
fenn | i can't think of any other field where massive amounts of effort is spent tailoring the input around the bugs of the software | 00:09 |
fenn | maybe government work | 00:09 |
@kanzure | have you ever watched a web developer do his thing? | 00:09 |
fenn | i guess not | 00:10 |
@kanzure | stalkerwatch someone next time you're in a coffee shop | 00:10 |
fenn | what's a web developer? | 00:10 |
@kanzure | you'll either laugh or cry | 00:10 |
@kanzure | it's the person using 10 devices and 100 browsers | 00:10 |
fenn | is this where you take the photoshop drawing the client has given you and chop it into pieces? | 00:10 |
Lemminkainen | and then you animate it with now.js, fenn | 00:10 |
@kanzure | that's sometimes one part of it, but at this point i just outsource that to psd2html.com | 00:10 |
Lemminkainen | can't forget that part | 00:10 |
Lemminkainen | popcorn.js is also very important, it makes your web apps smell like butter | 00:11 |
fenn | so, i want to get rid of all that shit and just use LCARS for everything | 00:11 |
fenn | how do i reverse all the shit into clean structured data? | 00:11 |
Lemminkainen | I have a flock of highly trained pigeons to sell you, then | 00:11 |
@kanzure | well, you can start by not writing for the web | 00:11 |
fenn | i'm thinking some kind of evolutionary classifier, similar to a spam filter | 00:12 |
fenn | "this is the content, this is a nav menu header, this is a nav element, this is a caption" etc | 00:13 |
@kanzure | i already don't like where this is going | 00:13 |
@kanzure | LCARS was crap too | 00:13 |
fenn | i disagree, it was way ahead of its time | 00:14 |
fenn | the 3d graphics were crap, but only because they didn't have the knowhow to make it fit with the rest of the interface | 00:15 |
@kanzure | why is this interesting http://upload.wikimedia.org/wikipedia/commons/a/aa/Lcars_wallpaper.gif | 00:15 |
@kanzure | what the fuck are the numbers on the left | 00:15 |
@kanzure | and why are there curved edges? | 00:15 |
fenn | sorry i am neck deep in swap atm | 00:17 |
-!- delinquentme [~dingo@74.61.157.78] has quit [Ping timeout: 252 seconds] | 00:17 | |
@kanzure | you have like 2 MB of RAM | 00:17 |
@kanzure | after you account for 800 MB of nsa.img | 00:17 |
fenn | hah | 00:17 |
fenn | i always wondered why chrome needed 190MB to show a one paragraph wikipedia entry | 00:18 |
@kanzure | let me know when your baseband processor stops being a secret-sharing monster | 00:18 |
-!- Darius [~quassel@adsl-69-231-35-194.dsl.irvnca.pacbell.net] has quit [Read error: Connection reset by peer] | 00:18 | |
-!- Darius [~quassel@69.231.35.194] has joined ##hplusroadmap | 00:20 | |
fenn | less arbitrary crap http://www.oocities.org/tana100750/ | 00:20 |
fenn | the font is a little bit much, but other than that it's a very elegant design | 00:20 |
@kanzure | look at all those wasted pixels by curves | 00:20 |
fenn | it's a touchscreen interface | 00:20 |
fenn | fat fingers waste pixels | 00:20 |
fenn | so i am thinking you have something like ranger, where each window opens another bar, and the previous bar falls off the side of the screen (but you can go back to it if necessary) | 00:21 |
@kanzure | i think i'll be keeping my keyboard interface for now | 00:22 |
fenn | like a vertical tab bar in a browser | 00:22 |
@kanzure | i carry that keyboard in my pocket | 00:22 |
fenn | in fact it would be a web browser too | 00:22 |
fenn | have you tried the steno app for ipad? | 00:22 |
@kanzure | no, also haven't seen it | 00:23 |
fenn | ah. well, it exists | 00:23 |
fenn | the developer sells a silicone thingy that goes over the screen so you hold your hands in the right place, or something | 00:23 |
fenn | but you can use it without the thingy | 00:23 |
fenn | anyway my point was that onscreen keyboards suck | 00:24 |
@kanzure | i wonder if ALS patients have control of that ear membrane muscle | 00:24 |
fenn | i also like the automatic callouts on diagrams. remember filelight? http://en.wikipedia.org/wiki/Filelight i'm going to re-do that with d3: http://bl.ocks.org/mbostock/4063423 | 00:27 |
fenn | there's a pretty good android disk usage app (called "disk usage" believe it or not) that is similar to how filelight worked, but with rectangles | 00:28 |
fenn | thrash thrash thrash | 00:29 |
-!- sheena [~home@d108-180-136-217.bchsia.telus.net] has quit [Ping timeout: 264 seconds] | 00:29 | |
fenn | https://code.google.com/p/diskusage/ | 00:31 |
-!- ebowden [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has joined ##hplusroadmap | 00:32 | |
-!- nsh [~nsh@wikipedia/nsh] has quit [Ping timeout: 252 seconds] | 00:42 | |
-!- nsh [~nsh@wikipedia/nsh] has joined ##hplusroadmap | 00:47 | |
fenn | oh yes there is also a tricorder app with no rounded corners https://f-droid.org/repository/browse/?fdid=org.hermit.tricorder http://www.appszoom.com/android_applications/tools/tricorder_bsf_screenshots.html | 00:52 |
fenn | probably others like that | 00:52 |
fenn | it seems like most of the stuff on the internet is either a toy (completely not functional for anything but blinkenlights) or the developer hasn't put much thought into a coherent system that spans more than a single app | 00:57 |
fenn | there was a java widget toolkit but i dunno what came out of that | 00:57 |
-!- Darius [~quassel@69.231.35.194] has quit [Read error: Connection reset by peer] | 01:11 | |
-!- Darius [~quassel@adsl-69-231-35-194.dsl.irvnca.pacbell.net] has joined ##hplusroadmap | 01:12 | |
-!- Darius [~quassel@adsl-69-231-35-194.dsl.irvnca.pacbell.net] has quit [Remote host closed the connection] | 01:12 | |
-!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 01:28 | |
fenn | the nook looks so much like the star trek PADD because it reflects the ambient light off a flat surface instead of glowing and bouncing light around inside it http://fennetic.net/irc/nook_padd.jpg http://images.auctionworks.com/hi/3/3282/padd-29.jpg | 01:31 |
fenn | an lcd looks a lot different http://wiki.maemo.org/images/1/15/ThemeShot3.jpg | 01:33 |
ebowden | Oh, is this functioning? | 01:33 |
ebowden | The non-LCD one. | 01:34 |
ebowden | (The non-LCD one.) | 01:34 |
fenn | nook_padd.jpg is how i'm chatting right now | 01:34 |
-!- nsh_ [~nsh@wikipedia/nsh] has joined ##hplusroadmap | 01:35 | |
fenn | padd-29 is a prop made by laser printing onto a transparency sheet | 01:35 |
ebowden | Ah, ok. | 01:36 |
ebowden | I wonder if it would be possible to make one look exactly like the printed one. | 01:36 |
fenn | yeah there are zillions of replicas, that might even be a replica | 01:37 |
fenn | there is a reflective color display technology in perpetual vaporware stage from pixel qi | 01:38 |
-!- nsh [~nsh@wikipedia/nsh] has quit [Ping timeout: 252 seconds] | 01:38 | |
fenn | i've never actually seen one though | 01:39 |
ebowden | Ok. | 01:39 |
-!- augur [~augur@216-164-48-148.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Quit: Leaving...] | 01:39 | |
fenn | oops not pixel qi, i meant Mirasol: http://techcrunch.com/2010/02/15/color-e-reader-uses-butterfly-based-technology-to-save-power/ | 01:41 |
ebowden | Ok. | 01:43 |
fenn | something that doesn't come through so well in pictures is how light the nook is. it weighs less than my phone despite having 4 times the surface area | 01:46 |
fenn | i double checked and the phone actually weighs 40% less. but it FEELS heavier | 01:50 |
-!- ebowden [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has quit [Remote host closed the connection] | 01:50 | |
-!- augur [~augur@216-164-48-148.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined ##hplusroadmap | 01:51 | |
delinquentme | http://www.mensjournal.com/magazine/print-view/do-men-suck-at-friendship-20140422 | 01:56 |
delinquentme | true! | 01:56 |
delinquentme | brownies, kanzure , fenn ParahSailin you guys are my fronds :D | 01:56 |
delinquentme | I hope you die a long time after everyone else <# | 01:56 |
fenn | what happened to that heart? | 01:58 |
-!- ebowden [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has joined ##hplusroadmap | 01:59 | |
delinquentme | I stepped on the top half | 02:00 |
delinquentme | Its how the umbuttu tribesmen relay man-appreciation over instant messaging | 02:00 |
archels | you want your friends to die? gee whiz | 02:00 |
Lemminkainen | mutation before death, dammit | 02:02 |
-!- nsh_ [~nsh@wikipedia/nsh] has quit [Ping timeout: 252 seconds] | 02:03 | |
delinquentme | ^^^^ | 02:16 |
-!- nsh_ [~nsh@wikipedia/nsh] has joined ##hplusroadmap | 02:17 | |
-!- ebowden [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has quit [Read error: Connection reset by peer] | 02:19 | |
-!- ebowden [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has joined ##hplusroadmap | 02:20 | |
-!- nsh_ [~nsh@wikipedia/nsh] has quit [Ping timeout: 240 seconds] | 02:25 | |
archels | don't suppose anyone has a digital copy of Longevitize lying around? | 02:25 |
-!- EnLilaSko [EnLilaSko@unaffiliated/enlilasko] has joined ##hplusroadmap | 02:27 | |
-!- AshleyWaffle [~waffle@gateway/tor-sasl/anastasiawyatt] has quit [Ping timeout: 272 seconds] | 02:37 | |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 02:38 | |
-!- AshleyWaffle [~waffle@gateway/tor-sasl/anastasiawyatt] has joined ##hplusroadmap | 02:47 | |
-!- mosasaur [~mosasaur@95.98.89.217] has joined ##hplusroadmap | 02:49 | |
-!- kuldeepdhaka [~kuldeepdh@unaffiliated/kuldeepdhaka] has joined ##hplusroadmap | 03:07 | |
-!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has quit [Quit: Lost terminal] | 03:09 | |
mosasaur | Is MIRI that bad? I know I share the feelings, but probably for completely different reasons. I mean I am against academia *and* billionaires. | 03:17 |
-!- Adifex is now known as Adifex|zzz | 03:19 | |
archels | paperbot: http://www.sciencedirect.com/science/article/pii/S0262407912627559 | 03:19 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/9542a0ea6c79693d3c44a08fd6ebbf97.txt | 03:19 |
archels | paperbot: http://www.sciencedirect.com/science/article/pii/S0262407912627559/pdfft?md5=bad43cabdd6be43e92c389edbb61ae5f&pid=1-s2.0-S0262407912627559-main.pdf | 03:20 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/a3166d720c2d175befa4fc95ae51fd82.pdf | 03:20 |
archels | <3 | 03:20 |
* archels wonders what that MD5 is for | 03:20 | |
-!- mosasaur1 [~mosasaur@95.99.48.20] has joined ##hplusroadmap | 03:22 | |
-!- mosasaur [~mosasaur@95.98.89.217] has quit [Ping timeout: 252 seconds] | 03:22 | |
-!- mosasaur1 is now known as mosasaur | 03:23 | |
-!- kuldeepdhaka [~kuldeepdh@unaffiliated/kuldeepdhaka] has quit [Ping timeout: 265 seconds] | 03:35 | |
-!- ThomasEgi [~thomas@185.5.8.81] has joined ##hplusroadmap | 03:38 | |
-!- ThomasEgi [~thomas@185.5.8.81] has quit [Changing host] | 03:38 | |
-!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap | 03:38 | |
-!- Viper168_ [~Viper@unaffiliated/viper168] has quit [Ping timeout: 252 seconds] | 03:41 | |
-!- kuldeepdhaka [~kuldeepdh@unaffiliated/kuldeepdhaka] has joined ##hplusroadmap | 03:44 | |
-!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 03:49 | |
-!- delinquentme [~dingo@74.61.157.78] has quit [Ping timeout: 240 seconds] | 03:55 | |
-!- sheena [~home@d108-180-136-217.bchsia.telus.net] has joined ##hplusroadmap | 04:28 | |
-!- AshleyWaffle [~waffle@gateway/tor-sasl/anastasiawyatt] has quit [Write error: Connection reset by peer] | 04:33 | |
-!- AshleyWaffle [~waffle@gateway/tor-sasl/anastasiawyatt] has joined ##hplusroadmap | 04:36 | |
-!- sheena [~home@d108-180-136-217.bchsia.telus.net] has quit [Ping timeout: 255 seconds] | 04:38 | |
-!- kuldeepdhaka [~kuldeepdh@unaffiliated/kuldeepdhaka] has quit [Ping timeout: 265 seconds] | 04:50 | |
-!- sheena [~home@d108-180-136-217.bchsia.telus.net] has joined ##hplusroadmap | 05:03 | |
-!- FourFire [~fourfire@67-149-15.connect.netcom.no] has joined ##hplusroadmap | 05:04 | |
-!- eudoxia [~eudoxia@r179-25-152-221.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap | 05:05 | |
-!- nsh [~nsh@host86-158-32-127.range86-158.btcentralplus.com] has joined ##hplusroadmap | 05:18 | |
* pasky scratches head at http://arxiv.org/abs/1401.1219 | 05:26 | |
-!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Remote host closed the connection] | 05:33 | |
-!- nmz787_i [~nmccorkx@192.55.54.42] has joined ##hplusroadmap | 05:39 | |
-!- nmz787_i [~nmccorkx@192.55.54.42] has quit [Client Quit] | 05:39 | |
-!- nsh [~nsh@host86-158-32-127.range86-158.btcentralplus.com] has quit [Ping timeout: 255 seconds] | 05:54 | |
-!- Lemminkainen [uid2346@gateway/web/irccloud.com/x-lnldsqvowfkrvorx] has quit [Quit: Connection closed for inactivity] | 06:16 | |
-!- ebowden_ [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has joined ##hplusroadmap | 06:17 | |
-!- ebowden [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has quit [Ping timeout: 252 seconds] | 06:20 | |
-!- marciogm [~marciogm@179.126.61.80] has joined ##hplusroadmap | 06:20 | |
mosasaur | pasky: I think they're still trying to solve the problem that quantum measurements require an observer. | 06:36 |
-!- marciogm [~marciogm@179.126.61.80] has quit [Quit: Computer has gone to sleep.] | 06:41 | |
-!- FourFire [~fourfire@67-149-15.connect.netcom.no] has quit [Read error: Connection reset by peer] | 06:46 | |
ebowden_ | paperbot: http://www.sciencedirect.com/science/article/pii/S0271531705001041 | 06:47 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/ad2643a7c38c723359b9773d838f7b59.txt | 06:47 |
ebowden_ | paperbot: http://www.nrjournal.com/article/S0271-5317(05)00104-1/pdf | 06:47 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/d010a7e0119e15952a7d1e1d2777a7c.txt | 06:47 |
-!- FourFire [~fourfire@67-149-15.connect.netcom.no] has joined ##hplusroadmap | 06:47 | |
ebowden_ | Oh, hello FourFire. | 06:48 |
FourFire | ello | 06:48 |
ebowden_ | paperbot: http://www.journals.elsevierhealth.com/periodicals/ntr/article/PIIS0271531705001041/abstract | 06:49 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/65eed9b74197844031fb7fb611ba259f.txt | 06:49 |
ebowden_ | paperbot: http://www.nrjournal.com/article/S0271-5317(05)00104-1/pdf | 06:50 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/a61172a0bfd230e00134d62372fb488b.txt | 06:50 |
ebowden_ | Damnit. | 06:50 |
ebowden_ | I am not good at paperbot. | 06:51 |
ebowden_ | paperbot: http://download.journals.elsevierhealth.com/pdfs/journals/0271-5317/PIIS0271531705001041.pdf | 06:52 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/2db8ec5f4524416ddc08562c59b3b358.txt | 06:52 |
ebowden_ | :( | 06:52 |
ebowden_ | Does anyone here know how to properly use paperbot? | 06:54 |
mosasaur | ebowden_: is it supposed to extract the javascript generated content? | 06:55 |
ebowden_ | I have absolutely no idea how it works, but normally it gives you a nice pretty PDF. | 06:56 |
* mosasaur wonders what it does with a google groups page | 06:56 | |
mosasaur | paperbot: https://groups.google.com/forum/#!forum/endgame-singularity | 06:58 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/3a73de0d7d8f7aaac21d63537a91056a.txt | 06:58 |
mosasaur | nope :-( | 06:58 |
ebowden_ | Oh, no, it extracts a PDF text file. | 06:59 |
ebowden_ | Doesn't create one. | 06:59 |
-!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 240 seconds] | 07:04 | |
-!- Viper168_ [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 07:05 | |
-!- mosasaur [~mosasaur@95.99.48.20] has quit [Quit: Leaving.] | 07:06 | |
@kanzure | ebowden_: instructions are here, https://github.com/kanzure/paperbot | 07:14 |
ebowden_ | Ah, thanks. | 07:14 |
ebowden_ | Well, night. | 07:14 |
-!- ebowden_ [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has quit [Remote host closed the connection] | 07:16 | |
@kanzure | http://info.5amsolutions.com/blog/bid/157752/Venter-s-Human-Longevity-Academic-vs-Private-Genomics-Round-2 | 07:20 |
@kanzure | http://www.humanlongevity.com/ | 07:20 |
@kanzure | "hey plan to sequence between 40,000 and 100,000 genomes a year. To do that, they’ve purchased two of Illumina’s new high-end DNA sequencing devices, the HiSeq X Ten, which is advertised as able to sequence a complete human genome for the long-sought-after price of $1,000." | 07:20 |
@kanzure | "So far they’ve raised $70M for this venture. $20M will go for those two sequencers, and $40M will go to sequence those first 40,000 genomes, so no doubt they’ll be looking for other sources of funding soon. Given Venter’s success at Celera in building interest in the potential commercial value of the human genome, I’d say he’ll have little trouble raising more funds - Celera pulled in over $1B from their public offerings of stock." | 07:21 |
@kanzure | "The company, according to their press release, is “focused on extending the healthy, high performance human life span” and is going to “tackle the diseases associated with aging-related human biological decline”." | 07:21 |
cluckj | o_O | 07:22 |
@kanzure | diamandis is on the board of directors haha | 07:22 |
cluckj | deCode part 2? | 07:22 |
@kanzure | "HLI has entered into a collaboration with Metabolon Inc., the world’s leader in the field of metabolomics. Metabolon has developed proprietary technology to quickly identify and measure the body’s biochemicals and will provide such biochemical profiling services to HLI. In the initial term of the agreement Metabolon will carry out small molecule analysis of 10,000 subjects and collaborate with HLI to map changes in the small molecules to ... | 07:23 |
@kanzure | ... end points of disease and gene mutations." | 07:23 |
cluckj | so craig venter wants to live forever, now? :) | 07:23 |
@kanzure | oh cool they have a public phone number: 858-249-7500 | 07:24 |
@kanzure | totally going to prank call them | 07:24 |
cluckj | ask if their $10 million sequencer is running | 07:24 |
FourFire | kanzure, I'm all for cheaper sequencing | 07:25 |
FourFire | do they have orders, or are they just doing it on random people? | 07:26 |
FourFire | (first 40 000) | 07:26 |
@kanzure | dunno, so far the pitch seems sort of fucked up | 07:26 |
@kanzure | give us $70M so we can buy some dna sequencers and do lots of sequencing? | 07:26 |
FourFire | who are they sequencing though? | 07:27 |
@kanzure | why does that matter? | 07:27 |
cluckj | there's probably more going on there than their pitch suggests | 07:27 |
@kanzure | craig has cancer? | 07:27 |
@kanzure | hrmm no he would have done something cancer-specific | 07:27 |
FourFire | well, if they could offer a, say 80% off sequencing, 250$ per genome after they are already funded, they'll be turning a profit | 07:27 |
cluckj | he's getting old; wants to live forever | 07:28 |
cluckj | I dunno | 07:28 |
@kanzure | FourFire: i don't think this is a generic sequencing company | 07:28 |
@kanzure | FourFire: nothing about this says "we want to sell direct-to-consumer sequencing services" | 07:28 |
FourFire | 250$ is near where I would consider it worthwhile to get my genome sequenced | 07:29 |
@kanzure | cluckj: i guess that $600 million from exxon is about empty | 07:29 |
cluckj | lol | 07:29 |
FourFire | kanzure, which 600 million? | 07:29 |
cluckj | craig's M.O. is to find things to patent | 07:29 |
cluckj | so what is patentable there? | 07:29 |
@kanzure | FourFire: craig venter got a $600 million deal with exxon for his company, synthetic genomics inc. | 07:30 |
FourFire | oh, right | 07:30 |
-!- catern [~catern@catern.com] has quit [Ping timeout: 240 seconds] | 07:30 | |
@kanzure | fenn: the entire world is a strange loop that folds back into itself roughly every 10 years https://www.allegro.cc/forums/thread/345403 | 07:35 |
-!- Qfwfq [~Qfwfq@cm113.kappa36.maxonline.com.sg] has joined ##hplusroadmap | 07:36 | |
-!- Qfwfq [~Qfwfq@cm113.kappa36.maxonline.com.sg] has quit [Changing host] | 07:36 | |
-!- Qfwfq [~Qfwfq@unaffiliated/washirving] has joined ##hplusroadmap | 07:36 | |
-!- catern [~catern@catern.com] has joined ##hplusroadmap | 07:37 | |
@kanzure | "Hellman: I smile when you say Merkle, I mean Merkle is just someone who makes you smile. He’s a comic. He comes and plops down in your office—have you met Ralph?" | 07:55 |
@kanzure | "Hellman: He plops down in your chair and says, ‘Hi!’ I remember, this was after the discovery of public key cryptography, maybe fifteen years ago, he comes into my office and plops down and says, ‘Hi, I’m building a human brain.’ He’s one of the stars in nanotechnology. Building human brains and repairing human brains on dead people so you can bring them back to life some time in the distant future is one of his passions." | 07:55 |
-!- nsh [~nsh@host217-43-193-5.range217-43.btcentralplus.com] has joined ##hplusroadmap | 07:59 | |
@kanzure | i think i need some soot goggles for interneting from now on http://www.entertainmentearth.com/images/AUTOIMAGES/EP828410lg.jpg | 08:00 |
@kanzure | jgarzik bitcoin satellite groupies https://groups.google.com/forum/#!forum/bitsat-project | 08:02 |
@kanzure | http://www.google.com/url?q=http%3A%2F%2Fwww.dunveganspace.com%2Fgoals%2Fbitsat%2FBitSatUpdate1.pdf&sa=D&sntz=1&usg=AFQjCNFzqytlkxLDkAz0J39nZCB-pbY9gA | 08:02 |
@kanzure | oops i mean http://www.dunveganspace.com/goals/bitsat/BitSatUpdate1.pdf | 08:02 |
@kanzure | http://www.dunveganspace.com/goals/bitsat/BitSatArchitecture-0.1.pdf | 08:03 |
-!- nsh_ [~nsh@wikipedia/nsh] has joined ##hplusroadmap | 08:05 | |
-!- nsh [~nsh@host217-43-193-5.range217-43.btcentralplus.com] has quit [Disconnected by services] | 08:05 | |
-!- nsh_ is now known as nsh | 08:05 | |
FourFire | lol, Kanzure you saying "back in the day" what are you, early 30s ? | 08:07 |
@kanzure | 24 | 08:07 |
FourFire | late 20s? | 08:07 |
FourFire | so when you were 14, you were saying "back in the day" | 08:08 |
@kanzure | internet accelerates your aging by a factor of like 1000x | 08:08 |
FourFire | I doubt it, I should be much more mature by now if that was the case | 08:08 |
@kanzure | or your curve is different | 08:08 |
FourFire | or perhaps "this here web 2.0 bullshit ain like it used to be... back in the days" | 08:08 |
@kanzure | yep, first phase of web 2.0 was the real bullshit | 08:09 |
FourFire | "I used to know 10% of the internet, by name! back in the day!" | 08:09 |
-!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 08:10 | |
@kanzure | 14 year olds can't have days to talk back about? | 08:10 |
-!- Viper168_ [~Viper@unaffiliated/viper168] has quit [Ping timeout: 276 seconds] | 08:12 | |
-!- marciogm [~marciogm@179.126.61.80] has joined ##hplusroadmap | 08:16 | |
-!- catern [~catern@catern.com] has quit [Ping timeout: 276 seconds] | 08:21 | |
-!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has joined ##hplusroadmap | 08:24 | |
FourFire | well, sure they can. it just seems pretentious to me, but then I probably said similar things when I was 14. | 08:28 |
-!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has quit [Quit: Lost terminal] | 08:30 | |
@kanzure | (not an actual recommendation) http://www.amazon.com/The-Knowledge-Rebuild-World-Scratch/dp/159420523X | 08:30 |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Remote host closed the connection] | 08:37 | |
-!- catern [~catern@catern.com] has joined ##hplusroadmap | 08:37 | |
@kanzure | long video about ants https://www.youtube.com/watch?v=Z-gIx7LXcQM | 08:40 |
@kanzure | eudoxia: http://web.archive.org/web/19980109011052im_/http://extropians.mit.edu/extropy.gif | 08:44 |
eudoxia | kanzure: it's rather sad that the MIT extropians were run by a bunch of racists and fell apart | 08:47 |
@kanzure | huh? | 08:48 |
eudoxia | http://web.mit.edu/observer/www/1-1/articles/ad1.html | 08:49 |
eudoxia | there was some drama and they disbanded or something | 08:50 |
-!- HashNuke [uid12117@gateway/web/irccloud.com/x-kjqbppokrvrdkvfj] has joined ##hplusroadmap | 08:53 | |
-!- QuadIngi [~fourfire@22-135-15.connect.netcom.no] has joined ##hplusroadmap | 08:55 | |
@kanzure | neat, chrome has been handling 326 tabs open for 15 days | 08:57 |
-!- FourFire [~fourfire@67-149-15.connect.netcom.no] has quit [Ping timeout: 264 seconds] | 08:58 | |
-!- nmz787_i [nmccorkx@nat/intel/x-psojhvhjxdqquddc] has joined ##hplusroadmap | 09:13 | |
-!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 09:15 | |
pasky | hmm it seems my 360-tab firefox is running for 11 days now | 09:30 |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 09:30 | |
pasky | sooner or later it ends up eating 100% CPU for no easily discernable reason and I have to restart it though | 09:31 |
-!- nmz787_i [nmccorkx@nat/intel/x-psojhvhjxdqquddc] has quit [Quit: Leaving.] | 09:32 | |
dingo | i recommend the tree style tab plugin for FF | 09:40 |
-!- nmz787_i [~nmccorkx@192.55.54.40] has joined ##hplusroadmap | 09:40 | |
@kanzure | tree style tab wont solve his memory problems | 09:41 |
nmz787_i | paperbot: http://ar.iiarjournals.org/content/33/11/4757.full.pdf | 09:43 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/b7cc5d5fdf484e246df172676a9922e9.txt | 09:43 |
dingo | i recommend a strict diet of lynx | 09:44 |
xmj | xombrero is nice for that | 09:45 |
-!- mosasaur [~mosasaur@95.99.48.20] has joined ##hplusroadmap | 09:45 | |
@kanzure | i would be okay with not using tabs if page caching actually worked | 09:45 |
dingo | if you really want to get hardcore, try a web browser written by a blind guy (Dave Khale, cool dude) http://the-brannons.com/edbrowse/ | 09:46 |
@kanzure | braille interfaces are obscenely expensive | 09:46 |
dingo | karl dahlke rather, he does a good math site i think | 09:47 |
dingo | http://eklhad.net/ | 09:47 |
@ParahSailin | yeah i think blind people generally use modern browsers with screen readers | 09:49 |
nmz787_i | kanzure: with that cefpython module... do you think we could actually write a chrome-based tab caching browser? Or would that work as a chrome extension? | 09:53 |
@kanzure | i would use webkit/blink before cefpython | 09:53 |
@kanzure | webkit has a .haf format but i don't know if it stores all in-memory objects correctly | 09:54 |
@kanzure | another option is to use mitmproxy to capture all resources and just replay it in the future | 09:54 |
@kanzure | so you don't have to keep the tab loaded in memory, and later can reload it without actually hitting the real interwebs | 09:54 |
@kanzure | if the web page is non-deterministic then it might cause other requests to be made when the page is reloaded in the future | 10:01 |
nmz787_i | isn't blink what chromium uses already? | 10:04 |
@kanzure | chromium has a bunch of other stuff on top of blink | 10:04 |
nmz787_i | aren't those desirable things? | 10:04 |
nmz787_i | 'features' | 10:04 |
@kanzure | dunno, the answer is most likely no, because all those 'features' are why you can't have 10000 tabs in firefox | 10:05 |
nmz787_i | firefox doesn't use chromium though | 10:05 |
@kanzure | whatever :) | 10:05 |
@kanzure | "Functional electrical stimulation-facilitated | 10:10 |
@kanzure | "Functional electrical stimulation-facilitated proliferation and regeneration of neural precursor" | 10:11 |
@kanzure | paperbot: http://www.sjzsyj.org/CN/article/openArticlePDF.jsp?id=862 | 10:11 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/7ba419edd532526e0b5359e818fcc72b.pdf | 10:11 |
@kanzure | "Prehensile traction test results showed that, at 14 days, prehension function of rats in the functional electrical stimulation group was significantly better than in the placebo group. These results suggest that functional electrical stimulation can promote endogenous neural precursor cell proliferation in the brains of acute cerebral infarction rats, enhance expression of basic fibroblast growth factor and epidermal growth factor, and ... | 10:12 |
@kanzure | ... improve the motor function of rats." | 10:12 |
@kanzure | "yo dawg we got 100 rats and we cut their brain blood flow for a bit, then we shocked them with electricity for a few weeks because science, it's all good" | 10:15 |
-!- Adifex|zzz is now known as Adifex | 10:18 | |
nmz787_i | so how do i contact langton labs to see about sleeping there? | 10:19 |
@kanzure | take blood oath at burning man? | 10:19 |
nmz787_i | that is after makerfaire | 10:20 |
nmz787_i | so won't work | 10:20 |
@kanzure | .tell fenn nmz787_i would like information about contacting or staying at langton labs | 10:21 |
yoleaux | kanzure: I'll pass your message to fenn. | 10:21 |
-!- marciogm [~marciogm@179.126.61.80] has quit [Quit: Computer has gone to sleep.] | 10:24 | |
-!- kuldeepdhaka [~kuldeepdh@unaffiliated/kuldeepdhaka] has joined ##hplusroadmap | 10:26 | |
pasky | hmm seems like switching from vertical tabs to tree-style tabs is not so trivial | 10:27 |
-!- juri_ [juri@funkykitty.net] has quit [Ping timeout: 240 seconds] | 10:29 | |
nmz787_i | this is pretty cool https://www.sparkfun.com/products/12042 | 10:31 |
@kanzure | .title | 10:31 |
nmz787_i | Z-axis conductive tape | 10:31 |
yoleaux | Z-Axis Conductive Tape | 10:31 |
nmz787_i | gtfo yoleaux, this is my link! | 10:31 |
@kanzure | i was impatient | 10:31 |
nmz787_i | psh | 10:31 |
nmz787_i | stupid latent bot | 10:31 |
nmz787_i | all 3 seconds slower than me | 10:32 |
nmz787_i | is yoleaux a play on YOLO? | 10:36 |
@kanzure | yes, ask the #swhack people | 10:36 |
@kanzure | huh, opencascade "community edition" still has activity https://github.com/tpaviot/oce | 10:39 |
-!- marciogm [~marciogm@186-210-025-170.xd-dynamic.ctbcnetsuper.com.br] has joined ##hplusroadmap | 10:43 | |
-!- QuadIngi [~fourfire@22-135-15.connect.netcom.no] has quit [Ping timeout: 240 seconds] | 10:44 | |
@kanzure | that guy with the ABI 310 w/o software is going to have problems | 10:58 |
-!- mosasaur [~mosasaur@95.99.48.20] has quit [Quit: Leaving.] | 10:58 | |
@kanzure | i think i'm going to recommend he sets up linux in front of the machine | 10:58 |
@kanzure | and then sets up an ssh server for me to login to | 10:58 |
@kanzure | and then i'll poke around the machine | 10:59 |
@kanzure | but i think i need to get a ROM dump of whatever microcontroller is inside the machine | 10:59 |
-!- fireprfHydra [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has joined ##hplusroadmap | 10:59 | |
nmz787_i | Whoo! http://www.nature.com/news/start-up-investor-bets-on-biotech-1.15096 | 11:04 |
nmz787_i | Y combinator application here I come! | 11:04 |
-!- delinquentme [~dingo@74.61.157.78] has quit [Ping timeout: 276 seconds] | 11:04 | |
-!- fireprfHydra_ [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has joined ##hplusroadmap | 11:05 | |
-!- fireprfHydra [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has quit [Ping timeout: 265 seconds] | 11:05 | |
-!- kyknos_ [~kyknos@89.233.130.143] has quit [Ping timeout: 252 seconds] | 11:05 | |
chris_99 | cool, good luck nmz787 | 11:06 |
-!- chido [chidori@pasky.or.cz] has joined ##hplusroadmap | 11:09 | |
-!- yorick [~yorick@oftn/member/yorick] has joined ##hplusroadmap | 11:13 | |
-!- ElixirVitae [~Shehrazad@unaffiliated/shehrazad] has quit [Ping timeout: 252 seconds] | 11:16 | |
@kanzure | heh i should have expected substack to have a thing for technical debt https://github.com/substack/codebux | 11:16 |
-!- eudoxia [~eudoxia@r179-25-152-221.dialup.adsl.anteldata.net.uy] has quit [Quit: leaving] | 11:25 | |
-!- ElixirVitae [~Shehrazad@95.5.71.53] has joined ##hplusroadmap | 11:27 | |
-!- ElixirVitae [~Shehrazad@95.5.71.53] has quit [Changing host] | 11:27 | |
-!- ElixirVitae [~Shehrazad@unaffiliated/shehrazad] has joined ##hplusroadmap | 11:27 | |
-!- fireprfHydra_ [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has quit [Ping timeout: 265 seconds] | 11:27 | |
-!- marciogm [~marciogm@186-210-025-170.xd-dynamic.ctbcnetsuper.com.br] has quit [Quit: Computer has gone to sleep.] | 11:36 | |
-!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 265 seconds] | 11:41 | |
-!- fireprfHydra [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has joined ##hplusroadmap | 11:44 | |
-!- marciogm [~marciogm@186-210-025-170.xd-dynamic.ctbcnetsuper.com.br] has joined ##hplusroadmap | 11:47 | |
-!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 11:50 | |
-!- AshleyWaffle [~waffle@gateway/tor-sasl/anastasiawyatt] has quit [Ping timeout: 272 seconds] | 11:51 | |
-!- AshleyWaffle [~waffle@gateway/tor-sasl/anastasiawyatt] has joined ##hplusroadmap | 11:59 | |
-!- fireprfHydra_ [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has joined ##hplusroadmap | 11:59 | |
-!- fireprfHydra [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has quit [Ping timeout: 265 seconds] | 12:02 | |
-!- Lemminkainen [uid2346@gateway/web/irccloud.com/x-lxjgcbfktqsfuzlh] has joined ##hplusroadmap | 12:03 | |
@kanzure | this is very cool https://github.com/twolfson/phantomjsify | 12:07 |
@kanzure | oh, not much is shimmed. hrm. | 12:08 |
-!- augur [~augur@216-164-48-148.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Remote host closed the connection] | 12:14 | |
-!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 12:17 | |
@kanzure | "IBM’s chip business needs help. So the company has opened up the technology of its Power microprocessors, inviting others to modify and manufacture Power-based designs pretty much as they see fit. This open, liberal licensing initiative is conducted under the auspices of the OpenPower Foundation, which was incorporated in December." | 12:25 |
@kanzure | http://openpowerfoundation.org/ | 12:26 |
dingo | they did that with opensparc, nobody did anything with it unfortunately | 12:26 |
@kanzure | hmm no files. | 12:26 |
@kanzure | what's up with a separate foundation for each idea? | 12:27 |
@kanzure | surely ibm owns a non-profit about something | 12:27 |
-!- nsh [~nsh@wikipedia/nsh] has quit [Ping timeout: 264 seconds] | 12:29 | |
-!- mode/##hplusroadmap [-o kanzure] by kanzure | 12:38 | |
-!- nsh [~nsh@wikipedia/nsh] has joined ##hplusroadmap | 12:38 | |
-!- ThomasEgi [~thomas@185.5.8.81] has joined ##hplusroadmap | 12:44 | |
-!- ThomasEgi [~thomas@185.5.8.81] has quit [Changing host] | 12:44 | |
-!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap | 12:44 | |
-!- nmz787_i [~nmccorkx@192.55.54.40] has quit [Ping timeout: 255 seconds] | 12:46 | |
-!- nsh_ [~nsh@wikipedia/nsh] has joined ##hplusroadmap | 12:47 | |
-!- nsh [~nsh@wikipedia/nsh] has quit [Ping timeout: 255 seconds] | 12:48 | |
-!- xentrac [~kragen@panacea.canonical.org] has quit [Ping timeout: 276 seconds] | 12:50 | |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Leaving] | 13:03 | |
-!- augur [~augur@129-2-129-33.wireless.umd.edu] has joined ##hplusroadmap | 13:04 | |
-!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 252 seconds] | 13:06 | |
-!- fireprfHydra_ [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has quit [Ping timeout: 264 seconds] | 13:09 | |
-!- HashNuke [uid12117@gateway/web/irccloud.com/x-kjqbppokrvrdkvfj] has quit [Quit: Connection closed for inactivity] | 13:12 | |
-!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 13:13 | |
-!- fireprfHydra [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has joined ##hplusroadmap | 13:22 | |
-!- delinquentme [~dingo@74.61.157.78] has quit [Ping timeout: 276 seconds] | 13:38 | |
-!- augur [~augur@129-2-129-33.wireless.umd.edu] has quit [Remote host closed the connection] | 13:49 | |
-!- nmz787_i [~nmccorkx@134.134.137.75] has joined ##hplusroadmap | 13:52 | |
-!- augur [~augur@129-2-129-33.wireless.umd.edu] has joined ##hplusroadmap | 13:56 | |
-!- nsh_ is now known as nsh | 14:01 | |
-!- fireprfHydra [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has quit [Ping timeout: 265 seconds] | 14:13 | |
-!- xentrac [~kragen@panacea.canonical.org] has joined ##hplusroadmap | 14:15 | |
-!- chido [chidori@pasky.or.cz] has quit [Quit: Lost terminal] | 14:21 | |
-!- juri_ [~juri@vpn166.sdf.org] has joined ##hplusroadmap | 14:25 | |
-!- marciogm [~marciogm@186-210-025-170.xd-dynamic.ctbcnetsuper.com.br] has quit [Quit: Computer has gone to sleep.] | 14:27 | |
-!- QuadIngi [~fourfire@245-201-15.connect.netcom.no] has joined ##hplusroadmap | 14:31 | |
-!- yorick [~yorick@oftn/member/yorick] has quit [Remote host closed the connection] | 14:33 | |
-!- EnLilaSko [EnLilaSko@unaffiliated/enlilasko] has quit [Remote host closed the connection] | 14:38 | |
-!- kardan [~kardan@2a02:810d:1100:af8:95b1:6157:5d37:95e2] has joined ##hplusroadmap | 14:51 | |
-!- fireprfHydra [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has joined ##hplusroadmap | 14:54 | |
kanzure | "Back in the 60s NASA considered a plan for a manned flyby of Venus using Apollo-derived hardware. I've wondered if a modern day manned flyby mission might make sense. A rover could be landed on Venus more easily than on Mars, but it would, by necessity, have a very short lifetime on the surface. Once it's out of coolant, the mission is over. So if you were going to land a rover (or rovers) on the surface of Venus, you'd need to maximize ... | 14:59 |
kanzure | ... their effectiveness. One way to do this would be to teleoperate them in near real-time, which could be done if rover landing were coordinated with a manned flyby." | 14:59 |
kanzure | "A manned flyby of Venus would be substantially cheaper and less risky than a manned mission to Mars. There's no manned landing component, and the flyby mission itself would take a little over a year, which is pushing the limit for human exposure to micogravity, but within the range that we have actual experience. This kind of mission will, of course, only make sense if you are an advocate of manned space exploration in the first place." | 14:59 |
kanzure | https://news.ycombinator.com/item?id=7633515 | 14:59 |
-!- kyknos_ [~kyknos@89.233.130.143] has joined ##hplusroadmap | 15:00 | |
kanzure | 15-20% margins on reselling modafinil http://www.gwern.net/Modafinil#margin-estimation | 15:02 |
kanzure | also there's a simple assay mentioned down the page | 15:03 |
kanzure | http://www.pharmacyreviewer.com/forum/discussion-online-pharmacies-featured-general-reviews-section/16335-edandmore-feedbacks-10.html#post139862 | 15:03 |
kanzure | "Modafinil is a sulfa drug, containing a certain chemical group called a sulfonamide group66. And lo and behold I stumbled across a simple pair of chemical tests for sulfonamides." | 15:03 |
kanzure | "I took about a third of a pill and placed it in a test tube, adding a few ml of dilute NaOH. I then mixed it up and heated it over an alcohol burner. This should produce ammonia, which has a really characteristic odor. I got a whiff of my tube and, indeed, it was ammonia! I also tested the fumes with a piece of litmus paper - it turned blue, as expected [litmus turns red for acids & blue for bases; ammonia is a base]. Then I put another ... | 15:03 |
kanzure | ... third of a pill in another test tube and added dilute HCl. Upon heating it, sulfur dioxide should be produced - another gas with a characteristic, pungent odor. So I sniffed my tube - it smelled awful! Further, my litmus paper turned red, which is what I’d expect because SO2 is acidic." | 15:04 |
kanzure | "Finally, one other test - I wanted to make sure this isn’t a characteristic of pill binder substances or anything. So I took half a caffeine pill and did the NaOH -> NH3 test on it. No ammonia whatsoever." | 15:04 |
kanzure | "This test is crude and likely produces many false positives and negatives67, but may still be worth using. Checking, 50+ strips of pH/litmus paper is ~$5; hydrochloric acid is harder to find, but seems to be obtainable at $10-20 online; and sodium hydroxide similarly (and no doubt purchasable cheaper locally), for a worst-case cost of $45. This is roughly a third of EcstasyData.org’s price, and enough to test 50+ samples at a worst-case ... | 15:04 |
kanzure | ... cost of ~$1 per sample." | 15:04 |
QuantumG | ha, I didn't know it was a sulfa drug.. good I never got my hands on it, eh? (apparently I'm allergic.) | 15:04 |
kanzure | how apparent is apparently | 15:04 |
QuantumG | like, it's written on my birth certificate, but I've never experienced the symptoms | 15:05 |
kanzure | allergies are written on birth certs? | 15:05 |
QuantumG | weird eh | 15:05 |
nmz787_i | apparently you can get a passport if you don't have a birth certificate, but have your name and birthdate written in a 'family bible' (you send the passport office your bible, potentially more evidence) | 15:06 |
nmz787_i | just what i heard recently | 15:06 |
kanzure | so does that mean i can mail them a book of the jedi order? | 15:07 |
-!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has joined ##hplusroadmap | 15:07 | |
nmz787_i | i dont think os | 15:08 |
nmz787_i | so | 15:08 |
QuantumG | mailing the three volumes of Knuth would be expensive. | 15:10 |
-!- chido [chidori@pasky.or.cz] has joined ##hplusroadmap | 15:11 | |
-!- kuldeepdhaka [~kuldeepdh@unaffiliated/kuldeepdhaka] has quit [Ping timeout: 265 seconds] | 15:17 | |
kanzure | "The origin of turning dates to around 1300 BCE when the Ancient Egyptians first developed a two-person lathe. One person would turn the wood work piece with a rope while the other used a sharp tool to cut shapes in the wood. Ancient Rome improved the Egyptian design with the addition of a turning bow. In the Middle Ages a pedal replaced hand-operated turning, freeing both the craftsman's hands to hold the woodturning tools. The pedal was ... | 15:17 |
kanzure | ... usually connected to a pole, often a straight-grained sapling. The system today is called the "spring pole" lathe. Spring pole lathes were in common use into the early 20th century." | 15:17 |
kanzure | http://en.wikipedia.org/wiki/Lathe | 15:17 |
kanzure | http://historicgames.com/lathes/Egyptstone.jpg | 15:18 |
kanzure | http://argenteriedesbauges.free.fr/1egyptien.JPG | 15:19 |
kanzure | pfft obviously they should have invented the four-person lathe | 15:19 |
-!- xentrac [~kragen@panacea.canonical.org] has quit [Ping timeout: 255 seconds] | 15:20 | |
-!- chido [chidori@pasky.or.cz] has quit [Quit: Lost terminal] | 15:23 | |
fenn | pixtronix optical MEMS shutter transflective color low power display, another neat technology buried in the depths of qualcomm's dungeons: http://www.the-digital-reader.com/2011/12/08/qualcomm-bought-pixtronix-a-mirasol-competitor/ http://ko.com.ua/files/Pixtronix_Direct_View_MEMS_%20Display_Whitepaper.pdf | 15:24 |
yoleaux | 17:21Z <kanzure> fenn: nmz787_i would like information about contacting or staying at langton labs | 15:24 |
fenn | the video is too long; summary: looks like a plasma tv, sizes as big or small as you want, about 25% of backlit lcd power consumption, and works in full sunlight with reduced color gamut, or black and white with no backlight | 15:27 |
-!- xentrac [~kragen@panacea.canonical.org] has joined ##hplusroadmap | 15:28 | |
fenn | this is why we need intellectual property taxes | 15:28 |
nmz787_i | do you still have contacts at langton fenn?\ | 15:28 |
fenn | i don't know; i haven't checked my email for about a year | 15:29 |
fenn | i suspect you'd be welcome | 15:29 |
fenn | you just want to stay for a weekend ish? | 15:30 |
nmz787_i | yeah, 4 folks total, we'll have tents so yard/slab is fine also... 2 of us will share abed | 15:30 |
-!- kuldeepdhaka [~kuldeepdh@unaffiliated/kuldeepdhaka] has joined ##hplusroadmap | 15:30 | |
nmz787_i | so 3 'beds' | 15:30 |
nmz787_i | fenn: do you know Cadence Allegro (PCB software) and want a job in Oregon for a bit? | 15:31 |
nmz787_i | the job also involves Python, and would be a contract with Intel through a temp agency | 15:32 |
-!- fireprfHydra [~fireprfHy@pool-173-70-216-225.nwrknj.fios.verizon.net] has quit [Ping timeout: 252 seconds] | 15:32 | |
kanzure | don't you think he has better things to be doing than bothering with a temp agency? | 15:32 |
QuantumG | I know someone who might | 15:32 |
nmz787_i | some kind of PCB routing thin, i think | 15:32 |
nmz787_i | kanzure: no idea, seems like his kind of thing though, and it would be a reason to live here during the summer | 15:33 |
nmz787_i | QuantumG: get them to send me a resume nmz787 at gmail | 15:33 |
kanzure | once you accumulate any non-negligble amount of skill, you want to avoid temp agencies like the plague | 15:34 |
nmz787_i | have you been to Oregon during the summer? | 15:34 |
nmz787_i | having any reason to come here is worth it | 15:34 |
nmz787_i | especially since its a contract | 15:34 |
nmz787_i | come visit, see the sights, GTFO if you want | 15:35 |
cpopell | /query nmz787_i | 15:36 |
cpopell | whoops! | 15:36 |
cpopell | Anyway, nmz787_i, are you involved with a staffing agency up there, or did you just hear about it? | 15:36 |
nmz787_i | no, my manager is reviewing resumes now | 15:36 |
cpopell | oh, k. | 15:37 |
-!- xentrac [~kragen@panacea.canonical.org] has quit [Ping timeout: 240 seconds] | 15:37 | |
fenn | nmz787_i: there is no yard/slab at langton. 4 people is the full capacity of the guest room so that may not be possible, especially with other people in town for maker faire | 15:40 |
fenn | i've never used Cadence | 15:41 |
fenn | what kind of temp agency hires EDA professionals? | 15:42 |
fenn | that's like, a real job | 15:42 |
kanzure | i assume there's a split in the industry | 15:43 |
kanzure | where there's some who are doing vsli stuff | 15:43 |
kanzure | and then others who are like blender/maya monkeys | 15:43 |
kanzure | and i don't mean the "write beautiful code for dreamworks inc to automate 3d production pipelines", i mean "clicking buttons in blender all day for weeks" | 15:43 |
kanzure | *vlsi | 15:44 |
fenn | that's what blender is for though | 15:44 |
kanzure | with programming you can take advantage of efficiencies that you introduce, with point and click this is less true | 15:44 |
kanzure | when you need a bunch of labor to click, you go pick a temp agency to supply you with labor | 15:45 |
kanzure | just theorizing here, re: your observation about a "real" job | 15:45 |
kanzure | +. | 15:46 |
-!- xentrac [~kragen@panacea.canonical.org] has joined ##hplusroadmap | 15:46 | |
fenn | I respect 3d artists like http://www.mikeanash.com/p/hard-surface-bust.html who do nothing but click on buttons all day | 15:48 |
dingo | i'm not seeing it in the buffer, which specific industry? | 15:48 |
dingo | just 3d modelling? | 15:48 |
fenn | he's just ranting about point and click interfaces, pay no attention | 15:48 |
fenn | kanzure: okay but the clicking is to _DO_ something, and it's the thought that goes into where you click that is important | 15:49 |
kanzure | i don't think i've claimed that thinking is unimportant | 15:49 |
dingo | well i dunno, i used to make a lot of quake 1 levels, its a lot of point and clicking, but its also "programming", not my cup of tea really... no opinion ... i did write plug-ins and know ppl who do plug-ins for 3d modelling software -- now thats great fun, working with vectors and nodes and such | 15:49 |
QuantumG | how about some division of labor? Hire the artist to do artistic shit and the programmer to automate his crappy workflow. | 15:49 |
fenn | it takes skill, experience, and education to design and "capture" an electronic circuit/layout | 15:49 |
-!- augur [~augur@129-2-129-33.wireless.umd.edu] has quit [Remote host closed the connection] | 15:50 | |
kanzure | also, i'd point out that nate specifically mentioned a python requirement, so there's obv. some programming expectations | 15:50 |
dingo | many gaming and movie studios have tirtuary tooling -- mostly the junior programmers | 15:50 |
kanzure | the junior programmers build their tools? | 15:50 |
fenn | tirtuary? | 15:50 |
kanzure | tertiary? | 15:50 |
QuantumG | then the artist can get more arty stuff done, resulting in more product | 15:50 |
dingo | like i know a guy at uhh whats it called, rockstar for gta, right? | 15:50 |
dingo | all he does is make gui interfaces for backend processing tools, none of it is in the game, just used to build the game | 15:51 |
fenn | we call that development tools | 15:51 |
nmz787_i | it is a real job, but they might not want to keep you around to maintain the software after its written | 15:51 |
fenn | heh that's fine with me :P | 15:51 |
fenn | wait is this software or electronics | 15:52 |
nmz787_i | software | 15:52 |
-!- xentrac [~kragen@panacea.canonical.org] has quit [Ping timeout: 240 seconds] | 15:52 | |
kanzure | fenn: why would you be more interested in that instead of doing your robocompany | 15:52 |
fenn | what does cadence have to do with it then | 15:52 |
-!- xentrac [~kragen@panacea.canonical.org] has joined ##hplusroadmap | 15:52 | |
nmz787_i | it needs scripted with Python | 15:52 |
fenn | define "it" | 15:52 |
nmz787_i | cadence :P | 15:52 |
nmz787_i | i really don't know | 15:52 |
QuantumG | the other guy says he'll do it remote, but nuts to Oregon. | 15:53 |
nmz787_i | i can ask for more info, but its some automagic generation via scripting thing | 15:53 |
nmz787_i | but I know they want Cadence Allegro and Python experience | 15:54 |
fenn | i wonder if wiping fabric softener on my screen will keep dust off | 15:54 |
nmz787_i | kanzure: the job market is weird around here for tech, as the economy crash thing cut a lot of highly skilled workers from salaried positions, and temp agencies are their way of getting something done and evaluating you for potential salaried position if they like you and have money for maintenance or other longer term work | 15:56 |
juri_ | cadence. ick. | 15:57 |
-!- augur [~augur@129-2-129-33.wireless.umd.edu] has joined ##hplusroadmap | 15:58 | |
-!- augur [~augur@129-2-129-33.wireless.umd.edu] has quit [Remote host closed the connection] | 15:59 | |
fenn | juri_: please elaborate | 15:59 |
juri_ | fenn: whats to say? non-free software is non-free, and there is an intire industry in users that do not know how to perform jobs, using the non-free tools to perform said jobs. | 16:00 |
nmz787_i | juri_: do you use an intel processor? | 16:01 |
fenn | oh, i agree completely | 16:01 |
juri_ | i maintained cadence for a college of engineering for 3 years. during that time, i learned enough EE to need to route my own boards. i soon discovered the combined might of the professors of the entire college could not route a PCB to save their lives. | 16:01 |
juri_ | nmz787_i: depends on the task. | 16:01 |
fenn | "i designed the block diagram, layout is left as an exercise for the student" (literally) | 16:02 |
fenn | my microwave runs an intel Z80 | 16:02 |
kanzure | why does intel have a z80? | 16:02 |
fenn | uh, nevermind | 16:03 |
kanzure | "The Z80 came about when Federico Faggin, after working on the 8080, left Intel at the end of 1974 to found Zilog" | 16:03 |
fenn | i guess the z80 was compatible in many ways with the intel 8080 | 16:04 |
kanzure | http://en.wikipedia.org/wiki/Zilog_Z80 | 16:04 |
dingo | they used to have Z80 add-on cards for the intel | 16:05 |
dingo | so you could boot cp/m | 16:05 |
kanzure | http://upload.wikimedia.org/wikipedia/commons/d/d1/Z80A-HD.jpg | 16:06 |
fenn | so has anyone made a microchip CPU in their garage yet | 16:08 |
-!- Viper168 [~Viper@unaffiliated/viper168] has quit [Quit: Leaving] | 16:08 | |
kanzure | i believe azonenberg has in #homecmos http://code.google.com/p/homecmos/ | 16:08 |
fenn | wow really? | 16:08 |
kanzure | but he may have been cheating because he definitely stil has access to university lab stuff | 16:08 |
kanzure | *still | 16:08 |
nmz787_i | i don't think so | 16:09 |
nmz787_i | he's probably made a proc on a fpga for sure | 16:09 |
kanzure | i would be extremely surprised if he hasn't done a transistor | 16:09 |
nmz787_i | but that's more design than fab | 16:09 |
fenn | anyone can download an fpga core | 16:09 |
kanzure | nah, fpgas are standard curriculum these days, not interesting | 16:09 |
kanzure | i mean, no points | 16:09 |
nmz787_i | its a common course excercise to design the proc yourself, so I think he's probably done more than just download some existing one | 16:10 |
kanzure | i'm sure. but the ability to design a 4-bit chip is irrelevant (high school students can do that bored in math classes). | 16:10 |
kanzure | he was asking about manufacturing/etching | 16:10 |
kanzure | and i assume operating | 16:10 |
fenn | i wonder if it's possible to turn off prerendering in chrome so things dont use so much ram | 16:11 |
kanzure | fenn: have fun http://code.google.com/p/chromium/wiki/AndroidBuildInstructions | 16:12 |
fenn | i meant my laptop | 16:12 |
nmz787_i | he did some fpga circuit editing recently | 16:12 |
nmz787_i | but that was with a FIB in school | 16:12 |
fenn | at least i guess it's due to prerendering, i can't imagine why else a wikipedia page would need 100MB of ram | 16:14 |
kanzure | closing the last chrome tab may not kill the original chrome process | 16:14 |
juri_ | fenn: we're working on diamond like carbon deposition here at hacDC. | 16:15 |
fenn | juri_: why? | 16:15 |
fenn | kanzure: i'm killing tabs in the chrome task manager | 16:15 |
juri_ | potential integrated circuit applications, plus 3d printer hotend applications. | 16:15 |
fenn | what are those applications | 16:16 |
juri_ | (i'm crazy. i'm trying to print aluminium. mirage is a different type of crazy, and wants to print chips. | 16:16 |
fenn | (i've only ever seen CVD used as a wear resistant coating on machine tool cutters) | 16:16 |
fenn | okay so why diamond and not, i dunno, steel | 16:17 |
juri_ | steel is conductive, so won't work on my heater core. | 16:17 |
fenn | stainless steel lasts quite a while in oxidizing environments at 500C | 16:17 |
fenn | diamond is also conductive | 16:17 |
juri_ | we're going to use steel underneath the diamond, to handle molten aluminium. | 16:17 |
juri_ | DLC is not conductive of electricity, is conductive of heat. exactly what we need. | 16:18 |
fenn | investigating | 16:18 |
juri_ | this makes good sense, since its used under some chips nowadays. if it was electricly conductive, they would have problems. ;) | 16:19 |
fenn | hm. from http://en.wikipedia.org/wiki/Diamond#Electrical_conductivity "Substantial conductivity is commonly observed in nominally undoped diamond grown by chemical vapor deposition. This conductivity is associated with hydrogen-related species adsorbed at the surface, and it can be removed by annealing or other surface treatments." | 16:22 |
fenn | best of both worlds | 16:22 |
juri_ | selectively conductive. | 16:22 |
juri_ | ;D | 16:22 |
fenn | apparently they can be semiconductors too | 16:22 |
juri_ | we're refurbishing a SEM at the space, as well. | 16:23 |
juri_ | me and mirage are excited, as this will let us verify our work. | 16:23 |
fenn | cool | 16:23 |
fenn | how often is the space open these days? is it likely it would be open if i dropped by randomly? | 16:24 |
juri_ | its not quite that often, but its getting to about 50% chance. | 16:24 |
juri_ | that said, if you want to drop by, i'm willing to open it up for a tour, or to get some work done. just give me the word. | 16:25 |
-!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 16:31 | |
nmz787_i | kanzure: what was the bitcoin donation thing? | 16:31 |
nmz787_i | the easy one, to start a donation campaign thing | 16:31 |
kanzure | uh, might have been anything, coinbas.come? blockchain.info? | 16:32 |
kanzure | *coinbase.com? | 16:32 |
-!- eudoxia [~eudoxia@r186-48-167-145.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap | 16:32 | |
nmz787_i | is there one with a counter though, like "# bitcoin raised" | 16:34 |
-!- Viper168_ [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 16:34 | |
kanzure | nmz787_i: http://bitcoinchipin.com/ | 16:34 |
fenn | juri_: have you considered something like the wax inkjet but for aluminum? spraying microdroplets that fuse on contact with the workpiece | 16:35 |
nmz787_i | kanzure: thanks | 16:35 |
-!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 252 seconds] | 16:36 | |
fenn | i'm still not convinced diamond is necessary for this application. surely there are other things that don't dissolve in aluminum, have high temperature stability and oxidation resistance | 16:36 |
nmz787_i | alumina? | 16:38 |
nmz787_i | i guess that's already oxidized | 16:39 |
fenn | i wasn't sure if alumina would stand up to molten aluminum for extended periods | 16:39 |
fenn | aluminum is highly reactive | 16:39 |
fenn | this means alumina is very stable, but in a fight with aluminum, they come out even | 16:40 |
QuantumG | do it in a noble gas? | 16:40 |
* fenn looks at a periodic table | 16:41 | |
fenn | QuantumG: also a good option, but it complicates logistics | 16:41 |
QuantumG | or semi-noble.. aka nitrogen | 16:41 |
QuantumG | heck, even co2 would probably do, if you're just trying to avoid oxidization | 16:41 |
fenn | yes you could use getters and molecular sieve pumps | 16:42 |
nmz787_i | none of those damn peasant gasses | 16:42 |
QuantumG | for sho | 16:42 |
fenn | pressure swing oxygen concentrators i mean | 16:42 |
fenn | http://en.wikipedia.org/wiki/Reactivity_series MgO, CaO ceramics are probably good options (why is Si not on this chart?) | 16:46 |
-!- augur [~augur@129-2-129-33.wireless.umd.edu] has joined ##hplusroadmap | 16:51 | |
juri_ | fenn: honestly, i haven't. i figgure go for the brass ring. | 16:52 |
juri_ | microwave induced carbon vapour deposition looks like a 'simple enough' procedure. | 16:53 |
juri_ | plus, it gets us on our way to using the same procedure for integrated circuit applications. | 16:53 |
juri_ | two birds. one (admitedly heavy) stone. | 16:54 |
nmz787_i | kanzure: I can't tell if its working https://bitcoinchipin.com/widgets/u/takeitapartdan/send-takeitapart-to-maker-fair | 16:54 |
-!- nmz787_i [~nmccorkx@134.134.137.75] has quit [Quit: Leaving.] | 16:55 | |
-!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 16:56 | |
-!- Viper168_ [~Viper@unaffiliated/viper168] has quit [Ping timeout: 276 seconds] | 17:10 | |
fenn | i meant for aluminum printing, not diamond | 17:11 |
QuantumG | so did I | 17:11 |
-!- kuldeepdhaka [~kuldeepdh@unaffiliated/kuldeepdhaka] has quit [Quit: Leaving] | 17:12 | |
fenn | i just saw 10 minutes of "silicon valley" - probably the most depressing piece of distilled reality i've seen in a while | 17:12 |
QuantumG | I've seen ever episode.. people complaining that it isn't realistic are missing the purpose of FICTION | 17:13 |
-!- marciogm [~marciogm@179.126.61.80] has joined ##hplusroadmap | 17:13 | |
fenn | what's not realistic about it? | 17:14 |
QuantumG | they seem to think Silicon Valley is "weirder than that" | 17:14 |
jrayhawk | "Perhaps the idea you mean is that the Web should be a distributed programming environment rather than just a way to deliver documents. True, but it's sort of too late to fix that. It's like expecting MS-DOS to magically morph into Unix. If you built a global namespace from scratch you would not reuse names (no mutable resources), and the namespace would be rooted in a secure identity model. Refactoring the Web to include these ... | 17:14 |
jrayhawk | ... points is sort of like redesigning a horse so that it's a motorcycle." | 17:15 |
fenn | maybe they are confusing "silicon valley" with "san francisco" | 17:15 |
jrayhawk | huh, a comment from 2007 actually sells me on | 17:15 |
jrayhawk | Urbit | 17:15 |
kanzure | i was staying in downtown sf a few weeks ago | 17:15 |
kanzure | and i forgot about the screams of the undead during the night | 17:15 |
dingo | SJ claims to be *the* silicon valley, which is pretty silly | 17:15 |
fenn | jrayhawk: they web has a global namespace though | 17:15 |
kanzure | very weird place. | 17:15 |
kanzure | san jose is boring | 17:16 |
QuantumG | ya | 17:16 |
dingo | san francisco smells like garbage | 17:16 |
QuantumG | Palo Alto baby | 17:16 |
dingo | i don't like either, i'm a country boy myself | 17:16 |
kanzure | that's because of the garbage | 17:16 |
jrayhawk | No, DNS is not a part of the web. | 17:16 |
dingo | vim | 17:16 |
fenn | google hq is literally built next to a landfill.. go figure | 17:16 |
fenn | jrayhawk: what? | 17:17 |
jrayhawk | And neither is IP. | 17:17 |
fenn | what? | 17:17 |
jrayhawk | HTTP is a socket. | 17:17 |
fenn | define "the web" plz | 17:17 |
jrayhawk | http://en.wikipedia.org/wiki/World_Wide_Web | 17:18 |
QuantumG | I enjoyed the 8 months I spent at VMware, but I don't nitpick every little thing in a tv show that's different to my experience. | 17:18 |
kanzure | when was your stay at vmware? | 17:18 |
dingo | i don't know where this conversation is headed but i'm pulling out :-) | 17:18 |
fenn | hm. define:www and the first result is amazon.com - long live the web! | 17:18 |
QuantumG | 'specially seeing as it was 13 years ago. | 17:18 |
kanzure | was that before or after vmware started making crap like vcloudsphere | 17:18 |
kanzure | or vcow | 17:19 |
QuantumG | before.. but the writing was on the wall | 17:19 |
kanzure | what went wrong? | 17:19 |
fenn | they became a large organization | 17:20 |
QuantumG | I don't know how they ever made money. | 17:20 |
QuantumG | lots of sales dudes hitting the phones at the end of financial year, it seemed | 17:20 |
kanzure | weren't they venture backed and going for the "long haul" of no sales? | 17:20 |
QuantumG | nope | 17:20 |
QuantumG | they were desperately trying to find a market beyond "I wanna run Windows in a VM on my linux box" from day one. | 17:21 |
QuantumG | including selling hardware at one point.. | 17:22 |
kanzure | how about, "lots of big enterprises have terrible software running on terrible old hardware that could instead be virtualized because ???" | 17:22 |
kanzure | i forget the rest | 17:23 |
juri_ | quantumg: i have found there is a class of people that is bought into the "free means its cheap" mindset. maybe that's how they make money. | 17:23 |
QuantumG | yeah, I think they thought x86 (and later x64) was the hardest thing to virtualize so they had to focus on that to keep competitors out.. or something | 17:24 |
juri_ | i'm slowly converting one of those who spends WAYTOOMUCH to run free software fast, simply by setting up light debian instances, and trouncing his huge boxes with machines thrown away by others. | 17:25 |
QuantumG | plus, they didn't want to confuse people by explaining that virtualization is essentially dynamic recompilation | 17:25 |
-!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Quit: Leaving] | 17:26 | |
-!- eudoxia [~eudoxia@r186-48-167-145.dialup.adsl.anteldata.net.uy] has quit [Read error: Connection reset by peer] | 17:28 | |
fenn | why is it recompilation if it's compiled for the same instruction set as your host hardware? | 17:32 |
fenn | am i missing something | 17:33 |
kanzure | maybe it's using the virtualization extensions on chip | 17:34 |
-!- delinquentme [~dingo@74.61.157.78] has quit [Ping timeout: 240 seconds] | 17:34 | |
kanzure | or it's not using the same instruction set | 17:34 |
fenn | but windows and linux use the same instruction set, no? | 17:34 |
QuantumG | this was pre-chip | 17:34 |
QuantumG | most instructions you can just copy, some have to be "virtualized" | 17:35 |
fenn | i want a chrome extension that colors my tabs by CPU usage | 17:35 |
-!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 17:35 | |
-!- nmz787_i [nmccorkx@nat/intel/x-tbpvognrguxsiszg] has joined ##hplusroadmap | 17:37 | |
fenn | juri_: how about titanium boride instead of diamond: http://en.wikipedia.org/wiki/Aluminium_carbide#Occurrence | 17:37 |
QuantumG | those x86/x64 virtualization instructions did basically prove how stupid it was for VMware to focus on those architectures because they're hard to virtualize. | 17:38 |
fenn | TiB2 is resistant to oxidation in air at temperatures up to 1100 °C,[2] and to hydrochloric and hydrofluoric acids, but reacts with alkalis, nitric acid and sulfuric acid | 17:39 |
fenn | the melting point is about 2970 °C, and, thanks to a layer of titanium dioxide that forms on the surface of the particles of a powder, it is very resistant to sintering. Admixture of about 10% silicon nitride facilitates the sintering, | 17:40 |
fenn | electroplating of TiB2 layers possess two main advantages compared with physical vapor deposition or chemical vapor deposition: the growing rate of the layer is 200 times higher (up to 5 μm/s) and [you can plate inside small holes, important for your application] | 17:41 |
fenn | so you'd want to make the bulk of your nozzle out of graphite and electroplate TiB2 onto that | 17:42 |
fenn | or maybe even silicon would work | 17:42 |
fenn | then you could just use a pre-made inkjet nozzle | 17:43 |
-!- HashNuke [uid12117@gateway/web/irccloud.com/x-egcplrxinrchzrjl] has joined ##hplusroadmap | 17:43 | |
juri_ | i like the way you're thinking; now to source the materials. | 17:43 |
fenn | "Solution phase reaction of NaBH4 and TiCl4" <- gnarly chemicals | 17:46 |
-!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 17:46 | |
fenn | paperbot: http://www.sciencedirect.com/science/article/pii/S0167577X06008081 | 17:47 |
fenn | .title | 17:47 |
yoleaux | Preparation of the TiB2 coatings by electroplating in molten salts | 17:47 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/700d72474d9db8c0016026c3e8e922a1.txt | 17:47 |
kanzure | paperbot: http://www.sciencedirect.com/science/article/pii/S0167577X06008081/pdfft?md5=29ca21048e348f649287e28ea51b2d0d&pid=1-s2.0-S0167577X06008081-main.pdf | 17:48 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/4bd972db526fcaca0c537123e04ffb78.pdf | 17:48 |
fenn | .title | 17:48 |
yoleaux | fenn: Sorry, that command (.title) crashed. | 17:48 |
kanzure | doesn't work on pdfs | 17:48 |
fenn | if it were a pdf you wouldn't have to use paperbot | 17:49 |
fenn | .title http://www.sciencedirect.com/science/article/pii/S0167577X06008081/pdfft?md5=29ca21048e348f649287e28ea51b2d0d&pid=1-s2.0-S0167577X06008081-main.pdf | 17:49 |
yoleaux | Preparation of the TiB2 coatings by electroplating in molten salts | 17:49 |
fenn | huh | 17:49 |
kanzure | the reason .title works there is because yoleaux doesn't have access | 17:49 |
kanzure | so sciencedirect.com redirects the bot to an html page | 17:49 |
fenn | i saw a robot that used a mig welder to 3d print stuff. you can get amazingly high aspect ratio shapes because of the stiffness and lack of mechanical force | 17:51 |
QuantumG | yeah, automated mig wire art | 17:51 |
kanzure | i want a wire keyboard. | 17:52 |
-!- Auctwo [~Auctus@122-57-138-207.jetstream.xtra.co.nz] has joined ##hplusroadmap | 17:53 | |
fenn | the mig process uses a thin stream of "shield gas" that surrounds the electrode; this should probably be used around the stream of aluminum, even though it is only going a few mm | 17:56 |
fenn | if only to prevent oxide buildup around the electrode and eventual inclusion in the part | 17:56 |
fenn | er, s/electrode/nozzle/ | 17:56 |
-!- Auctus [~Auctus@unaffiliated/auctus] has quit [Ping timeout: 276 seconds] | 17:57 | |
fenn | may need some kind of oxide skimmer because your feed material will have an oxide coating before it gets melted | 17:58 |
-!- AshleyWaffle_ [~waffle@gateway/tor-sasl/anastasiawyatt] has joined ##hplusroadmap | 17:59 | |
-!- AshleyWaffle [~waffle@gateway/tor-sasl/anastasiawyatt] has quit [Disconnected by services] | 18:00 | |
-!- AshleyWaffle_ is now known as AshleyWaffle | 18:00 | |
fenn | there's been a lot of noise in the past about syntactic(?) metal foams but no way to make them; a high resolution 3d metal printer could actually do it though | 18:02 |
fenn | thrash thrash thrash | 18:02 |
fenn | inexpensive heatsinks with built in heat pipes would lead to some interesting possibilities | 18:04 |
kanzure | machinewiki should be dumped into diyhpluswiki.git | 18:05 |
fenn | what is machinewiki | 18:05 |
fenn | the gingery things? | 18:05 |
kanzure | oops machines | 18:05 |
kanzure | http://fennetic.net/machines/ | 18:05 |
kanzure | i have been on a wiki merging spree lately | 18:05 |
fenn | ya i was hoping to come up with something better than wikis | 18:05 |
fenn | but that was a long time ago and it hasn't happened | 18:05 |
kanzure | los angeles biohackers lost their mediawiki install | 18:05 |
kanzure | cory emailed me an sql dump today | 18:06 |
kanzure | most of these wikis are going to evaporate | 18:06 |
kanzure | isn't tmp2 a wiki? | 18:06 |
fenn | every wiki should have a "dump database" button imho | 18:06 |
kanzure | mediawiki has a kinda-dump button, but you have to stare at it sternly | 18:06 |
QuantumG | reminds me of what I asked you the other day kanzure.. | 18:06 |
fenn | an end-user-accessible button | 18:06 |
QuantumG | got a list of biotechnology machines? | 18:07 |
fenn | of course you also need a merge button, thus ikiwiki | 18:07 |
kanzure | i gave you one | 18:07 |
kanzure | http://diyhpl.us/cgit/skdb/tree/doc/BOMs/diybio-equipment.yaml | 18:07 |
fenn | QuantumG: cows, pigs, sheep, corn, e. coli | 18:07 |
kanzure | hah i forgot about this http://diyhpl.us/cgit/skdb/tree/doc/tech-tree | 18:07 |
fenn | like the list? i made it myself | 18:07 |
kanzure | "I spent over a day writing a silly Python program to read in a Civilization 2 Technology Tree." | 18:08 |
kanzure | uggghhh | 18:08 |
QuantumG | have you got one that isn't ridiculous? | 18:08 |
kanzure | "The Civilization 2 Technology Tree has five errors, including two “Destroyer” units and a bunch of redundant dependencies, such as Fusion Power doesn’t need to depend on Nuclear Power." | 18:08 |
kanzure | why is it ridiculous? | 18:08 |
kanzure | http://www.charlesmerriam.com/blog/2008/04/fun-with-programming-a-technology-tree/ | 18:09 |
QuantumG | I'm after a list of commercial lab machinery | 18:09 |
kanzure | oh you mean the equipment they wont sell you? | 18:09 |
QuantumG | ya | 18:09 |
QuantumG | what they do, what they're for, etc | 18:09 |
-!- Darius [~quassel@adsl-69-231-35-194.dsl.irvnca.pacbell.net] has joined ##hplusroadmap | 18:09 | |
kanzure | ABI {100..900}{A..Z} | 18:09 |
@ParahSailin | you can build an ironclad before iron working | 18:09 |
@ParahSailin | and gunpowder | 18:09 |
fenn | iirc openwetware had a list of their preferred commercial bio lab equipment | 18:09 |
@ParahSailin | there are two destroyers? | 18:10 |
kanzure | drew endy approved equipment? | 18:10 |
kanzure | oh yeah, a backup of openwetware too.. god damn wikis. | 18:10 |
fenn | one wiki to rule them all, one wiki to bind them | 18:10 |
kanzure | i dunno if i want to dump openwetware straight into diyhpluswiki.git | 18:10 |
kanzure | it's not like anyone other than me edits diyhpluswiki.git | 18:10 |
@ParahSailin | oh, the second destroyer "steel" is supposed to be cruiser | 18:11 |
@ParahSailin | electricity is destroyer | 18:11 |
fenn | pierre baldi ChemDB, about 5 million chemical isomers in the database are available for download, including all of their primary chemical annotations. They are available for download as a collection of gzipped SDF molecular format files, each about 100 MB large. http://cdb.ics.uci.edu/ | 18:14 |
-!- Darius [~quassel@adsl-69-231-35-194.dsl.irvnca.pacbell.net] has quit [Ping timeout: 265 seconds] | 18:15 | |
gnusha | https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=9ea01bce Bryan Bishop: list of wikis to merge into this wiki | 18:15 |
kanzure | http://diyhpl.us/wiki/wikis/ | 18:16 |
fenn | isn't openwetware still an active wiki (it's like their primary public-facing website) | 18:17 |
kanzure | i believe the only reason it's active is because igem teams are coerced into posting their projects | 18:18 |
fenn | i mean why would you even consider copying it | 18:18 |
kanzure | i merged their diybio pages into diyhpluswiki.git a while back | 18:18 |
kanzure | so maybe some of the other content is wort htaking | 18:18 |
kanzure | *worth taking | 18:18 |
fenn | signal to noise ratio is important | 18:18 |
kanzure | i said some :) | 18:18 |
fenn | there are so many bio people i've never heard of | 18:20 |
kanzure | that's why i started writing them down. who the hell are these people? | 18:20 |
fenn | maybe it's just so hard to get anything done with biology that nobody makes a lasting impression | 18:21 |
fenn | the gap between brilliant idea and practice is too large | 18:21 |
kanzure | the parts of biology that require secret incantations should be thrown out in favor of anything that works well with the words "solid state" | 18:21 |
fenn | are microfluidics solid state? | 18:22 |
fenn | yes i know you also said "works well" | 18:22 |
kanzure | what's wrong with regular fluidics | 18:22 |
kanzure | do you really always need laminar flow? | 18:23 |
fenn | regular fluidics is big | 18:23 |
fenn | there is certainly a case to be made for mesofluidics | 18:23 |
kanzure | bigness hasn't been one of the constraints as far as i know? | 18:23 |
kanzure | erm, size | 18:23 |
kanzure | "you know, all this biology stuff is great, but what i really need is a place to put my shoes" | 18:24 |
fenn | someone was coming up with a chemistry reactor compiler (3d print a silicone thing to synthesize a batch of chemicals) | 18:24 |
fenn | kanzure: size of lab equipment determines where you can and can't put a lab | 18:24 |
fenn | for example, can you have a lab in your backpack? in your closet? in your guest room? or do you need to rent a storage unit or workspace | 18:25 |
fenn | this has obvious implications for who can and can't do lab work | 18:25 |
fenn | if anyone with a backpack can contribute to your research then your pool of potential contributors goes up by orders of magnitude | 18:26 |
kanzure | those implications don't seem to be a limiting factor at the moment | 18:26 |
kanzure | there seem to be other limiting factors that are extering more influence | 18:26 |
fenn | like what | 18:26 |
fenn | (i honestly dont know) | 18:26 |
fenn | why are people contributing to wikipedia but not diybio | 18:26 |
kanzure | i was going to complain about everyone who built the original equipment being dead after swallowed up by big bio in ancient history | 18:27 |
kanzure | but, the thesis that there are limiting factors that are preventing biology things, might also be flawed | 18:27 |
QuantumG | biometh labs | 18:28 |
kanzure | what about them | 18:28 |
fenn | also consider workflow, if your stuff is all on one desk then you dont have to run laps around a huge lab just to get anything done | 18:29 |
QuantumG | just a visual | 18:29 |
QuantumG | if your lab is (smoothly) automated you don't have to get up from your desk | 18:29 |
kanzure | fenn: the reason i decided that microfluidics was too much work was because the number of variables increases so dramatically compared to non-microfluidics | 18:29 |
fenn | QuantumG: also if you pay someone else to do it | 18:29 |
kanzure | fenn: engineering is about working on known constraints and within known tolerances, not adding arbitrary parts to your project until you hope it works | 18:30 |
QuantumG | that too | 18:30 |
kanzure | the amount of debugging necessary per additional buggy component that you add into your system makes "make a 1000-gate microfluidic device for a 12 step chemical reaction" almost intractable | 18:30 |
kanzure | if possible i prefer to focus on tractable things | 18:31 |
QuantumG | people are pretty buggy too | 18:31 |
kanzure | i wasn't using that as an argument against lab automation, sigh | 18:31 |
fenn | its the same amount of debugging, it's just harder to inspect | 18:32 |
kanzure | i was using that as an argument for why the microfluidic dna synthesizer is a huge project | 18:32 |
QuantumG | total quality management is like inevitable, man | 18:32 |
fenn | microfluidics _enables_ massive combinatorial explosions, but they're not required | 18:32 |
kanzure | why is adding arbitrary amounts of unknowns into a project good? | 18:32 |
QuantumG | usually it's the benefits you can't get from avoiding them.. like using C++ | 18:33 |
kanzure | QuantumG: context is http://diyhpl.us/~bryan/nucleic/fbi-diybio-dna-v1.pdf | 18:33 |
fenn | ok i see what this argument is about. option 1) replicate known working devices from the literature. option 2) build new devices based on principles learned from the literature | 18:33 |
kanzure | 1) those don't reaaaally exist | 18:33 |
QuantumG | whenever someone asks me why we use C++ I just cut to the chase and say "because we're idiots", it's what they want to hear. | 18:33 |
fenn | right because those people are all long dead | 18:34 |
fenn | and they didn't write down their secrets | 18:34 |
kanzure | no, they aren't dead | 18:34 |
kanzure | ParahSailin spent a bunch of time on microfluidics stuff | 18:34 |
fenn | so we're actually stuck with option 2 either way | 18:34 |
fenn | what does microfluidics have to do with anything | 18:34 |
kanzure | you were trying to convince me to focus on small devices | 18:34 |
kanzure | so i skipped ahead in the conversation | 18:34 |
fenn | i thought you wanted to make a 1980s DNA synthesizer with the big glass bottles hanging off the front | 18:34 |
@ParahSailin | i think cambrian folks are going to have the most sort of success doing dna synth | 18:34 |
kanzure | that doesn't really help anyone but their customers | 18:35 |
-!- Viper168_ [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 18:35 | |
kanzure | i mean, they are explicitly mail order only iirc | 18:35 |
fenn | QuantumG: C++ wasn't anti-hacker enough so they invented java? | 18:36 |
-!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 252 seconds] | 18:36 | |
fenn | question, is cambrian "solid state"? | 18:37 |
kanzure | well they are certainly using a laser somewhere in their pipeline | 18:37 |
fenn | there's also lots of shit literally flying around | 18:37 |
fenn | how many other machines have literally millions of moving parts | 18:38 |
fenn | yet still somehow it's more reliable than microfluidics? | 18:38 |
fenn | what am i missing here | 18:38 |
-!- augur [~augur@129-2-129-33.wireless.umd.edu] has quit [Remote host closed the connection] | 18:40 | |
fenn | maybe it's the difference between uncorrected error and feedback | 18:40 |
QuantumG | so, stupid question, ya never got anywhere on this project? | 18:40 |
kanzure | it's interesting that you and i have a different perspective on picking (in)tractable projects | 18:40 |
kanzure | QuantumG: nmz787 is still going at it, but i backed out when i realized the existential horror of it all | 18:41 |
fenn | i think you are over-exaggerating | 18:41 |
kanzure | well, then let's re-calibrate me (in general) | 18:41 |
fenn | and i think nmz787_i likes to fiddle with things | 18:41 |
-!- marciogm [~marciogm@179.126.61.80] has quit [Ping timeout: 240 seconds] | 18:41 | |
QuantumG | any interest left in the "enzymatic synthesis by direct control of DNA polymerase" version? | 18:42 |
kanzure | as you can increase the number of parts in a system you increase (at minimum) the number of possible failures | 18:42 |
kanzure | QuantumG: sure, http://diyhpl.us/~bryan/papers2/polymerase/ has some research notes | 18:42 |
fenn | QuantumG: it's interesting but theoretically stalled | 18:42 |
kanzure | yes | 18:42 |
kanzure | or, theoretically waiting on more theory stuff | 18:43 |
kanzure | "Protein conformational dynamics probed by single-molecule electron transfer.pdf" | 18:43 |
fenn | something like the inverse of the pacific bio synthesizer should be possible | 18:43 |
kanzure | "review - Making contact - Connecting molecules electrically to the macroscopic world.pdf" | 18:43 |
fenn | i mean nano volumes saturated by various frequencies of light that turn on a specific reaction | 18:44 |
kanzure | i wonder if i should estimate the engineering difficulty of that polymerase enzyme as either more or less than the potential bug/iteration count of a large microfluidic device | 18:44 |
@ParahSailin | protein engineering is awful | 18:45 |
-!- marciogm [~marciogm@179.126.69.68] has joined ##hplusroadmap | 18:45 | |
kanzure | the fake polymerase enzyme is much more fantasy so probably can't be estimated at all (the solution will probably end up being "directed evolution, and then magic happens, and oh look you got it on the 3rd month") | 18:45 |
fenn | the difference is that the microfluidic device problem can be decomposed into different working parts | 18:45 |
kanzure | same with proteins actually | 18:45 |
kanzure | ... to some extent. not always in the way that you need. | 18:46 |
fenn | i disagree | 18:46 |
kanzure | there are literally specific domains in the enzyme itself | 18:46 |
fenn | it's true you can combine functional subunits with linkers | 18:46 |
kanzure | not even subunits | 18:46 |
-!- Vutral [~ss@mirbsd/special/Vutral] has quit [Ping timeout: 246 seconds] | 18:46 | |
fenn | but when all your functional subunits work on the same active site, they have to be at least designed with an awareness of each other | 18:46 |
@ParahSailin | you might be able to get a polymerase that would halt if you changed its conformation by tugging on it with a magnetic bead | 18:46 |
kanzure | subunits as in other proteins (like from other genes) that form together to build the overall thing, right? | 18:46 |
@ParahSailin | but that would be a hell of a protein engineering project | 18:46 |
QuantumG | I completely agree that it's a different problem | 18:46 |
kanzure | dna polymerase often doesn't have subunits, but it does have functionally-specific components | 18:46 |
fenn | by subunit i was thinking "here's the part that adds 'A', here's the part that adds 'G' etc" | 18:47 |
kanzure | quick model of a polymerase is a hand with the thumb feeding into the inner chamber | 18:47 |
QuantumG | the enzyme-version is a science project.. the microfluidics-version is an engineering problem (but a pretty science project-y one), and the traditional-chemistry-with-some-modern-automation is a straight engineering problem | 18:47 |
kanzure | and then some weirdo chute where nucleotides get matched up | 18:48 |
QuantumG | the payoffs are also different though. | 18:48 |
@ParahSailin | most dna polymerases except for phage ones have subunits | 18:48 |
fenn | if separate polymerases have to fall off and reconnect to the strand for each base, the kinetics are too slow | 18:48 |
kanzure | ParahSailin: clearly i have to go kill myself now.. | 18:48 |
@ParahSailin | there are a lot of viral polymerases though | 18:48 |
kanzure | QuantumG: i'm curious about how you would characterize the straightforwardness in those last two.. | 18:48 |
kanzure | the good thing about polymerase is that it's one of the really good molecules for directed evolution | 18:49 |
kanzure | but then you need to figure out if your problem is evolutionarily-solvable | 18:49 |
QuantumG | well, no-one would argue that the last project is doable.. whereas the microfluidics project has unknowns | 18:49 |
fenn | microfluidics you're inventing both the process and the device; with the bulk fluidic synthesizer you're just inventing the device (process is known) | 18:49 |
kanzure | disagree, but yes obviously microfluidics has more work | 18:50 |
fenn | somehow making it smaller changes the process (don't ask me) | 18:50 |
kanzure | my original statement was that the amount of debugging and iterations for the microfluidic version would be way higher | 18:50 |
kanzure | it's not just making it smaller | 18:50 |
QuantumG | maybe I'm not understanding the last project though.. automate-a-manual-process is how I saw it, no? | 18:51 |
kanzure | it's different materials, chemistries, flow, valves, etc. | 18:51 |
fenn | why does it have to be different materials and chemistries | 18:51 |
QuantumG | I wonder if you're not taking that literally enough | 18:51 |
kanzure | QuantumG: yes, lots of that equipment exists. basically think stereotypical 1970s chemistry process, "hey let's sell a machine", "okay let's make something that looks like our glassware setup".. | 18:51 |
fenn | we can make glass chips; the chemistry scales down (no?) | 18:51 |
QuantumG | I mean, you *can* make reasonably long DNA sequences with glassware, right? | 18:52 |
fenn | QuantumG: yes | 18:52 |
kanzure | manual dna synthesis definitely works, if you have iron patience | 18:52 |
QuantumG | a Baxter is only $22k ;) | 18:52 |
kanzure | although the yields might be less than you'd hope | 18:52 |
kanzure | and this will ultimately constrain the maximum length | 18:52 |
QuantumG | sounds expensive, until you automate a few different processes with the same robot. | 18:52 |
kanzure | besides the usual types of errors: http://diyhpl.us/~bryan/papers2/DNA/Origin%20of%20impurities%20in%20oligonucleotides.pdf | 18:53 |
fenn | QuantumG: a Boxter is only $9k http://www.ebay.com/itm/Porsche-Boxster-Roadster-Convertible-2-Door-2000-porsche-boxter-/121322965246 | 18:54 |
QuantumG | yeah, synthesize your own DNA, I'm going cruisin' | 18:55 |
fenn | ok so what is the major difference between automating it with pumps and valves and stirrers, and automating it with micropumps and microvalves and micromixers | 18:57 |
@ParahSailin | its possible that the idea of trying to make a dna polymerase that you can control from the macro scale is like putting a beak on an aircraft so that it will fly like a bird | 18:58 |
-!- nmz787_i [nmccorkx@nat/intel/x-tbpvognrguxsiszg] has quit [Quit: Leaving.] | 18:58 | |
@ParahSailin | and that the optimal way to synthesize long dna will not resemble nature at all | 18:58 |
kanzure | ligation? | 18:58 |
-!- marciogm [~marciogm@179.126.69.68] has quit [Quit: Computer has gone to sleep.] | 18:58 | |
kanzure | oh wait, that's from nature | 18:58 |
fenn | ParahSailin: surely you mean flapping wings, not "beak" | 18:58 |
kanzure | yeah, i haven't seen much hypothesizing about extremely long dna synthesis | 18:59 |
kanzure | screw 1 gigabase genomes, what about 1 terabase etc | 18:59 |
@ParahSailin | fenn: im sure some people tried beaks when wings didnt work | 18:59 |
fenn | the truth is big birds fly more by catching thermals than by flapping their wings; we just didn't pay attention | 18:59 |
fenn | it's possible that da vinci built a hang glider and was able to catch thermals | 19:00 |
fenn | nobody really knows | 19:00 |
kanzure | obviously the correct way to synthesize extremely long dna molecules is through particle acceleration | 19:00 |
fenn | anyway, de novo DNA synthesis doesn't really happen in nature | 19:00 |
kanzure | GNA and other variants of DNA are an interesting concept too | 19:01 |
fenn | what's GNA | 19:01 |
kanzure | because you could possibly find something that is easier to do chemistry on, and then convert it to DNA | 19:01 |
kanzure | it's just a modified backbone, "simpler" chemistry or something | 19:01 |
fenn | eh who cares | 19:01 |
kanzure | and apparently it was polymerase-compatible | 19:01 |
kanzure | well, because if you get to pick your chemistry, then you have more options | 19:01 |
@ParahSailin | glycerol phosphate or something? | 19:01 |
kanzure | http://en.wikipedia.org/wiki/Glycol_nucleic_acid | 19:02 |
-!- marciogm [~marciogm@179.126.69.68] has joined ##hplusroadmap | 19:02 | |
kanzure | "The 2,3-dihydroxypropylnucleoside analogues were first prepared by Ueda et al. (1971). Soon thereafter it was shown that phosphate-linked oligomers of the analogues do in fact exhibit hypochromicity in the presence of RNA and DNA in solution (Seita et al. 1972). The preparation of the polymers was later described by Cook et al. (1995, 1999) and Acevedo and Andrews (1996). The GNA-GNA self-pairing described by Zhang and Meggers is however ... | 19:02 |
kanzure | ... novel, and the specificity of interaction well-demonstrated." | 19:02 |
kanzure | "DNA and RNA have a deoxyribose and ribose sugar backbone, respectively, whereas GNA's backbone is composed of repeating glycol units linked by phosphodiester bonds. The glycol unit has just three carbon atoms and still shows Watson-Crick base pairing. The Watson-Crick base pairing is much more stable in GNA than its natural counterparts DNA and RNA as it requires a high temperature to melt a duplex of GNA. It is possibly the simplest of the ... | 19:02 |
fenn | how about changing the surface chemistry of an array of microwells to accept various bases (conveniently lined up and waiting for ligation) | 19:02 |
kanzure | ... nucleic acids, so making it a hypothetical precursor to RNA." | 19:02 |
-!- augur [~augur@216-164-48-148.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined ##hplusroadmap | 19:02 | |
fenn | different charge patterns in the wells attract different bases | 19:02 |
kanzure | why? | 19:03 |
kanzure | what if instead you just have different nucleotides available in separate containers | 19:03 |
fenn | because it's parallel instead of serial | 19:03 |
kanzure | oh, well, single-nucleotide ligation doesn't happen | 19:03 |
fenn | why not | 19:03 |
kanzure | grabs longer sequences only iirc | 19:03 |
fenn | how long does it have to be | 19:03 |
@ParahSailin | i think cambrian is promising because phosphothioate is pretty effective chemistry that we have a good handle on | 19:04 |
@ParahSailin | and you can build fancy error checking and ligation on top of that | 19:04 |
fenn | "Oligonucleotides as short as 8 nucleotides can be efficiently assembled using T4 DNA Ligase" | 19:04 |
fenn | note that's in free solution, not bound to a plate | 19:05 |
fenn | i'd expect activity to be higher when bound to a plate (for enzymes that like the plate) because it's 2D reaction kinetics instead of 3D | 19:05 |
kanzure | i've seen lots of stuff about enzymes being bound to ATP n' stuff | 19:06 |
kanzure | and sometimes surfaces, like biotinylated | 19:06 |
fenn | if you can get it down to ~4 bp oligos, then it's cominatorially feasible to just dump in oligos instead of bases | 19:06 |
kanzure | "Generation of biotin/avidin/enzyme nanostructures with maskless photolithography" stuff like that | 19:06 |
QuantumG | damn wget, walking up your directory tree | 19:06 |
kanzure | wget -m -np | 19:06 |
kanzure | plz | 19:06 |
@ParahSailin | oh dang, i didnt know that one existed, i always was using -I | 19:07 |
-!- Darius [~quassel@adsl-69-231-35-194.dsl.irvnca.pacbell.net] has joined ##hplusroadmap | 19:08 | |
fenn | 4^8 = 65536, 4^7 = 16384, 4^6 = 4096, 4^5 = 1024, 4^4 = 256 so maybe 6 bp sequences is reasonably | 19:08 |
kanzure | oligomer libraries are a thing, you know | 19:08 |
-!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap | 19:08 | |
kanzure | but it doesn't require surface binding | 19:08 |
kanzure | paperbot: http://pubs.acs.org/doi/abs/10.1021/ac970957t | 19:09 |
kanzure | .title | 19:09 |
yoleaux | Microelectrode Control of Surface-Bound Enzymatic Activity | 19:09 |
kanzure | oh, hrm. | 19:09 |
kanzure | yeah these are all going to be electrode chemistry | 19:09 |
fenn | patent 5,650,489 june 1991 library of bio-oligomers of defined size and known composition, in which the library contains all of the possible sequences of the bio-oligomers, and a method of synthesis thereof. | 19:10 |
paperbot | XMLSyntaxError: None (file "/home/bryan/code/paperbot/phenny/modules/scihub.py", line 51, in _go) | 19:10 |
fenn | yeah this patent is basically what I was trying to describe | 19:11 |
kanzure | the problem with in-solution oligomer libraries is that you need to either have enough wings to avoid wrong additions, | 19:11 |
kanzure | or if you do it once per cycle then you need a bazillion wash reactions | 19:11 |
kanzure | and that's an imperial bazillion | 19:11 |
fenn | i guess they were doing random sequences though | 19:12 |
fenn | what's a "cycle" in this context | 19:12 |
kanzure | phosphoramidite/oligonucleotide synthesis involves a few steps that get repeated a bunch: wash, insert chemicals, wait for reaction, cap, wash, repeat | 19:13 |
@ParahSailin | why dont find efficient ways to error check and ligate phosophoramidite microarrays | 19:14 |
fenn | this method would be reduced to: insert oligos, wash, insert ligase, wash, repeat | 19:14 |
kanzure | so robotic pipetting of microarray? ala what george is doing? | 19:14 |
fenn | even better if your oligo solution has anti-ligase properties | 19:15 |
fenn | gotta keep the hot side hot and the cool side cool | 19:15 |
@ParahSailin | isnt he using beads instead of pipettes? | 19:15 |
QuantumG | phrasing | 19:17 |
-!- entelechios [~elysium@mail.2bett.com] has joined ##hplusroadmap | 19:17 | |
fenn | oh, so the moving bead is coated with the complementary oligo, and the stationary reaction site holds the strand you're working on? | 19:17 |
fenn | this seems slow | 19:17 |
fenn | moving beads around takes time | 19:17 |
kanzure | i think it was this one | 19:18 |
kanzure | http://diyhpl.us/~bryan/papers2/DNA/High-fidelity%20gene%20synthesis%20by%20retrieval%20of%20sequence-verified%20DNA%20identified%20using%20high-throughput%20pyrosequencing%20-%20supplementary.pdf | 19:18 |
fenn | are you talking about what i'm talking about? or is this unrelated? | 19:18 |
kanzure | i'm talking about ParahSailin's suggestion of micropipetting | 19:18 |
@ParahSailin | yeah thats it | 19:18 |
@ParahSailin | you could probably come up with faster ways to move beads around | 19:19 |
fenn | that's what cambrian's laser thingy is for | 19:19 |
kanzure | they laser at a microarray? | 19:19 |
fenn | basically | 19:19 |
@ParahSailin | they use two lasers | 19:19 |
kanzure | but where does the bead go? | 19:19 |
@ParahSailin | one for microarray, one for beads | 19:20 |
kanzure | so bead goes up into the air, another laser pushes it? | 19:20 |
kanzure | by... acoustic vibration? | 19:20 |
heath | hmm. | 19:21 |
heath | i think i'm going to switch back to python | 19:21 |
fenn | they grow oligos on a plate with DLP and photosynthesis, release the oligos into dilute solution, bind to beads such that statistically only one sequence per bead, do surface-bound PCR to amplify the oligos on the beads, then pop the beads off into a sequencer well | 19:21 |
@ParahSailin | beads are in water | 19:22 |
@ParahSailin | theres no sequencer well, its sequencing illumina style | 19:22 |
kanzure | i still don't know where the laser part is | 19:22 |
fenn | ok they have like 96 sequencers running in parallel then | 19:23 |
fenn | the laser is directly above the sequencer input | 19:23 |
fenn | the plate with the beads is upside down over the input, being moved around by an XY table | 19:23 |
fenn | i think both the laser and the plate are moving, so they can get the desired bead into the desired sequencer input | 19:24 |
fenn | either that or the sequencer moves, i haven't actually seen it | 19:24 |
@ParahSailin | theres no 96 sequencers, in illumina sequencing you have one synchronized reaction and one or multiple microscope cameras scanning a plate in a raster pattern | 19:25 |
-!- yashgaroth [~ffffff@cpe-76-167-105-53.san.res.rr.com] has joined ##hplusroadmap | 19:27 | |
fenn | " separate DNA strands from the microarray (biochips). They attach each strand to its very own bead, where they replicate it 100,000 times. Next, they attach the beads to a glass surface, and optically sequence them—one color for each of the four letters in a DNA sequence. In this way, they are able to read out the sequence, and pick those with the highest quality score. Cambrian Genomics then | 19:29 |
fenn | uses a laser printer to retrieve the DNA sequences. Using laser light, Heinz is able to quickly read one billion strands; find and print the correct ones;" i dont really understand that last sentence, maybe the journalist was improvising | 19:29 |
kanzure | "uses a laser printer to retrieve the DNA" | 19:30 |
fenn | there used to be a better explanation of this online, it's probably in the logs somewhere | 19:31 |
fenn | okay so the point of this is to do gene synthesis; eventually you need to get a few oligos in a reaction well to start doing your gibson assembly | 19:32 |
kanzure | the better explanation is anselm's second email over here https://groups.google.com/forum/#!searchin/diybio/cambrian$20genomics/diybio/YqGUf3A9AKo/NCFQ12T_8SwJ | 19:32 |
fenn | just pretend i said "reaction well" instead of "sequencer well" | 19:33 |
kanzure | well, nevermind, i don't think he points out the full process | 19:33 |
@ParahSailin | i wonder if you'd even have more luck using the poxvirus dna pol instead of gibson | 19:33 |
fenn | kanzure: hey why isnt this on the hplus wiki | 19:33 |
@ParahSailin | poxvirus pol only needs overlap of 15nt or so, and it doesnt chew things up willy nilly | 19:33 |
@ParahSailin | only problem is that you have to sacrifice a child infected with chickenpox to isolate this enzyme | 19:34 |
@ParahSailin | so its not commonly used | 19:34 |
fenn | i really doubt that's the reason | 19:34 |
@ParahSailin | heh | 19:35 |
fenn | fuck you too google *grump* | 19:35 |
@ParahSailin | it is pretty expensive, but only because it requires mammal cell culture | 19:35 |
kanzure | why does anselm call micropipetting "very high throughput"? | 19:36 |
@ParahSailin | clontech sells it for like $20/rxn which is not that expensive in the whole scheme of things | 19:36 |
kanzure | if you have a 1024x1024 microarray, the only way you're going to have high throughput is with a 1024x1024 micropipetting device | 19:36 |
fenn | ParahSailin: just use recombinant e. coli to make your enzyme. what's the problem | 19:36 |
@ParahSailin | fenn: nope they tried that | 19:36 |
@ParahSailin | viruses are evil, they never do what you want them to do | 19:37 |
kanzure | terrorize children? | 19:37 |
@ParahSailin | free north korea? | 19:37 |
fenn | ya gotta believe me | 19:37 |
kanzure | how fast can you drive a micropipette over a million wells | 19:38 |
fenn | kanzure: that's what the laser is for | 19:38 |
kanzure | stoping at each well to do the necessary | 19:38 |
fenn | how many times do i have to say it | 19:38 |
kanzure | *stopping | 19:38 |
kanzure | apparently you need to say it again | 19:38 |
@ParahSailin | paperbot: https://www.sciencedirect.com/science/article/pii/S0042682299997052 | 19:38 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/335a64ef75c98c46d28e301479744549.txt | 19:39 |
-!- ebowden [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has joined ##hplusroadmap | 19:39 | |
fenn | since apparently i need google chrome to look at a google groups page, i'm still waiting to read the diybio post | 19:39 |
kanzure | fenn: how does the laser speed up the pipetting? | 19:39 |
kanzure | oh wait, is it a game of 2048? | 19:40 |
-!- Adifex is now known as Adifex|pub | 19:41 | |
ebowden | paperbot: http://diyhpl.us/~bryan/papers2/paperbot/3a73de0d7d8f7aaac21d63537a91056a.txt | 19:41 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/99cb84d56d5df36e73c5ced10d36d30e.txt | 19:41 |
kanzure | what | 19:42 |
@ParahSailin | paperbot: http://diyhpl.us/~bryan/papers2/paperbot/99cb84d56d5df36e73c5ced10d36d30e.txt | 19:42 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/bbdc9be894f39be6b87612761b47809c.txt | 19:42 |
kanzure | paperbot: file:// | 19:43 |
kanzure | oh right, i anticipated thta one | 19:43 |
fenn | paperbot hacks into your system and liberates the paper | 19:43 |
kanzure | paperbot is the source of all knowledge (and some anti-knowledge) | 19:44 |
kanzure | anyhow: how does the laser speed up pipetting a million wells? | 19:44 |
QuantumG | Light controlled synthesis of nucleic acids - Pinheiro - 2010 .. looks like a winner to me | 19:44 |
kanzure | yeah they use photolithography and DMDs for the synthesis of oligos | 19:45 |
@ParahSailin | you dont have to move that many beads around to be orders of magnitude better than contemporary techniques | 19:45 |
fenn | kanzure: i just re-read all that and i'm honestly not understanding what you're missing | 19:46 |
@ParahSailin | a million beads to make all the yeast chromosomes, but most people would be happy with 100 beads | 19:46 |
kanzure | okay, so the pipette visits each of the 100 wells each step? | 19:46 |
fenn | 10^5 oligos -> 10^6 beads, sequence the beads (in place apparently), pop off the beads you want into the desired well for assembly | 19:47 |
fenn | there is no pipette anywhere | 19:47 |
@ParahSailin | there exists at least one pipette somewhere | 19:47 |
fenn | maybe there are pipettes once you get into the assembly stuff | 19:48 |
yashgaroth | when you find the good bead you pipette it out | 19:48 |
kanzure | "pop off" okay.. but it's still in the same well. | 19:48 |
fenn | yashgaroth: NO | 19:48 |
yashgaroth | argh | 19:48 |
@ParahSailin | might not even be wells, it could be emulsion bubbles | 19:48 |
kanzure | the point is, it's still separated from all the other molecules | 19:48 |
kanzure | (including sequencing reactants) | 19:48 |
fenn | kanzure: pop off means it literally goes flying through the air | 19:49 |
kanzure | to land where? | 19:49 |
fenn | in the assembly reaction well | 19:49 |
QuantumG | in the punter's drink, obviously | 19:49 |
fenn | which is conveniently placed directly underneath the bead | 19:50 |
fenn | pew pew | 19:50 |
fenn | plop plop | 19:50 |
ebowden | paperbot: file:http://www.nrjournal.com/article/S0271-5317(05)00104-1/pdf | 19:51 |
kanzure | underneath the bead was the bottom of the well | 19:51 |
QuantumG | I expect ya need to drop the file: | 19:52 |
fenn | the bead is hanging from the bottom side of a glass plate | 19:52 |
fenn | the laser shoots at it, dislodging it, and it falls straight down | 19:52 |
kanzure | so there are two entrances to each well? | 19:53 |
kanzure | you know, wells usually have at least one side | 19:53 |
kanzure | call 'em holes pls | 19:53 |
fenn | i dont know what the well looks like. i assumed it was a 4096 well plastic plate | 19:53 |
ebowden | paperbot: http://www.nrjournal.com/article/S0271-5317(05)00104-1/pdf | 19:54 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/5b9aea9f64767417d97210f5f97bf8a2.txt | 19:54 |
ebowden | paperbot: file:http://diyhpl.us/~bryan/papers2/paperbot/5b9aea9f64767417d97210f5f97bf8a2.txt | 19:54 |
kanzure | uh.. | 19:54 |
fenn | er, apparently 4096 well plates dont exist | 19:55 |
fenn | anyway, some relatively large yet still manageable number | 19:55 |
fenn | 4096 seems to be where i'm at today | 19:55 |
kanzure | so they are straight-through holes | 19:56 |
fenn | no | 19:56 |
fenn | let's say it's a 96 well plate | 19:56 |
kanzure | a 96 well plate usually has a bottom | 19:57 |
kanzure | such that you can only access each well through the top | 19:57 |
fenn | correct | 19:57 |
kanzure | now, if it's upside down, gravity takes your stuff | 19:57 |
fenn | no, the bead is hanging off a glass microscope slide. it's adhered by (unknown, irrelevant) | 19:57 |
kanzure | yes, but you wanted your bead to react to chemicals etc | 19:58 |
fenn | right | 19:58 |
fenn | it has dna growing on it | 19:58 |
yashgaroth | at like 1536 wells you can probably rely on capillary action | 19:58 |
ebowden | kanzure, paperbot does not appear to work with that site, does it work with sciencedirect? | 19:59 |
yashgaroth | against gravity that is | 19:59 |
kanzure | paperbot does not have 100% access to sciencedirect | 19:59 |
ebowden | Ok. | 19:59 |
kanzure | yashgaroth: so how do you insert the new reactants? | 19:59 |
fenn | yashgaroth: 4096 is only 64 by 64; for a 100mm plate that's like a 1mm diameter well | 19:59 |
fenn | maintaining +- 0.5mm is not that hard | 20:00 |
yashgaroth | fuck knows how you insert stuff into that, unless you're okay with using a pipette *cough* | 20:00 |
fenn | hell, even reprap can do it | 20:00 |
kanzure | yashgaroth: are you as confused as i am | 20:00 |
ebowden | what about elsevierhealth? | 20:00 |
yashgaroth | mhm | 20:00 |
fenn | i am confused about why you guys are confused | 20:00 |
kanzure | ebowden: try it, who cares? | 20:00 |
kanzure | fenn: have you ever used a microarray | 20:00 |
yashgaroth | how did you get to pew-pewing beads off of a plate | 20:01 |
kanzure | fenn: or, seen it enough to have a good understanding of how it works | 20:01 |
fenn | yashgaroth: we are talking about how the cambrian genomics "dna laser printer" works | 20:01 |
kanzure | here's a big plate: http://arrayit.com/Products/Microarray_Slides/Microarray_ProPlate_Glass/microarray_proplate_assemble.jpg | 20:01 |
fenn | why, i dont know | 20:01 |
yashgaroth | well it mostly runs on bullshit and VC money but that's just my opinion | 20:01 |
fenn | kanzure: what the hell is that | 20:02 |
kanzure | so is it capillary action? | 20:02 |
yashgaroth | that's a 96 well | 20:02 |
kanzure | plate gets inverted, capillary action to hold chemical solution in each well ? | 20:02 |
kanzure | so molecular feedstock is added prior to flipping plate? | 20:02 |
fenn | how does it even seal the liquid gap between wells | 20:02 |
fenn | or is contamination not important | 20:03 |
yashgaroth | if they're not filled up all the way you're fine | 20:03 |
yashgaroth | presumably if you've got beads attached to the bottom somehow, contamination isn't an issue anyway | 20:03 |
* kanzure nods | 20:04 | |
fenn | so each of those 96 wells has its own pattern with a zillion different dna strands? | 20:04 |
fenn | like a silicon wafer before it's cut into chips | 20:04 |
yashgaroth | per cambrian, each has one bead with 100k copies of an identical clone | 20:05 |
fenn | i'm confused what this picture is | 20:05 |
fenn | microarray_proplate_assemble.jpg | 20:05 |
yashgaroth | that's a plate with 96 wells in it | 20:06 |
fenn | no, it's a clear smooth piece of glass and a black plastic thing with 96 holes in it | 20:06 |
yashgaroth | glass optional, but yes | 20:06 |
kanzure | the holes only go through less than 100% of the material | 20:06 |
fenn | okay the glass being there is the confusing part | 20:07 |
fenn | god i should just stop | 20:07 |
QuantumG | DNA (or RNA) typewriter is a nice term | 20:08 |
yashgaroth | I remember us having the same discussion about how the hell cambrian's technology worked, some months ago | 20:08 |
fenn | "Arrayit offers ProPlate Microarray technology for users wishing to print microarrays on glass plates and subsequently attach the printed plates to adhesive microplates for 96-well and 384-well sample processing." | 20:09 |
fenn | Arrayit offers ProPlate Microarray technology for users wishing to print microarrays on glass plates and subsequently attach the printed plates to adhesive microplates for 96-well and 384-well sample processing. | 20:09 |
fenn | er, The adhesive provides a water-tight seal between microplate wells. | 20:09 |
yashgaroth | okay microarrays is more apt for this discussion, same deal where they're not separated but each reaction has a specific spot on the array | 20:10 |
yashgaroth | since they're bound to that spot you can isolate them with lasers or whatever, and contamination isn't an issue since the interesting stuff is stuck to a particular place | 20:10 |
kanzure | "Synthesis,1249034,NSF,NSF,SBIR,2013,1,150000.00,Cambrian Genomics Inc" | 20:11 |
fenn | okay i dont really care about this proplate thing | 20:11 |
kanzure | 665 Third StSuite 425San Francisco | 20:13 |
kanzure | phone: 6075920123 | 20:13 |
fenn | you can just watch the tv show, bloomberg brink on netflix | 20:13 |
kanzure | "This Small Business Innovation Research (SBIR) Phase I project attempts to radically reduce the cost of error-free oligonucleotides for use with gene synthesis. DNA Synthesis of large molecules is done by the assembly of many short oligonucleotide fragments of DNA 60-100bp in length. Currently, each DNA fragment is synthesized in relatively small numbers at an excessive macroscopic scale that incurs a large manufacturing overhead in its ... | 20:13 |
kanzure | ... production costs. This sets a high cost floor for the entire DNA synthesis process. One route to making cheaper oligonucleotides is to synthesize them in large sets on microarrays. However, array-synthesized DNA is both extremely error-prone and produced as dilute, complex mixtures. This proposed project will use massively parallel sequencers to sequence clonally amplified copies of DNA species sampled from the microarray in order to sort ... | 20:13 |
kanzure | ... every species apart from one another as well as to identify correctly synthesized oligonucleotides from incorrect ones. Further, it is proposed to use focused laser pulses in a custom laser ejection device to eject and recover desired subsets of perfect oligos from micron-scale sequenced colonies into multiwell plates for assembly into genes. The goal is to be able to recover tens of thousands of sequence-verified oligonucleotides in ... | 20:13 |
kanzure | ... several hours from sequencer flowcells. The broader impact/commercial potential of this project is to achieve truly disruptive cost decreases in the DNA synthesis of arbitrary genetic information. With the abundance of sequencing data, it is possible to imagine entering a new era of "constructive" biology, where in addition to classical reductive experiments on components, it will be possible to test our understanding of genetic- and ... | 20:13 |
kanzure | ... protein-based circuits by synthesizing new designs and measuring their discrepancy from predicted behaviors. The low-cost industrial production of arbitrary synthetic DNA has the potential to change the practice of biology such that it becomes cost effective to engineer whole genetic pathways and even genomes, accelerating the development of bioengineering, synthetic biomaterials production, as well as medical and research applications." | 20:14 |
QuantumG | so, I guess we need light activated phosphoribosyltransferases that respond to four distinct spectra. | 20:16 |
fenn | this could be an interesting method for wild DNA sampling; load up a plate of beads each with a single wild bacterium (unculturable) and see what's in it | 20:16 |
yashgaroth | so the principle is that you could pull single molecules out, isolate them, and then amplify each one enough to determine its sequence | 20:16 |
fenn | yashgaroth: correct | 20:17 |
fenn | and do that a million times in parallel | 20:17 |
yashgaroth | and do it on a small enough scale that it's more cost effective than traditional synthesis | 20:17 |
yashgaroth | but, good luck with that shit | 20:18 |
yashgaroth | which is why I'm hopeful but not optimistic on their approach | 20:18 |
heath | omg, python, how i've missed you | 20:18 |
kanzure | heath: except its packaging sucks :( | 20:18 |
heath | so much productivicheese | 20:18 |
fenn | programming languages shouldn't have package managers IMHO | 20:19 |
heath | fenn would like component | 20:19 |
fenn | eh? | 20:19 |
kanzure | a web thing? hell no | 20:19 |
fenn | "component" is a software framework or something (why do people keep using these un-googlable names) | 20:20 |
kanzure | it's like bower except not | 20:20 |
fenn | it's like <thing you've never heard of> but not | 20:20 |
fenn | oh? | 20:20 |
heath | whoops, for some reason, i though holowaychuk inc. were creating a language agnostic package manager | 20:20 |
xentrac | like dpkg? | 20:21 |
kanzure | like dpkg except not | 20:21 |
xentrac | well, like dpkg and apt? | 20:21 |
fenn | yeah it's called {dpkg, gentoo} | 20:21 |
kanzure | the problem is that python's ability to load a module sucks too | 20:21 |
kanzure | so it's not just packages that blow | 20:21 |
fenn | what's wrong with modules | 20:21 |
kanzure | you can't even unload a module | 20:21 |
-!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Remote host closed the connection] | 20:21 | |
fenn | why would you do that | 20:21 |
@ParahSailin | sounds like a challenge | 20:22 |
@ParahSailin | matplotlib's module loading side effects are totally evil | 20:22 |
kanzure | maybe you no longer need the module loaded | 20:22 |
fenn | can't you do "import foo as bar ; del bar" | 20:22 |
kanzure | that doesn't unload the module | 20:22 |
kanzure | nice try though | 20:22 |
cpopell | don't modules have scope limited to the function they're called in if you call them in a function? I was chatting about this with a friend just earlier today | 20:23 |
-!- nmz787_i [~nmccorkx@192.55.54.42] has joined ##hplusroadmap | 20:23 | |
fenn | kanzure: i dont know what you're talking about | 20:23 |
kanzure | reload() is also broken | 20:23 |
xentrac | cpopell: no, they go into the global module namespace in sys.modules, and also ld.so's symbol table | 20:23 |
xentrac | I tried to fix reload() more than ten years ago | 20:23 |
xentrac | I realized I don't know how | 20:24 |
cpopell | xentrac: hmmm | 20:24 |
cpopell | this was the result of our discussion earlier | 20:24 |
xentrac | the local name for the module is indeed limited to the function that imports it | 20:24 |
cpopell | mmm. | 20:24 |
cpopell | I see what you mean | 20:24 |
xentrac | but if another function or module imports the same module later, it won't get loaded a second time | 20:24 |
kanzure | cpopell: this is wizard magic unrelated to your immediate issue of scope | 20:24 |
fenn | okay so i load foo as bar1, delete bar1, change foo, load foo as bar2, now bar2 == bar1? | 20:25 |
xentrac | yes, except tht bar1 is a NameError | 20:25 |
cpopell | is there a module to simplify the action of deleting modules? | 20:25 |
cpopell | ... | 20:25 |
fenn | is there any reason you would ever do that? | 20:25 |
xentrac | you can del things from sys.modules; that works for the Python part | 20:25 |
kanzure | reload() is very very nifty... if it would actually work. | 20:25 |
xentrac | it doesn't work for the ld.so part | 20:26 |
@ParahSailin | why though | 20:26 |
QuantumG | paperbot http://pubs.acs.org/doi/abs/10.1021/bi00633a029 | 20:26 |
-!- Viper168_ [~Viper@unaffiliated/viper168] has quit [Ping timeout: 255 seconds] | 20:26 | |
@ParahSailin | side effects in module loading should just not be done | 20:26 |
kanzure | "Please accept that Python indeed does not support unloading modules for severe, fundamental, insurmountable, technical problems, in 2.x." | 20:27 |
kanzure | oh is it fixed in 3.x? | 20:27 |
nmz787_i | kanzure: a fab service recently told me "We process about 5 metric tons of glass per year into micro fluidic devices, micro-valves and pumps, micro reactors and MEM's devices." | 20:28 |
nmz787_i | so that sounds like some engineering to build on top of | 20:28 |
kanzure | great, feel free to pay for $500/prototype yourself | 20:28 |
kanzure | ugh | 20:28 |
nmz787_i | i only have to make 114 bp/second to make 300k gross/year | 20:29 |
@ParahSailin | thats a lot of glass | 20:29 |
nmz787_i | reaction will def be less than 60s, scaling # of reactors still takes up less than 10 sq cm real estate | 20:29 |
nmz787_i | and prob not more than a meter or two sq otherwise | 20:30 |
nmz787_i | likewise, if you consider desktop synthesizers, you'd need about 57 sq meter real estate, and that doesn't include other stuff, doesn't include purification, doesn't include transformation | 20:30 |
kanzure | neither does your chip | 20:31 |
nmz787_i | ( reaction on macro scale needs 2 mins) | 20:31 |
fenn | you'll have overhead beyond 10 sq cm of microfluidic plate | 20:31 |
nmz787_i | fenn: yeah that's what the extra ~2 sq meters was for | 20:31 |
paperbot | XMLSyntaxError: None (file "/home/bryan/code/paperbot/phenny/modules/scihub.py", line 51, in _go) | 20:31 |
nmz787_i | kanzure: yes it does | 20:31 |
kanzure | your chip doesn't exist | 20:31 |
nmz787_i | kanzure: that has been my plan for 4-5 years | 20:31 |
@ParahSailin | poor paperbot | 20:31 |
nmz787_i | my CAD has a v 0.1 | 20:32 |
QuantumG | I just used the library anyway | 20:32 |
kanzure | you can't just handwave a lot of engineering effort into existence and claim it will magically work | 20:32 |
nmz787_i | you have to set requirements before you start a project, generally | 20:32 |
fenn | oligos or GTFO | 20:32 |
fenn | lol | 20:32 |
nmz787_i | fuck oligo | 20:32 |
nmz787_i | :p | 20:32 |
@ParahSailin | if you are doing novel chemistry, dont do novel apparatus at the same time | 20:33 |
kanzure | i need a mathematical model that corresponds to your estimate of amount of work per added component on a project | 20:33 |
nmz787_i | nope | 20:33 |
kanzure | same goes with you fenn | 20:33 |
nmz787_i | not novel chem | 20:33 |
fenn | work is force times distance | 20:33 |
kanzure | just because it's not novel chemistry doesn't mean your layout works the first time or the first 100 times | 20:33 |
@ParahSailin | phosphoramidite? | 20:33 |
nmz787_i | though i just put out a call for a chemist/polymerchemist/materialsscientist today | 20:33 |
nmz787_i | yeah some PDMS compatible version | 20:34 |
kanzure | fenn: you know what i meant | 20:34 |
-!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 20:34 | |
@ParahSailin | by pdms you mean this is going to be a disposable device? | 20:34 |
fenn | you know what i meant. "work" is undefined otherwise | 20:34 |
kanzure | engineering time | 20:34 |
kanzure | number of iterations, prototypes, critical bugs | 20:34 |
fenn | number of bugs is proportional to number of interacting systems squared | 20:35 |
kanzure | is the material of each wall considered an "interacting system" | 20:35 |
fenn | probably not | 20:35 |
kanzure | is the yield rate of a chemical reaction considered an "interacting system" | 20:35 |
fenn | yield rate is a parameter, not a system | 20:36 |
kanzure | yield can be impacted by all sorts of interesting factors, like ambient environment, temperature, light, ph, etc | 20:36 |
fenn | great | 20:36 |
xentrac | speaking of chemistry, it occurred to me today why buses smell kind of like matches | 20:37 |
fenn | also parameters | 20:37 |
xentrac | a substantial part of match smoke is phosphorus oxide | 20:37 |
kanzure | what is your maximum amount of debugging then | 20:37 |
xentrac | the bus smell is from burning sulfur | 20:37 |
fenn | the rubber in the floor tiles contains sulfur vulcanizing residue? | 20:37 |
kanzure | because apparently you think this idea is below some threshold amount | 20:37 |
xentrac | no, from sulfur in the fuel | 20:37 |
kanzure | and i'm really confused what your personal estimates are here | 20:37 |
xentrac | here in Argentina, not in the US | 20:38 |
@ParahSailin | paperbot: http://nar.oxfordjournals.org/content/40/3/e23.full | 20:38 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/Error%20correction%20of%20microchip%20synthesized%20genes%20using%20Surveyor%20nuclease.pdf | 20:39 |
nmz787_i | ParahSailin: yeah, i don't want to think about cleaning it... but I've got flush lines abounding, so I think it probably will until it gets clogged or something. | 20:39 |
nmz787_i | *probably will keep working* | 20:39 |
fenn | kanzure: reading various essays about programming management (ie. the mythical man month) has led me to believe that estimating these kinds of things is beyond my ability | 20:39 |
@ParahSailin | abandon all hope of regenerating, reconditioning, etc the pdms, but im sure its good for prototyping | 20:39 |
nmz787_i | nah, not really regenerating, just more like, use it until it dies | 20:40 |
nmz787_i | flushing is part of the synthesis, so its there already | 20:40 |
nmz787_i | cleaning is sort of built in I guess you could say | 20:40 |
kanzure | QuantumG: do you know of any nasa reports regarding failure estimation? | 20:41 |
nmz787_i | unless the PDMS is slowly swelling, or if some corner is too sharp and causes crystallization to start and clog up | 20:41 |
fenn | kanzure: it's one thing to estimate something you've done before (known unknowns) and another thing completely to estimate something you've never done before (unknown unknowns) and then yet another thing completely for something that has never been done before | 20:41 |
@ParahSailin | wow ^paper made a 723nt construct | 20:41 |
QuantumG | NASA does the world's worst statistics | 20:41 |
QuantumG | and no | 20:42 |
xentrac | I thought PDMS was pretty dimensionally stable | 20:42 |
nmz787_i | xentrac: depends on solvent absorbtivity | 20:42 |
@ParahSailin | pdms swells | 20:42 |
kanzure | the slide with "weighted technology impact rating" is weird (the ones before it are worthless) http://start1.jpl.nasa.gov/pubPro/Alan_Wilhite.pdf | 20:42 |
@ParahSailin | http://www.biotechniques.com/multimedia/archive/00036/BTN_A_04364PF01_O_36417a.pdf | 20:43 |
kanzure | QuantumG: how is that possible, don't they do the famous "1 failure in a billion" numbers | 20:43 |
xentrac | nmz787_i: thanks | 20:43 |
QuantumG | yeah, how do you think they come up with those numbers? | 20:44 |
kanzure | oh :( | 20:44 |
kanzure | this seems like something that large-scale engineering projects need to have had addressed though | 20:45 |
kanzure | like the LHC.. you don't just go put that sucker in the ground without knowing the constraints, tolerances, and expected failure rate. | 20:45 |
QuantumG | heheh.. ya do know they did exactly that right? | 20:45 |
kanzure | huh? | 20:45 |
QuantumG | hooked up all the superconducting doo-dahs and discovered that not only didn't it work but that it wasn't even close to a functioning design | 20:46 |
kanzure | that's pathetic | 20:46 |
nmz787_i | all those enzyme things seem silly though, DNA synthesis isn't really super crazy complex of a reaction... there are like 8 or 10 reagents involved, keep it dry of water, get your metering right and your purification in-line and that's that | 20:46 |
nmz787_i | block co-polymer synthesis is all around | 20:47 |
nmz787_i | ABS | 20:47 |
-!- QuadIngi [~fourfire@245-201-15.connect.netcom.no] has quit [Ping timeout: 264 seconds] | 20:47 | |
kanzure | "Early estimates by managers at NASA put the risk of catastrophic failure during a mission at 1 in 100,000; engineers, on the other hand, put the estimate at 1 in 100. A study recently released using all the data available from 30 years of flights reveals that the danger was in fact much higher. " | 20:47 |
nmz787_i | that's only 1 monomer type less than DNA code | 20:47 |
fenn | QuantumG: why didn't the LHC work as designed? | 20:48 |
QuantumG | I love the way history on Wikipedia for the LHC starts in 2008. | 20:48 |
* fenn is too lazy to read up on it | 20:48 | |
QuantumG | 10 Sep 2008 CERN successfully fired the first protons around the entire tunnel circuit in stages. | 20:48 |
QuantumG | yeah, cause the LHC just magically popped into existence. | 20:48 |
QuantumG | The construction of LHC was approved in 1995 | 20:49 |
xentrac | kanzure: they did run into a lot of unexpected failures actually | 20:49 |
fenn | the LHC was warped in via microsingularity transport | 20:49 |
xentrac | which is what usually happens when you do things nobody has done before | 20:49 |
kanzure | there are lots of ways to estimate likely types of failures | 20:50 |
kanzure | did they even have a failure estimation model | 20:50 |
QuantumG | http://en.wikipedia.org/wiki/Large_Hadron_Collider#Construction_accidents_and_delays | 20:50 |
kanzure | in a system with 1 million parts, it's probably a good idea to have at least a rough estimate of failure, or even make up fake lower bounds which tell you how much manpower you will need for maintenance or replacement etc | 20:50 |
QuantumG | it's the second last section before "Popular culture" and "See also". | 20:50 |
fenn | kanzure: and the actual measured catastrophic failure rate was 2 in 135 | 20:51 |
QuantumG | .. and doesn't even include the stuff I vaguely remember reading. | 20:52 |
fenn | plus or minus some hand waving | 20:52 |
kanzure | insert stereotypical "haha they used english feet and we used meters" story | 20:52 |
QuantumG | http://web.archive.org/web/20080616063402/http://www.photonics.com/content/news/2007/April/4/87089.aspx | 20:52 |
xentrac | kanzure: I'm sure they did, or they wouldn't have been able to ever get it to work | 20:52 |
QuantumG | .title | 20:52 |
yoleaux | Fermilab 'Dumbfounded' by Fiasco That Broke Magnet | 20:52 |
fenn | the mars orbiter screwup was actually much more complicated and hard to explain than the newtons/foot-poundals mixup that we've been told | 20:53 |
kanzure | "Fermilab designed and built nine of the magnet assemblies for CERN and conducted four engineering reviews between 1998 and 2002" | 20:53 |
kanzure | aka "someone looked at this only four times" :) | 20:53 |
xentrac | journalists always love it when "boffins" are "baffled" | 20:53 |
xentrac | they can safely be ignored | 20:54 |
QuantumG | in a project that large, the reviews are nothing more than rubber stamps anyway | 20:54 |
kanzure | i feel like freitas/merkle might have written about estimation in their book on kinematic self-replicating machines | 20:54 |
kanzure | oh wait, maybe AASM instead | 20:55 |
QuantumG | there's plenty of it in Nanosystems | 20:55 |
kanzure | AASM seems like the place they would do it | 20:55 |
* ParahSailin is disappointed that nobody looked at the awesome paper i linked | 20:55 | |
kanzure | http://www.islandone.org/MMSG/aasm/ | 20:55 |
xentrac | QuantumG: why do you think that projects being large requires that reviews be ineffective? | 20:56 |
kanzure | haha "Because of the already challenging design problem, it is highly desirable to keep all seed systems as simple as possible in both structure and function. This should help reduce the risk of partial or total system failure and make closure less difficult to achieve at all levels." | 20:56 |
-!- drewbot [~cinch@ec2-54-198-249-125.compute-1.amazonaws.com] has quit [Remote host closed the connection] | 20:56 | |
fenn | ParahSailin: i thought you were being sarcastic | 20:56 |
kanzure | that' sthe only mention of failure | 20:56 |
-!- drewbot [~cinch@ec2-54-198-29-24.compute-1.amazonaws.com] has joined ##hplusroadmap | 20:56 | |
QuantumG | ParahSalin: looks neat | 20:57 |
@ParahSailin | i think it has potential to be scaled | 20:57 |
fenn | in the diybio post made by anselm (linked here a couple hours ago) he says that mismatch endonucleases chew up all your DNA, leaving very little yield | 20:58 |
kanzure | also, a lot of this is part failure, like mechanical failure under load | 20:58 |
kanzure | rather than system design problems | 20:58 |
kanzure | i dunno what to call the system design type of problem | 20:58 |
@ParahSailin | fenn: apparently not all | 20:59 |
fenn | systemantics? | 20:59 |
-!- nmz787_i [~nmccorkx@192.55.54.42] has quit [Quit: Leaving.] | 20:59 | |
fenn | ParahSailin: if you have very low yield you need to amplify (probably with PCR) which has some finite error rate of its own | 20:59 |
* fenn shrugs | 21:00 | |
fenn | i guess there are fancy low error PCR enzymes now | 21:00 |
QuantumG | kanzure: NASA is a great place to propose a barely working concept, get billions spent on trying to figure out how it could possibly work (because failure isn't an option!) and then move onto the next thing (after not failing). | 21:00 |
fenn | the shuttle mentality | 21:00 |
@ParahSailin | fenn: expectations are really low right now for dna synthesis | 21:01 |
kanzure | i am a shuttle voom | 21:01 |
fenn | nyoom | 21:01 |
QuantumG | Ares I is what I had in mind, but the examples abound | 21:01 |
@ParahSailin | even with a couple pcr errors, a multi kb construct is sploosh | 21:01 |
fenn | the average nasa engineer during apollo was 26 | 21:01 |
QuantumG | everything is either a success or cancelled by those stupid policy makers at NASA. | 21:02 |
QuantumG | (sarcasm) | 21:02 |
fenn | average nasa age is now 47 | 21:02 |
QuantumG | if your spacecraft crashes, that's not a failure, that's.. like.. space is hard man | 21:02 |
QuantumG | if you put a rover on Mars that costs 10x more than you said it would at the start of the program.. that's a success. | 21:03 |
fenn | i'm more worried about not sending anything up at all | 21:03 |
-!- pads [~not@100.43.114.90] has joined ##hplusroadmap | 21:03 | |
justanotheruser | That's science! Their theory that their mission would be successful was wrong | 21:03 |
fenn | who cares if it crashes | 21:03 |
-!- pads is now known as Guest49564 | 21:04 | |
fenn | we still dont have an equatorial LEO laboratory | 21:04 |
fenn | stuff costs too much | 21:05 |
QuantumG | yeah, it costs too much because of that attitude | 21:05 |
fenn | soon we'll have unpiloted aerostats and the satellite industry will die, taking the rest of commercial space with it, unless something big changes | 21:06 |
QuantumG | "soon" has been soon for a long time when it comes to aerosats | 21:06 |
xmj | moin | 21:06 |
fenn | the FAA is dragging its knuckles | 21:06 |
QuantumG | I don't think that's the problem | 21:07 |
fenn | i think it is | 21:07 |
QuantumG | how do you figure? | 21:07 |
fenn | well, why don't we have aerostats flying around? | 21:07 |
fenn | they exist | 21:08 |
kanzure | 1) make a correct failure estimation model 2) proceed to building dyson sphere | 21:08 |
kanzure | *engineering design failure | 21:08 |
fenn | expected chance of success building dyson sphere: 1e-1337 | 21:08 |
kanzure | you have already failed | 21:08 |
fenn | you are already dead | 21:08 |
xentrac | we do have aerostats flying around for weather monitoring | 21:09 |
cpopell | muda muda muda | 21:09 |
QuantumG | I'm pretty sure you mean aerosats.. aerostats already exist and fly around the US all the time | 21:09 |
xmj | fenn: "still bigger than the last web startup i funded. here, have some VC money" | 21:09 |
kanzure | do you think the death punch is possible? you could inject some chemicals that cause a body to explode, right? | 21:09 |
fenn | xentrac: a balloon is not an aerostat because it moves randomly | 21:09 |
xmj | can aerostats be geostationary? | 21:09 |
cpopell | kanzure: what about the high pressure knives? | 21:10 |
xentrac | fenn: I think it is: https://en.wikipedia.org/wiki/Aerostat#Free_balloons | 21:10 |
fenn | whatever | 21:11 |
fenn | let's abuse language and make everything mean nothing and nothing everything | 21:12 |
xentrac | .ety aerostat | 21:12 |
yoleaux | Sorry, I couldn't find the etymology of that. | 21:12 |
xentrac | what's the FAA doing? | 21:12 |
QuantumG | in any case, there's plenty of FAA support for balloon delivered comms services.. I don't think that's the problem. | 21:12 |
QuantumG | I think the problem is that flying a permanent balloon platform is really hard. | 21:12 |
fenn | xentrac: the "stat" part means "static" as in "not moving" | 21:13 |
fenn | "aero" means it's in the air | 21:13 |
xentrac | there is a discussion of this at https://en.wikipedia.org/wiki/Talk%3AAerostat#Meaning | 21:13 |
fenn | a balloon without motive power. Contrasted with {aerodyne}. [1913 Webster | 21:14 |
xentrac | it turns out that "stat" means that its lift is not derived from moving, as opposed to aerodynes | 21:14 |
xentrac | "All three of those bulleted definitions mean exactly the same thing, which has been the conventional meaning for well over a century. The second meaning referred to is the recent narrow use of the "stat" bit to indicate static mooring in contrast to a free-flying airship, e.g. as used by the US General Accounting Office, but it is very much a minority usage by those who are not technically knowledgeable." | 21:14 |
fenn | so a kite is not an aerostat? | 21:16 |
fenn | what's the point of even having the word | 21:16 |
xentrac | no, it's lifted by aerodynamic lift | 21:16 |
fenn | why not just say "balloon" | 21:17 |
xentrac | is a zeppelin a balloon? | 21:17 |
fenn | yes | 21:17 |
@ParahSailin | greek roots sound more professional | 21:17 |
xentrac | not normally | 21:17 |
xmj | ParahSailin: moin | 21:17 |
xentrac | the point of having the word is probably to distinguish the two major types of aircraft: heavier-than-air and lighter-than-air | 21:18 |
fenn | stationkeeping flying thingy is too much | 21:18 |
QuantumG | anyway, I dunno how commsats are still around.. people like their soccer delivered that way I guess | 21:18 |
@ParahSailin | møin møin møin | 21:18 |
QuantumG | and people keep building LEO constellations despite the fact that none have ever turned a profit. | 21:19 |
fenn | and "statite" already has a meaning related to solar sails | 21:19 |
xentrac | there are probably classified funding source for LEO constellations | 21:19 |
-!- Adifex|pub is now known as Adifex | 21:19 | |
QuantumG | nah VCs just like the idea. | 21:20 |
QuantumG | this time it'll be different because SpaceX | 21:20 |
-!- Qfwfq [~Qfwfq@unaffiliated/washirving] has quit [Ping timeout: 252 seconds] | 21:20 | |
fenn | what does SpaceX do that radically changes things? | 21:20 |
QuantumG | cheaper launch, is the argument | 21:21 |
xentrac | control costs | 21:21 |
QuantumG | mostly cheaper because you don't have to take your payloads to Russia, but yeah | 21:21 |
fenn | wasnt this the point of sealaunch | 21:21 |
-!- delinquentme [~dingo@74.61.157.78] has quit [Ping timeout: 252 seconds] | 21:22 | |
QuantumG | yeah, and they got sued | 21:22 |
xentrac | didn't know that | 21:22 |
xentrac | .g sealaunch lawsuit | 21:22 |
yoleaux | http://www.reuters.com/article/2013/02/04/boeing-sealaunch-idUSL1N0B31GP20130204 | 21:22 |
fenn | sued for what | 21:22 |
QuantumG | itar | 21:22 |
fenn | huh? | 21:22 |
fenn | because sailing is "export"? | 21:23 |
xentrac | oddly enough, that article says: | 21:23 |
QuantumG | no, Boeing failed to keep their tech out of Russian/Ukranian hands | 21:23 |
xentrac | "Sea Launch sent its first satellite into space in 1999, but filed for bankruptcy in 2009 because of weaker demand, mounting debt and a failed launch that led to a $53.2 million arbitration award against the company" | 21:23 |
fenn | so, not itar? | 21:24 |
xentrac | .g "sea launch" itar | 21:24 |
yoleaux | http://www.alston.com/files/docs/3-23-11-Items-Insert.pdf | 21:24 |
QuantumG | yeah, part of that weaker demand was the supposed benefits of using them over Russian launch - not going through the ITAR process - was proven false.. graphically | 21:24 |
xentrac | "The assets owned by Sea Launch created other | 21:25 |
xentrac | regulatory and defense-related issues, including the fact that the equipment (the vessel and the platform | 21:25 |
xentrac | utilized to launch rockets) is subject to the International Trafficking in Arms Regulations (ITAR). The | 21:25 |
xentrac | ITAR regulations created issues in connection with obtaining DIP financing as lenders were concerned | 21:25 |
xentrac | about the ability to liquidate and monetize those assets. Today Sea Launch is now reorganized and | 21:25 |
xentrac | operating in Luxemburg and Switzerland (with the physical assets still in Long Beach, California, the | 21:25 |
xentrac | reorganization was successful and created significant value for the stakeholders and customers. | 21:25 |
xentrac | " | 21:25 |
-!- ebowden [~ebowden@CPE-60-231-177-134.lns4.dav.bigpond.net.au] has quit [Remote host closed the connection] | 21:25 | |
fenn | i think sealaunch was a great idea and SpaceX should take advantage of it | 21:26 |
fenn | it's practically in their back yard | 21:26 |
xentrac | Sea Launch was largely Russian-owned | 21:26 |
QuantumG | dear god why? | 21:26 |
xentrac | and Ukrainian-owned | 21:26 |
xmj | xentrac: that means essentially the same thing, doesn't it. | 21:27 |
QuantumG | the launching from the equator stuff was never the purpose of the partnership.. that was just some nonsense to avoid export regulations and it didn't work | 21:27 |
xmj | just use Soviet-owned. | 21:27 |
xmj | :] | 21:27 |
fenn | i never knew anything about the russian partnership | 21:27 |
QuantumG | The US now has "export free zones" like the Mid-Atlantic Regional Spaceport to do something similar. | 21:28 |
dingo | ☭ | 21:28 |
fenn | um except Virginia is nowhere near the equator | 21:28 |
QuantumG | all of this stuff stems from the time when it was considered crazy talk to say you were going to make your own rockets. | 21:29 |
QuantumG | launching from the equator gives you a tiny performance improvement which doesn't make up for the all the butt-hurt you get from having to go outside the US.. ya know, assuming you're a US company | 21:29 |
xmj | lol | 21:30 |
xentrac | cos(latitude(virginia)) is pretty close to 1 | 21:30 |
QuantumG | Orbital Sciences launches from MARS because 90% of their rocket is made in other countries. | 21:30 |
xmj | yes, because local subsidiaries to circumvent export regulation is... hard to do, really | 21:30 |
fenn | it also provides transportation to your launch site, a wide open launch range, a launch pad | 21:31 |
QuantumG | xmj: local subsidiaries full of non-US citizens? | 21:31 |
fenn | i thought orbital sciences used "MARS" (what a terrible name, it's WALLOPS FLIGHT FACILITY) because of a treaty that forbids launches of ICBMs except from vandenberg and virginia | 21:32 |
QuantumG | that could be why WALLOPS exists, but it's not why Orbital cares | 21:33 |
fenn | orbital was using a minotaur V (when i was paying attention) | 21:34 |
fenn | the LADEE launch | 21:34 |
fenn | Orbital's Minotaur V launches LADEE mission to the Moon ... | 21:34 |
fenn | Launch to Occur from NASA's Wallops Flight Facility in Virginia on Friday --(Wallops Island, VA 5 September 2013) | 21:34 |
@ParahSailin | also is pretty cool that railroads go to virginia, and not french guyana | 21:34 |
xentrac | well, LADEE DA | 21:34 |
QuantumG | yeah? and? | 21:35 |
xmj | ParahSailin: boats go | 21:35 |
fenn | sorry for the spam | 21:35 |
fenn | i watched that launch through binoculars, it was pretty cool | 21:35 |
fenn | very orange | 21:35 |
-!- entelechios [~elysium@mail.2bett.com] has quit [Ping timeout: 252 seconds] | 21:36 | |
xentrac | .g latitude of virginia | 21:36 |
yoleaux | http://www.worldatlas.com/webimage/countrys/namerica/usstates/valatlog.htm | 21:36 |
fenn | .wa latitude of virginia | 21:36 |
xentrac | .g cos(37.5 degrees) | 21:36 |
yoleaux | https://answers.yahoo.com/question/index?qid=20100515221911AA8dWT5 | 21:36 |
yoleaux | Virginia: center coordinates: 37° 31' 7"N, 78° 53' 28"W | 21:36 |
fenn | come on wolfram, you don't know what latitude means? | 21:36 |
xentrac | .c cos(37.5 degrees) | 21:36 |
yoleaux | cos(37.50°) = 0.793353... | 21:36 |
@ParahSailin | .wa circumference of earth | 21:37 |
yoleaux | Earth: equatorial circumference: 40075.017 km (kilometers); Unit conversions: 24901.461 miles; 4.0075017×10⁷ meters; 21638.778 nmi (nautical miles); 0.13367587 light seconds; Comparison as length: ~1.9 × length of the Great Wall of China (~21196 km); Corresponding quantities: Light travel time t in vacuum from t = x/c:: 134 ms (milliseconds); Light travel time t in an optical fiber t = 1.48x/c:: 198 ms (milliseconds) | 21:37 |
xentrac | .wa circumference of earth / 1 day | 21:37 |
yoleaux | Earth: equatorial circumference/(1 day): 40075.017 km/day (kilometers per day); Unit conversions: 463.83121 m/s (meters per second); 0.46383121 km/s (kilometers per second); 0.28821135 mi/s (miles per second); 17.292681 mi/min (miles per minute); 1037.5609 mph (miles per hour) | 21:37 |
fenn | .wa latitude of vandenberg california | 21:37 |
yoleaux | fenn: Sorry, no result! | 21:37 |
xentrac | 463 m/s | 21:37 |
fenn | .wa latitude of vandenberg air force base | 21:37 |
yoleaux | Vandenberg AFB: latitude: 34° 44' 14"N; Unit conversions: 606.3 mrad (milliradians); 0.6063 radians; 34 degrees 44 arc minutes 14.4 arc seconds; 2^h 18^m 56.96^s (right ascension); 38.6 grads (unit officially deprecated) | 21:38 |
xentrac | .wa leo speed | 21:38 |
yoleaux | Leo (male given name): Speed (male given name): Information for births: | Leo: Speed; rank: 134th: beyond 1000th; fraction: 1 in 704 people (0.14%): less than 1 in 12500 people (0.008%); number: 2855 people per year: < 200 people per year (US data based on 2012 births and other SSA registrations in the US) | 21:38 |
xentrac | .g leo speed | 21:38 |
yoleaux | http://en.wikipedia.org/wiki/Low_Earth_orbit | 21:38 |
fenn | oh but vandenberg can't do equatorial launches because it would kill those honest god fearing people in nebraska | 21:38 |
@ParahSailin | delta v starts at about 10km/s | 21:38 |
xentrac | yeah, I was trying to figure out whether I want to deduct the 21% of 463 m/s from delta v or orbital velocity: | 21:39 |
xentrac | "orbital velocity needed to maintain a stable low earth orbit is about 7.8 km/s, but reduces with increased orbital altitude." | 21:39 |
xentrac | .c 21% * 463 m/s | 21:39 |
yoleaux | 21%×463 m/s (meters per second) = 97.23 m/s (meters per second) | 21:40 |
xentrac | .c 21% * 463 m/s / (7.8 km/s) | 21:40 |
yoleaux | 21%×(463 m/s (meters per second))/(7.80 km/s (kilometers per second)) = 0.0125 | 21:40 |
fenn | from delta V; orbit is around the center of earth, not any arbitrary latitude line | 21:40 |
xentrac | .c 21% * 463 m/s / (7.8 km/s - 463 m/s) | 21:41 |
yoleaux | 21%×(463 m/s (meters per second))/(7.80 km/s (kilometers per second)-463 m/s (meters per second)) = 0.0133 | 21:41 |
xentrac | so launching from the equator instead of Virginia saves you 1⅓% of the orbital velocity you need | 21:41 |
xentrac | or more like 1% if the delta-v is actually the relevant number, as ParahSailin says it is | 21:42 |
QuantumG | .. and most launches are underperforming anyway. | 21:42 |
fenn | when 95% of your mass is fuel, 1% is a lot | 21:43 |
xentrac | how much, fenn? | 21:43 |
fenn | how much what? | 21:43 |
xentrac | how much is 1%? | 21:43 |
@ParahSailin | fuel requirement is geometric wrt delta v | 21:43 |
fenn | 0.01? | 21:43 |
xentrac | or is the answer to that question really "come back after you've played Kerbal Space Program for a couple of weeks"? | 21:44 |
@ParahSailin | .g tsiolkovsky equation | 21:44 |
yoleaux | http://en.wikipedia.org/wiki/Tsiolkovsky_rocket_equation | 21:44 |
Darius | what does .c use? | 21:45 |
kanzure | chinese children | 21:45 |
Darius | they're very quick | 21:45 |
kanzure | well there's a lot | 21:45 |
fenn | i hate wikipedia's equations, they are light colored staticky pixels around black letters on a gray background, pretty much ensuring i can't read them | 21:45 |
kanzure | switch the css theme, there's a dropdown box somewhere | 21:46 |
kanzure | the css theme might determine their latex generator image thing | 21:46 |
fenn | wikipedia has themes? | 21:47 |
kanzure | http://en.wikipedia.org/wiki/Wikipedia:Skin#How_to_change_the_skin | 21:47 |
xentrac | Darius: on #swhack, where yoleaux comes from, it's currently using Wolfram Alpha | 21:47 |
kanzure | "Once you have an account, go to your preferences and go to the "Skin" section of the Appearance tab. With the default skin, this page can be accessed at any time from the "my preferences" link placed on the top right corner. Choose your skin and then click Save. Then, all pages will be loaded with the new selected skin." | 21:47 |
@ParahSailin | Isp of kerosene/lox is 353s | 21:48 |
xentrac | but I don't know if yoleaux has been updated yet | 21:48 |
kanzure | "However, any user may change the skin of a page for them only, but only one each time, by typing ?useskin=skinname to the end of the URL (i.e.: ?useskin=vector)" | 21:48 |
fenn | xentrac: the simple answer is, weigh your rocket fully fueled, multiply its mass by 1% and that's the amount of extra payload you get by launching at the equator | 21:49 |
xentrac | fenn: ah, so 20%? | 21:50 |
xentrac | (in that 100%-95%-1% is 120% of 100%-95%) | 21:51 |
fenn | oh also " the minimum orbital inclination without a plane-change maneuver is the latitude of the launch site" | 21:51 |
xentrac | but ParahSailin says "fuel requirement is geometric wrt delta v" so perhaps not | 21:52 |
@ParahSailin | if you can get 6.4% in orbit with the balance as fuel, going for delta v of 9.5 km/s, you can get 6.2% launch mass in orbit at 9.5 * 1.01 delta v | 21:52 |
xentrac | that's a substantial improvement but not nearly as much as my other number | 21:55 |
fenn | kanzure: thanks i was guessing randomly ?skin=dark ?skin=black | 21:55 |
fenn | they are all black text on white background though? | 21:56 |
-!- agentsmith2 [agentsmith@cpe-75-80-110-69.socal.res.rr.com] has joined ##hplusroadmap | 21:59 | |
-!- yashgaroth [~ffffff@cpe-76-167-105-53.san.res.rr.com] has quit [Quit: Leaving] | 21:59 | |
xmj | fenn: did you see http://opensslrampage.com/ ? | 21:59 |
xmj | er | 22:00 |
xmj | fenn: did you see http://www.libressl.org/ ? | 22:00 |
xmj | This page scientifically designed to annoy web hipsters. Donate now to stop the Comic Sans and Blink Tags | 22:00 |
fenn | no, why should i care? | 22:01 |
fenn | SSL sucks anyway | 22:01 |
xentrac | fenn: where does your gray background come from? | 22:01 |
xentrac | my Wikipedia background is always white (so I can't see the staticky pixels) | 22:02 |
fenn | cat ~/.dillo/style.css | 22:02 |
fenn | * { background-color: #272727 !important; ... | 22:02 |
@ParahSailin | i've done some patches to hs-tls | 22:02 |
xentrac | ah | 22:03 |
xentrac | I didn't know Dillo even supported CSS | 22:04 |
fenn | it doesn't really | 22:04 |
fenn | it will do borders and text and colors, but not float or various box model things | 22:05 |
fenn | as for why i'm using dillo, see previous complaints about thrashing and ram and tabs | 22:06 |
fenn | kanzure: i thought you had a watermark removal script | 22:09 |
-!- entelechios [~elysium@181.194.129.225] has joined ##hplusroadmap | 22:09 | |
fenn | i mean, i thought it was running on paperbot's output | 22:09 |
xentrac | no, I totally understand wanting to use dillo | 22:09 |
fenn | but i clearly see a watermark from a paper downloaded today at <location redacted> | 22:09 |
fenn | xentrac: i recently discovered "netsurf" which, despite its odd origins and intended audience, is actually a damn fine first draft of a browser | 22:10 |
fenn | it just needs javascript and some minor bugfixes | 22:10 |
fenn | netsurf is as fast as dillo, but it properly renders CSS | 22:11 |
-!- helleshin [~talinck@66-161-138-110.ubr1.dyn.lebanon-oh.fuse.net] has quit [Read error: Connection reset by peer] | 22:12 | |
-!- helleshin [~talinck@66-161-138-110.ubr1.dyn.lebanon-oh.fuse.net] has joined ##hplusroadmap | 22:12 | |
-!- HashNuke [uid12117@gateway/web/irccloud.com/x-egcplrxinrchzrjl] has quit [Quit: Connection closed for inactivity] | 22:12 | |
xentrac | I think I tried using netsurf once but couldn't get it to stop crashing | 22:13 |
xentrac | but I might have confused it with something else | 22:13 |
kanzure | fenn: pdfparanoia does not have 100% coverage | 22:13 |
kanzure | fenn: https://github.com/kanzure/pdfparanoia/issues report a bug | 22:13 |
xentrac | presumably if it doesn't have JS it doesn't have the DOM either | 22:13 |
xentrac | which means that getting it to run modern JS would presumably involve implementing the DOM and restructuring its rendering around the DOM, which I presume would make it no longer as fast as dillo or in the same ballpark | 22:14 |
kanzure | servo is the only modern practical option at this point | 22:14 |
kanzure | webkit and blink are both massive | 22:14 |
xentrac | servo is still a proof of concept | 22:15 |
xmj | it's too bad opera stamped their engine and settled for webkit | 22:15 |
xmj | opera's was by far the fastest | 22:16 |
xentrac | I didn't realie that | 22:16 |
xentrac | realize | 22:16 |
fenn | oh, oxford journals is already listed as a bug | 22:16 |
xentrac | I stopped using opera in 2000 | 22:16 |
fenn | 90% of the time i don't care about javascript, i just want to see the page | 22:17 |
fenn | so maybe have a browser that loads webkit for the other 10% | 22:17 |
fenn | or i could just switch browsers i guess | 22:17 |
xentrac | I'm wondering at this point if it makes sense to write a browser that recognizes bootstrap and has a fast path for it | 22:18 |
fenn | actually midori worked fine before i "upgraded" to 13.10 | 22:18 |
xentrac | midori's still webkit, no? | 22:18 |
fenn | yes, but it was pretty fast still | 22:18 |
xentrac | I guess I figure that 95% of the stuff on current web sites is trying to accomplish pretty simple things | 22:18 |
fenn | and it didn't explode on more than 10 tabs, like all mainstream browsers do | 22:18 |
fenn | what do you mean "has a fast path for it"? | 22:19 |
xentrac | if you could stick in optimizations to recognize which things people are trying to accomplish, and then accomplish them in more efficient ways, you could probably save a lot of memory and CPU | 22:19 |
fenn | oh like a self optimizing JIT for javascript | 22:19 |
xentrac | no | 22:20 |
fenn | (i thought chrome did that with v8) | 22:20 |
xentrac | like a large library of extremely specific optimizations for not just JS but also CSS and HTML | 22:20 |
xentrac | mostly CSS and HTML in fact | 22:20 |
kanzure | "oh look, the billionth webpage that's using float:left" | 22:20 |
kanzure | "and jquery" | 22:20 |
xmj | xentrac: did you see the stackoverflow page telling you to use jquery for simple things like adding two numbers? | 22:21 |
fenn | i understand having a local cache of jquery, but how do you fix float:left? | 22:21 |
xentrac | I don't know specifically what the things are that make e.g. Twitter take multiple seconds to scroll down | 22:21 |
xentrac | on my netbook | 22:21 |
xentrac | but the end result is a pretty mundane change to what's on the screen | 22:22 |
kanzure | why are all the google scholar results for "design flaws" all about software? :( | 22:22 |
fenn | usually my problem is: i open a new tab, paste in a url, and nothing happens for 5-50 seconds | 22:22 |
xentrac | you would think that whatever particular chunk of DOM that Bootstrap is constructing and then twiddling, I could stick a Bootstrap-DOM-recognizing hack into the browser engine that recognizes this particular case as "oh yeah, that's Bootstrap DOM for a pulldown menu" | 22:23 |
fenn | somehow chrome is using 4GB of (virtual) RAM | 22:23 |
fenn | its like it doesn't understand the difference between ram and swap | 22:24 |
xentrac | or "this is going to scroll the page down" | 22:24 |
fenn | how is this different from memoizing function calls | 22:25 |
kanzure | "design errors" is also all software and vlsi stuff. so, what, design errors don't happen in buildings or supercolliders? | 22:25 |
fenn | i guess the DOM varies because the colors are different or something? | 22:25 |
fenn | so the function call is not exactly the same and you get a cache miss | 22:26 |
Lemminkainen | paperbot http://www.sciencedirect.com/science/article/pii/S0163725811001653 | 22:26 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/fd72b379e6676e13c477f45520c6534a.txt | 22:26 |
xentrac | I guess I'd have to dig into some DOM to find out | 22:26 |
fenn | xentrac: what do you think about throwing away all the webmaster-provided styling information and using LCARS instead? | 22:28 |
xentrac | <sadtrombone /> | 22:28 |
xentrac | I think I'd rather read Wikipedia with any of the Wikipedia skins than with LCARS | 22:29 |
fenn | but styling is just reinventing the UI wheel for every website in existence | 22:29 |
fenn | fine, s/LCARS/vector/ | 22:29 |
xentrac | it would be okay for some pages | 22:29 |
xentrac | and not others | 22:29 |
fenn | the way i see it, there's a page with some data (content) and some affordances (menus, buttons, etc) and i really dont give a shit about anything else | 22:30 |
fenn | all that stuff can just fuck off and leave me alone | 22:31 |
-!- QuantumG [~chatzilla@unaffiliated/quantumg] has quit [Remote host closed the connection] | 22:31 | |
fenn | custom mouse cursor that sprinkles fairy dust -> /dev/null | 22:31 |
xentrac | even if you just remove the inline style tag from e.g. http://canonical.org/~kragen/sw/aspmisc/my-very-first-raytracer.html it becomes substantially less readable | 22:32 |
fenn | aside from the large renderings breaking up the text flow, i dunno what i'm missing | 22:33 |
* fenn waits for chrome to do anything at all | 22:34 | |
fenn | processing request... | 22:34 |
fenn | processing... | 22:34 |
fenn | [Pages Unresponsive] | 22:35 |
xentrac | pasting the style tags from https://en.wikipedia.org/wiki/Wikipedia:Vital_articles into it improves the situation somewhat | 22:35 |
fenn | oh there it is. oh look, it has a logo in the upper left corner | 22:35 |
xentrac | at least the <pre> tags have a different background | 22:36 |
fenn | and the code text flows into the rendering making it hard to read | 22:36 |
-!- Darius [~quassel@adsl-69-231-35-194.dsl.irvnca.pacbell.net] has quit [Read error: Connection reset by peer] | 22:36 | |
xentrac | but all the text is run up against the left edge of the page with a single pixel of whitespace, which makes it hard to read | 22:36 |
xentrac | and the width of the text is way too wide | 22:36 |
kanzure | conceptual design flaws exist, right? | 22:37 |
kanzure | it's not all just "your load bearing loadbearer wasn't able to bear enough bears"? | 22:37 |
fenn | in dillo the <pre> text has a different font and font size and it has a box around it, quite clear it's a block of code | 22:37 |
xentrac | that's good | 22:37 |
-!- Darius [~quassel@adsl-69-231-35-194.dsl.irvnca.pacbell.net] has joined ##hplusroadmap | 22:37 | |
xentrac | I put CSS in there for that | 22:37 |
fenn | kanzure: doing the wrong thing faster is a moral hazard | 22:37 |
xentrac | what I meant was that even if I replace it with Wikipedia's CSS that's still there | 22:37 |
-!- Darius [~quassel@adsl-69-231-35-194.dsl.irvnca.pacbell.net] has quit [Remote host closed the connection] | 22:37 | |
xentrac | but I don't think Wikipedia's vector skin CSS is really an improvement | 22:38 |
-!- Vutral [ss@mirbsd/special/Vutral] has quit [Ping timeout: 250 seconds] | 22:38 | |
xentrac | the <pre> text overlapping the image is indeed a problem but I think it may actually be the least bad solution on a small screen | 22:38 |
fenn | xentrac: you're missing the point. i'm not talking about just swapping CSS themes, i mean wholly reorganizing the presentation of the underlying data structure and user interface so as to be consistent across websites and even programs that are not web browsers | 22:39 |
xentrac | oh, sure | 22:40 |
fenn | like an rss reader i guess | 22:40 |
xentrac | represent your data at a higher conceptual level so that you can automatically reformat it as desired | 22:40 |
xentrac | the thing is that people have been trying that since Scribe with limited success | 22:41 |
fenn | Scribe the book typography tool? | 22:41 |
fenn | (what is scribe?) | 22:41 |
xentrac | yeah, that one. 1979: https://en.wikipedia.org/wiki/Scribe_%28markup_language%29 | 22:42 |
-!- Vutral [~ss@31.7.56.131] has joined ##hplusroadmap | 22:42 | |
fenn | i know this is confusing because a lot of people think SGML is just HTML | 22:42 |
fenn | HTML was sort of worse-is-better | 22:43 |
-!- Vutral [~ss@31.7.56.131] has quit [Changing host] | 22:43 | |
-!- Vutral [~ss@mirbsd/special/Vutral] has joined ##hplusroadmap | 22:43 | |
xentrac | I know | 22:44 |
fenn | anyway the difference is that now we have an entire internet's worth of data and hacky webpage-as-UI that needs to get steamrollered into shape. i figure a bayesian classifier is a reasonable solution | 22:44 |
Lemminkainen | paperbot http://www.nature.com/nri/journal/v13/n3/pdf/nri3399.pdf | 22:45 |
paperbot | http://libgen.org/scimag/get.php?doi=10.1038%2Fnri3399 | 22:45 |
Lemminkainen | hellllllllyuss | 22:45 |
-!- marciogm [~marciogm@179.126.69.68] has quit [Quit: Computer has gone to sleep.] | 22:46 | |
fenn | then you have different pre-compiled presentation modules like "image gallery" or "long form text" or "comment thread" or "drop down menu" that all are end-user configurable | 22:46 |
xentrac | http://cr.yp.to/writing/visual.html describes how far we've come (as of 2007) with "The author should be concerned with the structure, not any particular visual representation." | 22:46 |
xentrac | after 28 years of work, success is still limited | 22:46 |
fenn | the user agent in this case can fetch pages that it thinks are relevant to the display, like say someone links you to a flickr photo in a set, it would fetch the thumbnails from the rest of the set and show you image gallery mode | 22:47 |
xentrac | so I think the problem may be harder than you think | 22:47 |
kanzure | "it's only the sum of the number of parts" | 22:48 |
kanzure | :grumpy: | 22:48 |
fenn | in this case i'm processing relatively unstructured data (arbitrary webpages) and attempting to transform it to the ideal structured data | 22:48 |
fenn | so it's a different problem than "trying to get everyone to use your nerdball view of the world" | 22:49 |
fenn | ontological assimilation | 22:50 |
fenn | i just know this is an irc log i am going to find during some future web search | 22:51 |
kanzure | hello future me, fuck you | 22:51 |
xentrac | well, I'm just saying that even if web page authors had taken the time to provide you the structured data in the first place, current software will still do a much worse job of presenting it as a two-dimensional interactive display | 22:52 |
xentrac | than the current mishmash of skilled and unskilled webmasters tweaking CSS | 22:52 |
kanzure | fenn's idea is to just use a default minimum terrible level of "two-dimensional display" | 22:52 |
fenn | bullshit, the webmasters have no idea what my requirements are and even if they did, they wouldn't care | 22:52 |
kanzure | then you don't have to be disappointed anymore | 22:52 |
fenn | for example the black text on white background sucks when you are in a dark room | 22:53 |
xentrac | like the readability bookmarklet, plus some classifiers to apply more specialized presentations? | 22:53 |
kanzure | not that i agree, i just want software and package managers | 22:53 |
xentrac | yeah, it does, fenn | 22:53 |
fenn | white text on a black background sucks when you're using an e-ink reader | 22:53 |
xentrac | I prefer red text on a black background most of the time | 22:53 |
fenn | this is what CSS was supposed to solve, but instead we've gone and turned it into a do-anything rendering API | 22:53 |
xentrac | yeah. and above I explained why I think that is | 22:54 |
fenn | i spent several days tweaking tumblr to get it all as light text on dark background; absolutely boggling how many things needed to be fiddled with to get there | 22:55 |
xentrac | and in the case of Wikipedia you need to rerender the equations | 22:55 |
xentrac | which btw you can do because the TeX source is present as alt text | 22:55 |
fenn | nah you can just do simple processing like "this has a white border and is transparent, invert the image" | 22:55 |
xentrac | <img src="//upload.wikimedia.org/math/9/4/5/945a66bb8ac5a46fd959ab6c12eebb00.png" alt="\Delta v = v_\text{e} \ln \frac {m_0} {m_1}" class="mwe-math-fallback-png-inline tex"> | 22:56 |
xentrac | do you mean a white background? (in the original page CSS?) | 22:56 |
fenn | nevermind, re-rendering the equation is probably the best general solution | 22:57 |
fenn | there will be cases when the TeX is not available, such as scanned images or "we converted this svg to a jpg for you" | 22:57 |
xentrac | it's not very general; it only applies to MediaWiki with TeX | 22:58 |
fenn | but consider what the equation represents, a mathematical function with symbols representing constants or parameters | 22:58 |
xentrac | hmm, it would be interesting if you could automatically do something interesting with that | 22:59 |
-!- HashNuke [uid12117@gateway/web/irccloud.com/x-ybszwnvfjexstzpx] has joined ##hplusroadmap | 22:59 | |
fenn | it's not too far-fetched | 22:59 |
fenn | really any text should be OCR'ed | 23:00 |
xentrac | like Project Naphtha | 23:00 |
fenn | even text in images that aren't clearly computer generated | 23:00 |
kanzure | ugh now my queries are returning me stuff written by matt campbell | 23:01 |
kanzure | well i have clearly looped back on myself | 23:01 |
kanzure | maybe i should ask him | 23:02 |
kanzure | conceptual design flaws seems like a thing he might think about | 23:02 |
xentrac | fenn: maybe you could build a search engine that returns you equations scraped from Wikipedia pages | 23:02 |
fenn | most of campbell's stuff was about "conceptual design" anyway | 23:02 |
xentrac | and integrate it into a computer algebra system | 23:02 |
kanzure | are you trolling him? i can't tell | 23:03 |
fenn | xentrac: actually i hate equations | 23:03 |
fenn | they're supposed to look pretty and be compact or something | 23:03 |
xentrac | no, not at all; I think this might actually be useful | 23:03 |
fenn | but there's a reason japanese computer programmers don't use kanji in their code... | 23:04 |
xentrac | case in point, yesterday my ex-girlfriend and I were 3-D printing a Helmholtz resonator for a music box | 23:04 |
xentrac | https://en.wikipedia.org/wiki/Helmholtz_resonator has some relevant equations and variable definitions | 23:04 |
fenn | what is "project naphtha" all i see is petrochemical stuff | 23:05 |
xentrac | sorry, they misspelled the name | 23:05 |
xentrac | .g project naptha | 23:05 |
yoleaux | http://projectnaptha.com/ | 23:05 |
xentrac | .t | 23:05 |
yoleaux | Thu, 24 Apr 2014 06:05:47 UTC | 23:05 |
xentrac | ...¿en serioooo? | 23:06 |
kanzure | weren't they in the news yesterday | 23:06 |
kanzure | their name screams "hey vegeta" to me | 23:06 |
fenn | napa = cabbage | 23:06 |
xentrac | anyway it turns out all the equations on that page use compatible naming of variables | 23:06 |
kanzure | .title https://www.youtube.com/watch?v=qslAuezumIA | 23:06 |
yoleaux | DBZ Abridged: The Best of Nappa | 23:06 |
-!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 23:06 | |
xentrac | and so if you pulled that into a computer algebra system and set some of the variables to known values, you could get a thing that let you graph how the others are functions of each other | 23:07 |
xentrac | turns out that a really crucial missing piece there is the Q factor of the resonator, which ought to be on that page but isn't | 23:08 |
fenn | oh btw xentrac in my-very-first-raytracer.html the font kerning is all off in the google chrome version (maybe because it's trying to substitute some low quality font i have installed) | 23:08 |
xentrac | fenn: sorry to hear that | 23:09 |
xentrac | could also be because I'm specifying that the text be justified, but browsers still don't hyphenate, which puts a lot of heavy demand on adjusting letter and word spacing | 23:10 |
fenn | oh right i remember dillo introduced all this fancy text hyphenation stuff | 23:11 |
fenn | and netsurf just left-justifies it | 23:12 |
fenn | man wikipedia's equations really look like crap even with a white background | 23:13 |
fenn | must be optimized for speed instead of quality | 23:13 |
kanzure | it's all cached anyway so i doubt they are bothered by an extra 10-100ms during rendering | 23:14 |
fenn | what's with all the visible pixels then | 23:15 |
fenn | (aliasing) | 23:15 |
-!- Vutral_ [~ss@vutral.net] has joined ##hplusroadmap | 23:17 | |
fenn | xentrac: as you've no doubt deduced the same system could be applied to PDFs | 23:19 |
kanzure | figure/diagram extraction is a pain in the butt, much less OCR.. | 23:19 |
fenn | you don't have to OCR the pdf because it's already text for the most part | 23:20 |
kanzure | i wish | 23:20 |
fenn | i've resorted to OCR'ing pdf files though | 23:20 |
fenn | this is the kind of problem that really shouldn't exist | 23:21 |
kanzure | it's not like i want to read a million pages of longform text to get effective at anything anyway | 23:21 |
kanzure | that's just not sustainable | 23:21 |
fenn | huh | 23:21 |
kanzure | i'm being grumpy about reading, this week | 23:22 |
fenn | you should be able to sort a table in a pdf by its headings | 23:22 |
xentrac | fenn: I don't see aliasing on Wikipedia equations on a white background | 23:22 |
kanzure | figshare does some of that but they are a company and i don't think they're giving away source/software | 23:22 |
fenn | and summarize columns, or whatever | 23:22 |
xentrac | I do see big pixels but that's because they're images and my zoom is >100% | 23:22 |
fenn | you dont see the huge pixels on the radical sign? https://upload.wikimedia.org/math/1/5/8/158e4d2f3b3610849e24263720118db6.png | 23:23 |
xentrac | fenn: I posted a bookmarklet for doing that on HTML to kragen-hacks a few years back | 23:23 |
xentrac | have you tried it? | 23:23 |
xentrac | oh, yeah, that's a little jaggy. wonder why! | 23:24 |
fenn | a bookmarklet will let me sort pdf tables? | 23:24 |
kanzure | /topic mathematical aliasing criticism group | 23:24 |
xentrac | no, "on HTML" | 23:24 |
fenn | i'm of the school of thought that equations are about as dumb as cursive or astrological diagrams | 23:25 |
xentrac | looking at the equation in the GIMP, it seems to be antialiased wrongly | 23:25 |
xentrac | I think all the pixels should be black, with varying levels of opacity | 23:25 |
fenn | is it a png8 or a png24? | 23:26 |
fenn | png8 has boolean transparency i think | 23:26 |
xentrac | instead they are many different shades of gray, with only boolean transparency | 23:27 |
xmj | 2^8 shades of gray | 23:27 |
xentrac | "4-bit colormap" | 23:27 |
xmj | :o | 23:27 |
fenn | you'd think there would be a boolean luminance channel png8 with 2^8 shades of transparency | 23:27 |
fenn | i mean this is a pretty common scenario isn't it? | 23:28 |
fenn | take a black and white image, scale it down, share it on the net | 23:28 |
fenn | the tyranny of diurnalists reaches into the furthest corners of our lives! | 23:29 |
xentrac | it seems like the cleanest solution to that is to colormap the rendered image of the web page | 23:30 |
xentrac | black to red, white to black, presumably red to white, and other corners of the color cube to themselves? | 23:31 |
xentrac | anyway, I find equations useful because they describe relationships between things | 23:31 |
xentrac | have you considered trying to describe e.g. population standard deviation without a formula? | 23:32 |
fenn | i think we should use either code or pseudocode | 23:32 |
xentrac | it's the square root of the sum of the squared differences between each data value and the average of all the data values | 23:32 |
xentrac | code is a lot less expressive than equations | 23:32 |
fenn | people say that but i don't believe it | 23:33 |
fenn | equations just leave out all the important details and that fucks you up if you aren't already a domain expert | 23:33 |
xentrac | maybe that's not exactly right | 23:33 |
xentrac | it's not that code is less expressive, per se | 23:33 |
fenn | and if you're a domain expert you don't need to be reminded of the equation | 23:33 |
xentrac | it's that it is too powerful | 23:34 |
xentrac | according to the principle of least power, when a language loses power, it gains manipulability | 23:34 |
xentrac | and that is certainly true of equations | 23:34 |
fenn | does not compute. power = expressiveness (from my reading of paul graham at least) | 23:34 |
xmj | ln(f(x)) makes it certainly lose power :( | 23:35 |
xentrac | haha | 23:35 |
xentrac | are you familiar with the principle of least power? | 23:35 |
fenn | no | 23:35 |
xentrac | .g principle of least power | 23:35 |
yoleaux | http://www.w3.org/DesignIssues/Principles.html | 23:35 |
fenn | i'm guessing it's the difference between source code and data format | 23:36 |
xentrac | yes, and that's a continuum | 23:36 |
xentrac | something that's "data" from one perspective is "code" from another | 23:36 |
-!- Adifex is now known as Adifex|zzz | 23:36 | |
xentrac | the equation whose rendering we are criticizing here tells us how to compute the necessary volume of the resonator to achieve a given frequency, given a fixed static pressure, neck size and shape (A and Vₙ), and gas (which determines γ and m) | 23:39 |
xentrac | it also tells us how to compute the necessary neck length if instead the volume and neck diameter and gas are fixed | 23:39 |
xentrac | or how to compute the resonant frequency if all those things are fixed | 23:39 |
fenn | i'll just note that omega is never defined as frequency anywhere on the page | 23:40 |
xentrac | no, it's angular frequency, not frequency | 23:40 |
xentrac | and it is in fact defined right up at the top: | 23:41 |
xentrac | It can be shown[2] that the resonant angular frequency is given by: | 23:41 |
xentrac | \omega_{H} = \sqrt{\gamma\frac{A^2}{m} \frac{P_0}{V_0}} (rad/s), | 23:41 |
fenn | how do you know that omega_H is angular frequency? | 23:41 |
xentrac | because it says right there | 23:42 |
fenn | no it doesn't | 23:42 |
xentrac | "the resonant angular frequency is given by: \omega_{H} =" | 23:42 |
fenn | it says "angular frequency is given by: (equation)" | 23:42 |
xentrac | it also allows us to compute the ambient pressure given a fixed resonator and gas, if we can measure the frequency | 23:42 |
xentrac | so that single equation gives us four different useful computations, which would need four different pieces of code to compute them | 23:43 |
fenn | now i know you've heard of prolog | 23:43 |
xentrac | yes. | 23:43 |
fenn | "prolog does that" | 23:44 |
xentrac | you would think you could do that kind of manipulation with prolog but in practice you cannot. | 23:44 |
xentrac | because prolog gave up that level of manipulability in order to be turing-complete. | 23:44 |
xentrac | very early arithmetic prolog programs from the 1970s can in fact be manipulated in this way. | 23:44 |
fenn | okay but does this mean we have to use funky hieroglyphics i can't even type with a keyboard or even a stream of characters? | 23:44 |
xentrac | no, of course not | 23:45 |
xentrac | you might, in a Turing-complete language, write a piece of code that can easily be manipulated in the same way | 23:45 |
xentrac | but as soon as you start sticking cuts or while-loops into your code, that kind of manipulation becomes very difficult | 23:46 |
fenn | it just seems like a case of lazy evaluation to me | 23:46 |
xentrac | ("cuts" are a Prolog thing) | 23:46 |
jrayhawk | hv3 does some javascript stuff | 23:46 |
xentrac | it seems like it, but it's actually a different phenomenon. | 23:46 |
fenn | don't invert the matrix until you need to, etc. | 23:46 |
xentrac | if you combine this equation with the equation of Q from https://en.wikibooks.org/wiki/Acoustics/Flow-induced_Oscillations_of_a_Helmholtz_Resonator | 23:48 |
xentrac | which is | 23:48 |
xentrac | Q = 2 \pi \sqrt{V (\frac{L'} {S})^3} | 23:48 |
fenn | jrayhawk: do you compile your own hv3? "ECMAScript (a.k.a. javascript) is not supported since the corresponding library (SEE) is [not?] installed (not available in Debian yet). | 23:48 |
xentrac | (which requires some variable renaming unfortunately) | 23:48 |
jrayhawk | No, I don't use hv3. | 23:48 |
jrayhawk | Automatic execution of unsigned code is a bug, not a feature. | 23:48 |
fenn | oh this was the tk thing. yeah, no. | 23:49 |
xentrac | then you can specify that you want a given Q and resonant frequency and gas and ambient pressure, and solve the system of two equations to give you a relationship with one degree of freedom between the length of the neck, the diameter of the neck, and the volume of the vessel | 23:50 |
fenn | xentrac: obviously non-functional (as in functional programming) code can't be treated the same as functional code | 23:50 |
xentrac | even Turing-complete functional code is not generally manipulable in this way | 23:51 |
xentrac | Turner is exploring the limits of non-Turing-complete functional code, which might be promising in this way | 23:51 |
fenn | what's the problem with a while loop exactly? | 23:52 |
xentrac | the biggest difficulty is that it's uncomputable whether it terminates or not, and if so, when | 23:54 |
fenn | how is this different from equations with no solution? | 23:55 |
xentrac | consider the Ackermann function | 23:55 |
fenn | (possibly) no solution | 23:55 |
xentrac | which is actually just a nested for loop | 23:57 |
fenn | the article on ackermann function starts out with a lot of phrases that mean nothing to me, so i'm going to call it a night :P | 23:57 |
xentrac | not even a while loop; it's provably terminating | 23:57 |
fenn | it looks like the factorial function | 23:58 |
xentrac | it does addition, multiplication, exponentiation, and so on | 23:59 |
--- Log closed Thu Apr 24 00:00:10 2014 |
Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!