--- Log opened Fri Mar 29 00:00:41 2013 | ||
@kanzure | oh look | 00:09 |
---|---|---|
@kanzure | there's a prusa3 | 00:09 |
@kanzure | https://github.com/josefprusa/Prusa3 | 00:09 |
@kanzure | and it has thingdoc documentation | 00:09 |
@kanzure | https://github.com/josefprusa/Prusa3/blob/master/box_frame/common.tdoc | 00:09 |
@kanzure | /box_frame/inc looks like a primitive library system :\ | 00:10 |
-!- sivoais [~zaki@unaffiliated/sivoais] has quit [Quit: leaving] | 00:40 | |
-!- sivoais [~zaki@unaffiliated/sivoais] has joined ##hplusroadmap | 00:41 | |
-!- BioGuy [~BioGuy@184-76-124-69.war.clearwire-wmx.net] has joined ##hplusroadmap | 01:39 | |
BioGuy | Anyone alive at this hour? | 01:40 |
-!- plur [~nonentity@121-73-87-49.cable.telstraclear.net] has quit [Remote host closed the connection] | 01:53 | |
-!- qu-bit [~shroedngr@unaffiliated/barriers] has quit [Ping timeout: 252 seconds] | 01:57 | |
-!- cpopell [~cpopell@pool-96-231-37-73.washdc.fios.verizon.net] has quit [Ping timeout: 258 seconds] | 02:23 | |
-!- BioGuy [~BioGuy@184-76-124-69.war.clearwire-wmx.net] has quit [Remote host closed the connection] | 02:44 | |
archels | only the Euros and kanzure | 03:13 |
-!- makoLime [~mako@103-9-42-31.flip.co.nz] has quit [Ping timeout: 255 seconds] | 03:46 | |
-!- zdap [~not@bas5-kingston08-1242447600.dsl.bell.ca] has quit [Read error: Connection reset by peer] | 03:51 | |
-!- zdap [~not@bas5-kingston08-1242447600.dsl.bell.ca] has joined ##hplusroadmap | 03:52 | |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 04:19 | |
-!- augur [~augur@208.58.5.87] has quit [Remote host closed the connection] | 04:25 | |
-!- augur [~augur@208.58.5.87] has joined ##hplusroadmap | 04:25 | |
-!- augur [~augur@208.58.5.87] has quit [Ping timeout: 264 seconds] | 04:29 | |
-!- ThomasEgi [~thomas@185.5.8.81] has joined ##hplusroadmap | 04:49 | |
-!- ThomasEgi [~thomas@185.5.8.81] has quit [Changing host] | 04:49 | |
-!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap | 04:49 | |
-!- augur [~augur@129-2-129-34.wireless.umd.edu] has joined ##hplusroadmap | 04:56 | |
-!- sivoais [~zaki@unaffiliated/sivoais] has quit [Read error: Connection reset by peer] | 05:02 | |
-!- sivoais [~zaki@unaffiliated/sivoais] has joined ##hplusroadmap | 05:03 | |
-!- EnLilaSko [EnLilaSko@unaffiliated/enlilasko] has joined ##hplusroadmap | 05:03 | |
-!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 264 seconds] | 06:15 | |
-!- yorick [~yorick@ip51cd0513.speed.planet.nl] has joined ##hplusroadmap | 06:47 | |
-!- yorick [~yorick@ip51cd0513.speed.planet.nl] has quit [Changing host] | 06:47 | |
-!- yorick [~yorick@oftn/member/yorick] has joined ##hplusroadmap | 06:47 | |
-!- diginet [~kevin@107-213-84-107.lightspeed.hstntx.sbcglobal.net] has joined ##hplusroadmap | 07:30 | |
diginet | I'm not interested in it for biological applications, but what is there in the way of DIY microfluidics | 07:31 |
superkuh | paperbot: http://www.nature.com/nature/journal/v495/n7442/full/nature11990.html | 08:02 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/X-ray%20analysis%20on%20the%20nanogram%20to%20microgram%20scale%20using%20porous%20complexes.pdf | 08:03 |
@ParahSailin | diginet: not much | 08:03 |
diginet | ParahSailin: couldn't you just use standard lithography techniques but on glass? (I guess HF as the etchant) | 08:04 |
diginet | not that litho is "easy," but it's not impossible, and we're not talking nanoscale here | 08:04 |
@ParahSailin | for glass you wouldnt even have to go to hf unless you were really in a hurry | 08:08 |
diginet | what else could you use? | 08:09 |
diginet | (for wet etching anyhow, not really interested in plasma etching) | 08:09 |
@ParahSailin | tmah could probably do it in a reasonable time, regular ammonia would do it a bit slower | 08:10 |
diginet | interesting | 08:10 |
@ParahSailin | strong acids will etch glass too | 08:10 |
@ParahSailin | just look up the etch rate you want to see which is most acceptable | 08:11 |
diginet | like which ones? I thought HF was the only common acid that etched glass | 08:11 |
diginet | although I know hot phosphoric acid will | 08:11 |
diginet | what about for the resist? PMMA has the advantage of being cheap, but IIRC it needs deep UV to be developed (or e-beam but that's obviously out of the picture) | 08:11 |
@ParahSailin | its a question of rate, a lot of stuff will etch glass, but HF is the fastest | 08:14 |
diginet | I see | 08:14 |
@ParahSailin | HF is probably the most practical | 08:16 |
diginet | well HF does have the advantage of not attacking plastics (in general), I've heard of people making DIY photopolymer plates | 08:19 |
diginet | maybe I could use those as a resist | 08:20 |
@ParahSailin | photoresist works | 08:21 |
diginet | like SU-8? isn't that really pricey though? | 08:22 |
@ParahSailin | the regular novolac kind | 08:24 |
diginet | where do you find that? | 08:24 |
diginet | I've been looking for novolac for ages (for other stuff) | 08:25 |
diginet | novolac, like bakelite right? | 08:28 |
@ParahSailin | what, no, like AZ[0-9]* | 08:29 |
diginet | bakelite was a phenolic resin, as I understand it so is novolac, right? | 08:30 |
@ParahSailin | yes | 08:47 |
@kanzure | https://github.com/SpiderLabs/BurpNotesExtension | 08:55 |
@kanzure | https://github.com/aglane for a UPnP browser soonish | 08:55 |
@kanzure | opencascade 6.5.5 has been released | 08:58 |
@ParahSailin | paperbot: http://pubs.acs.org/doi/abs/10.1021/bi9809425 | 09:14 |
paperbot | error: HTTP 500 http://diyhpl.us/~bryan/papers2/paperbot/Thermodynamic%20Parameters%20for%20an%20Expanded%20Nearest-Neighbor%20Model%20for%20Formation%20of%20RNA%20Duplexes%20with%20WatsonCrick%20Base%20Pairs.pdf | 09:14 |
@ParahSailin | i wonder if idt has done secret experiments more recently to make their oligoanalyzer more accurate | 09:15 |
-!- diginet [~kevin@107-213-84-107.lightspeed.hstntx.sbcglobal.net] has left ##hplusroadmap ["Leaving"] | 09:16 | |
@kanzure | maradydd would be the person to talk with about that (since she worked at idt for a while and then repented to diybio) | 09:16 |
@kanzure | paperbot: http://www.nature.com/news/taking-the-crystals-out-of-x-ray-crystallography-1.12699 | 09:18 |
paperbot | no translator available, raw dump: http://diyhpl.us/~bryan/papers2/paperbot/788fffaa32f2b8b8118aa1d28c521831.txt | 09:18 |
@kanzure | wait.. /news :( | 09:18 |
@kanzure | paperbot: http://www.nature.com/nature/journal/v495/n7442/full/nature11990.html | 09:18 |
paperbot | http://diyhpl.us/~bryan/papers2/paperbot/X-ray%20analysis%20on%20the%20nanogram%20to%20microgram%20scale%20using%20porous%20complexes.pdf | 09:18 |
@kanzure | "We demonstrate that as little as 80 nanograms of a sample is enough for X-ray single-crystal diffraction analysis." | 09:20 |
-!- jmil [~jmil@hive76/member/jmil] has quit [Quit: jmil] | 09:24 | |
@ParahSailin | wow thats pretty big | 09:36 |
@ParahSailin | oh, they only did small molecules | 09:38 |
superkuh | There is talk about proteins too. | 09:39 |
superkuh | http://www.sciencemag.org/content/336/6084/1018.full | 09:39 |
superkuh | paperbot: http://www.sciencemag.org/content/336/6084/1018.full | 09:39 |
paperbot | error: HTTP 300 http://diyhpl.us/~bryan/papers2/paperbot/Large-Pore%20Apertures%20in%20a%20Series%20of%20Metal-Organic%20Frameworks.pdf | 09:39 |
@ParahSailin | well there is no mention of peptide or protein in that nature one | 09:40 |
* superkuh nods. | 09:41 | |
@ParahSailin | paperbot: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1277807/ | 09:45 |
paperbot | no translator available, raw dump: http://diyhpl.us/~bryan/papers2/paperbot/b5f68ff1cd1f5306a7f3a0aab50f9fa3.txt | 09:45 |
@ParahSailin | paperbot: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1277807/pdf/gki918.pdf | 09:45 |
paperbot | no translator available, raw dump: http://diyhpl.us/~bryan/papers2/paperbot/7bcc57d4c29ce24a4b3d856414ae0089.pdf | 09:45 |
-!- jmil [~jmil@hive76/member/jmil] has joined ##hplusroadmap | 10:10 | |
-!- augur [~augur@129-2-129-34.wireless.umd.edu] has quit [Remote host closed the connection] | 10:13 | |
@ParahSailin | paperbot: http://pubs.acs.org/doi/abs/10.1021/bi034621r | 11:03 |
paperbot | error: HTTP 500 http://diyhpl.us/~bryan/papers2/paperbot/Effects%20of%20Sodium%20Ions%20on%20DNA%20Duplex%20Oligomers%3A%20Improved%20Predictions%20of%20Melting%20Temperatures.pdf | 11:03 |
@kanzure | wasn't there also a thing about hairpins increasing the melting temperature | 11:04 |
@ParahSailin | im not sure how that would affect it if you have the exact complement to duplex with | 11:13 |
@ParahSailin | should have some sort of effect | 11:14 |
@ParahSailin | this stuff is actually me using paperbot for work related stuff | 11:15 |
@kanzure | work is too cheap to buy subscriptions? | 11:17 |
-!- Mariu [Jimmy98@89.41.57.33] has joined ##hplusroadmap | 11:38 | |
@ParahSailin | paperbot: http://pubs.acs.org/doi/abs/10.1021/bi962590c | 11:56 |
paperbot | error: HTTP 500 http://diyhpl.us/~bryan/papers2/paperbot/Thermodynamics%20and%20NMR%20of%20Internal%20GT%20Mismatches%20in%20DNA.pdf | 11:56 |
@ParahSailin | i couldnt condone paying these crooks on principle | 11:59 |
-!- augur [~augur@208.58.5.87] has joined ##hplusroadmap | 12:09 | |
-!- undersco2 [~quassel@192.210.211.75] has quit [Ping timeout: 272 seconds] | 12:18 | |
-!- underscor [~quassel@192.210.211.75] has joined ##hplusroadmap | 12:22 | |
-!- underscor is now known as Guest17500 | 12:22 | |
-!- makoLime [~mako@103-9-42-31.flip.co.nz] has joined ##hplusroadmap | 12:23 | |
-!- jmil_ [~jmil@hive76/member/jmil] has joined ##hplusroadmap | 12:26 | |
-!- jmil [~jmil@hive76/member/jmil] has quit [Ping timeout: 246 seconds] | 12:28 | |
-!- jmil_ is now known as jmil | 12:28 | |
-!- OldCoder_ [~OldCoder_@209.237.22.146] has joined ##hplusroadmap | 12:33 | |
-!- plur [~nonentity@121-73-87-49.cable.telstraclear.net] has joined ##hplusroadmap | 12:50 | |
-!- plur [~nonentity@121-73-87-49.cable.telstraclear.net] has quit [Ping timeout: 245 seconds] | 13:10 | |
-!- plur [~nonentity@121-73-87-49.cable.telstraclear.net] has joined ##hplusroadmap | 13:21 | |
@ParahSailin | paperbot: http://pubs.acs.org/doi/abs/10.1021/bi702363u | 13:31 |
paperbot | error: HTTP 500 http://diyhpl.us/~bryan/papers2/paperbot/Predicting%20Stability%20of%20DNA%20Duplexes%20in%20Solutions%20Containing%20Magnesium%20and%20Monovalent%20Cations.pdf | 13:31 |
-!- jmil [~jmil@hive76/member/jmil] has quit [Quit: jmil] | 13:40 | |
@kanzure | bah ipython has closed my pull request | 14:01 |
@kanzure | https://github.com/ipython/ipython/issues/1286 | 14:01 |
@kanzure | https://github.com/ipython/ipython/issues/2895 | 14:01 |
-!- indigenous [~quassel@94-76-243-32.static.as29550.net] has quit [Changing host] | 14:37 | |
-!- indigenous [~quassel@pdpc/supporter/student/indigenous] has joined ##hplusroadmap | 14:37 | |
-!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 14:45 | |
@kanzure | "East Bay DIYbio meeting 6pm at Sudoroom" | 14:53 |
@kanzure | win 5 | 14:53 |
@kanzure | fidjklada | 14:53 |
-!- qu-bit [~shroedngr@unaffiliated/barriers] has joined ##hplusroadmap | 15:19 | |
@kanzure | http://news.yahoo.com/blogs/this-could-be-big-abc-news/teen-scientist-amazing-breakthrough-her-home-lab-165831291.html | 15:38 |
@kanzure | "In her research she developed a process of artificial selection where she killed off algae with low levels of acetyl-CoA carboxylase (ACCase), an enzyme crucial to lipid synthesis. Left behind in the beakers under her bed were an efficient bunch of high yield algae, which she discovered produced a significant increase in lipid production." | 15:38 |
@kanzure | i wonder what selection procedure this was | 15:38 |
-!- klafka [~klafka@204.101.190.178] has joined ##hplusroadmap | 15:58 | |
@kanzure | "Dear Colleagues, An assertion such as below is ridiculous. Pharmaceutical companies would not spend the $1,000,000,000 for research it takes to bring each product to market if it were not for the brief period of market exclusivity that the patent system provides. Signed, Research Professor of Chemistry." | 16:00 |
@kanzure | blah i can't remember the right links that debunk this | 16:00 |
-!- cpopell [~cpopell@pool-96-231-37-73.washdc.fios.verizon.net] has joined ##hplusroadmap | 16:01 | |
-!- jonathanc [~jonathanc@wsip-98-174-141-194.sd.sd.cox.net] has joined ##hplusroadmap | 16:07 | |
jonathanc | buy more guns | 16:07 |
chris_99 | ? | 16:09 |
@fenn | our sponsor, the NRA | 16:15 |
@kanzure | jonathanc: hello | 16:17 |
@kanzure | jonathanc: welcome back | 16:17 |
@kanzure | also i think your yaml>xml stance is respectable | 16:17 |
jonathanc | gcode vs. ? | 16:22 |
@kanzure | context? | 16:29 |
jonathanc | ""Cool presentation and I want one of those things to make an incubator | 16:36 |
jonathanc | for the repstrap I am building with Tecan/Biomek like attributes, I | 16:36 |
jonathanc | got a ... blah blah .. Anyway the repstrap firmware, I am planning on using Marlin which | 16:36 |
jonathanc | processes Gcode. I guess at first I will have the gcode thing going on | 16:36 |
jonathanc | for position control of steppers and some way to send commands ...."" | 16:36 |
@kanzure | ah, well at least he's not writing his own gcode firmware | 16:36 |
@kanzure | i think you can wire up linuxcnc directly to your actuators without a gcode intermediate | 16:38 |
jonathanc | i could just implement multiple command sets on my device. auto-detect g-code and then switch to that, easy. | 16:39 |
@kanzure | i don't think re-implementing gcode every time one of us makes anything that moves is a good use of our time | 16:39 |
jonathanc | meh, that only takes a day | 16:40 |
@kanzure | but then you end up with lots of implementations that have poor edge case handling | 16:41 |
@kanzure | what's wrong with repstrap? i haven't actually used it. the last i checked the main problem was that it was a java monstrosity? that might have been another reprap firmware though.. | 16:41 |
jonathanc | Seems to me that gcode is pretty simple as an API | 16:45 |
@kanzure | fenn: ping? | 16:45 |
jonathanc | simple/dumb | 16:46 |
@kanzure | oh wow linuxcnc is unable to find a grammar for gcode | 16:47 |
@kanzure | http://www.mailinglistarchive.com/html/emc-users@lists.sourceforge.net/2012-02/msg00984.html | 16:47 |
@kanzure | there are "extensions" that must be allowed in real-time? what is this. | 16:47 |
@kanzure | fenn is our resident linuxcnc developer and gcode user so i imagine he might know more specifics than i do | 16:48 |
jonathanc | is this what happens when meche's attempt to program? | 16:48 |
@kanzure | well afaik gcode was never really standardized with a grammar and it just "sort of happened" | 16:49 |
jonathanc | hence my curiousity regarding alternatives to gcode | 16:51 |
jonathanc | although any complexity pales in comparison to hacking proprietary workarounds into any networking protocol so... it really is not that complex of a problem at all, it is no big deal | 16:52 |
@fenn | jonathanc: have you considered stealing large chunks from grbl? | 16:53 |
@fenn | of code* | 16:53 |
@kanzure | also if you really do have that file it would be nice to put it somewhere on the internets | 16:54 |
@fenn | kanzure: the reason for lack of standardization is vendor lock-in | 16:54 |
@fenn | there's a perfectly good standard published by nist (RS274-NGC) but nobody implements it | 16:54 |
jonathanc | well i'd never re-use code written by someone for arduino but might use it as documentation so that's a good pointer to have, will look at grbl | 16:55 |
jonathanc | I assume devices do not have to implement all g-code commands if they do not support them anyway | 16:56 |
@fenn | here's the closest you'll get to a grammar for g-code: http://fennetic.net/irc/rs274ngc.ast.g although it's not particularly enlightening (g-code doesn't have much grammar to begin with) | 16:57 |
jonathanc | simply, gimme a break, these commands are just turning on a motor or turning it off... if the complex coordinate stuff is ignored | 16:57 |
@fenn | oh. in context http://fennetic.net/irc/emc3/src/emc/interpreter/rs274ngc/rs274ngc.ast.g | 16:58 |
jonathanc | excellent | 16:59 |
@fenn | i'm not sure if that's actually correct; it was in the early stages of development | 16:59 |
jonathanc | ok, i'll use it as a cross reference | 16:59 |
@fenn | i'm sort of confused what the goal is here. are you going to be running an embedded g-code interpreter on the robot? if so why not just use grbl running on an arduino? | 17:01 |
@kanzure | this doesn't seem to explain all the "gcode extensions" stuff. it seems to also support functions and it has to have a vm? | 17:01 |
jonathanc | because arduino is a toy? | 17:02 |
@fenn | well, any avr | 17:02 |
@fenn | if it's against your religion, i guess that's an answer | 17:02 |
jonathanc | well from superficial reading of gcode on wikipedia, gcode does allow registers (variables) to be set, so if you want to call that a vm, so be it | 17:02 |
@kanzure | "remap gcodes half way through a swarf-making job" | 17:03 |
@kanzure | "the language we have *is already runtime extensible* at the word level" | 17:03 |
jonathanc | g-code is just an api | 17:03 |
-!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 264 seconds] | 17:03 | |
@kanzure | http://www.linuxcnc.org/docs/devel/html/remap/structure.html#_introduction_extending_the_rs274ngc_interpreter_by_remapping_codes | 17:04 |
@kanzure | uhh "Embedded Python functions in the Interpreter started out as glue code, but turned out very useful well beyond that." | 17:04 |
@fenn | kanzure: various people in the emc project have tried to put lipstick on the pig, but custom functions and subroutines isn't generally considered part of the language | 17:04 |
@kanzure | is python the lipstick or the pig? | 17:04 |
@fenn | g-code is the pig, fancy programming tricks is the lipstick | 17:04 |
@kanzure | also, it would be nice to have something that is not subject to vendor-lockin-guided-evolution | 17:05 |
@kanzure | so even if you want to use gcode, why not just publish a grammar, then a specific standard, and just say let's use that subset | 17:05 |
@kanzure | fenn: is the "3.3.4. Advanced example: Remapped codes in pure Python" section sane? it seems to have some python that accesses the gcode's interpreter's state. | 17:06 |
@fenn | i dont know, i've never used any of the embedded python stuff | 17:07 |
@fenn | also you keep asking for a standard, but there is a standard, published by a standards organization, with a grammar, and a reference implementation of that | 17:08 |
@fenn | so i dont know what more you want | 17:08 |
@fenn | google "nist rs274 production rules", the PDF has basically the same as what was in that antlr file i linked to earlier | 17:09 |
@kanzure | well, there's a pile of complaints about gcode that i keep hearing and it was always about incompatibility or something. and how could that happen if there's an actual standard? | 17:09 |
@fenn | because nobody follows the standard | 17:10 |
@kanzure | because there's no reference implementation? | 17:10 |
@fenn | no, because the reference implementation wasn't good enough, so people extended it, and now nobody uses the original version | 17:10 |
@fenn | that's what linuxcnc is | 17:10 |
@fenn | then you have all the other vendors, whose names i forget now | 17:11 |
@fenn | anyway the core stuff like G0 G1 G2 is almost always the same | 17:12 |
@fenn | some machines require fixed decimal point input, or no spaces, or must have spaces, or capital letters, or only parameters on their own line, and various stupid rules like that | 17:12 |
@kanzure | because people chose not to use a parser parser? | 17:13 |
@fenn | so there's a thing called a post-processor, which takes a config file for your specific target machine and makes it work | 17:13 |
@fenn | it capitalizes the letters or adds zeroes to the end of numbers or whatever | 17:14 |
@kanzure | haha grand, i wasn't aware that there are gcode postprocessors | 17:14 |
@fenn | it's usually part of the CAM program | 17:14 |
@fenn | there are people who make a living installing these config files :\ | 17:15 |
@kanzure | i suppose there might be some embedded environments where embedding an entire parser that can work on standard input might be impractical, maybe, but parsing decimals vs integers or capitals.. that's some pretty extreme optimization? | 17:15 |
@fenn | it's not optimization, it's for compatibility with crap written 50 years ago | 17:15 |
@fenn | punch cards man | 17:16 |
@fenn | NC came first, then CNC. originally there was no computer in the loop | 17:16 |
@kanzure | i'm aware. | 17:16 |
@fenn | a "toy" arduino is a supercomputer in comparison | 17:17 |
jonathanc | well like I said, still simple compared to mucking with proprietary/standard internet protocol differences, so it is no big deal | 17:18 |
jonathanc | doesnt need lex yacc for that | 17:18 |
jonathanc | the trick would be to detect which dialect is required and then change on the fly | 17:19 |
jonathanc | by querying things on the device for example | 17:19 |
-!- gedankenstuecke [~bastian@phylomemetic-tree.de] has quit [Ping timeout: 245 seconds] | 17:19 | |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Leaving] | 17:20 | |
@fenn | how do you query a brain-dead serial protocol? | 17:20 |
jonathanc | well from my superficial reading, gcode has registers that you can read, right. so that's one way. | 17:21 |
-!- gedankenstuecke [~bastian@phylomemetic-tree.de] has joined ##hplusroadmap | 17:21 | |
jonathanc | alternatively, send it a command, if it responds with errors, then go down the list until a dialect variant succeeds | 17:22 |
jonathanc | if it's transmit only with no detected errors, then there's still tricks that can be done | 17:22 |
jonathanc | power cycle and read registers not yet initialized, see if there are characteristic values | 17:23 |
jonathanc | worst case, require the end user to select their device from a large list, which then selects the dialect | 17:23 |
jonathanc | it's no worse than any typical device driver on linux or something | 17:23 |
jonathanc | however.. if it's implemented on arduino I can see where there would be problems | 17:26 |
@kanzure | i still want a package manager for pulling in maintained libraries for microcontrollers (including arduinos) | 17:27 |
@kanzure | "oh just paste in the binary/source code once, you'll be fine" | 17:27 |
jonathanc | well you'll have to design a package manager for installing the cross compilers first | 17:28 |
jonathanc | some of which only run on windows | 17:28 |
@fenn | dpkg etc already does that | 17:28 |
jonathanc | used to be that 80% of all 8051 microcontroller code was written on keil compiler/toolchain | 17:29 |
* fenn puts his open-source zealot blinders on | 17:29 | |
jonathanc | plus those libraries wont run on arbitrary hardware | 17:29 |
jonathanc | there's 15 firmware engineers taking directions from me and guess how much reusable code they write and how many libraries are created even for in-house use? | 17:30 |
jonathanc | basically I am saying there is no such thing as a library for microcontrollers unless you want to say a math library is all you want | 17:31 |
jonathanc | there is no "application library" | 17:31 |
jonathanc | the firmware is monolithic | 17:31 |
DonnchaC | paperbot: https://www.thieme-connect.com/ejournals/abstract/10.1055/s-2004-815977 | 17:31 |
paperbot | no translator available, raw dump: http://diyhpl.us/~bryan/papers2/paperbot/96286a67a3ade429ef9d1804eadab404.txt | 17:31 |
@fenn | how is a FAT filesystem not reusable? | 17:31 |
jonathanc | why is there a FAT filesystem on a microcontroller project? | 17:32 |
jonathanc | that's pretty dumb imho | 17:32 |
@fenn | so you can read and write to an SD card? | 17:32 |
jonathanc | what are you gonna do, stick a floppy disc in your avr? | 17:32 |
@kanzure | jonathanc: opencv is very much reused everywhere in microcontrollers | 17:32 |
@kanzure | jonathanc: maybe your firmware engineers suck, by the way.. reusable code is important. | 17:33 |
@fenn | um, opencv? you must be thinking of something else | 17:33 |
@kanzure | opencv? no i'm thinking of opencv. | 17:33 |
jonathanc | in microcontrollers? no you are thinking of something else | 17:33 |
@kanzure | then what are all these embedded projects doing with opencv | 17:34 |
jonathanc | show me a FAT filesystem on an 8051 and I will show you a product that didn't sell | 17:34 |
@fenn | kanzure are you thinking of this? http://www.jrobot.net/Projects/AVRcam.html | 17:34 |
jonathanc | maybe you are thinking of firmware which isn't really firmware. | 17:35 |
@fenn | also today "embedded" could mean using a cellphone, i guess | 17:35 |
jonathanc | no, a smart phone is not really embedded | 17:35 |
@fenn | well i dont know what to call it | 17:35 |
@kanzure | take qualcomm for example, their firmware surely has at least a few internal libraries.. it seems crazy that libraries are valid everywhere except one area of computing because...? because it's annoying to debug ? | 17:36 |
@fenn | jonathanc: say you have a spectrometer in your lab, how do you get data off it? | 17:36 |
jonathanc | their libraries are machine specific | 17:37 |
@fenn | so you have a custom windows driver that plugs into the machine with a serial port? and that's the best we can ever hope for? | 17:37 |
@kanzure | cross compiling is fake, don't listen to the non-believers! | 17:37 |
@kanzure | i think i messed that up | 17:38 |
-!- yashgaroth [~ffffff@cpe-66-27-118-94.san.res.rr.com] has joined ##hplusroadmap | 17:38 | |
@fenn | i'm pretty sure there are lab instruments that write to a usb key when you plug it in | 17:38 |
jonathanc | not sure where you are going with this convo | 17:38 |
@kanzure | openpcr :( | 17:38 |
@kanzure | oh wait openpcr was by sdcard | 17:38 |
jonathanc | embedded has many meanings. microcontroller is typically something very small memory, lets say 32K | 17:39 |
@fenn | well anyway, my point was there are reusable chunks of code that have nothing to do with microcontroller-specific hardware, but still need to be implemented on a microcontroller | 17:39 |
jonathanc | sure, lets say usb stack | 17:39 |
jonathanc | or bluetooth stack | 17:40 |
jonathanc | or can bus protocol for ford autos | 17:40 |
jonathanc | or i2c protocol for lm temperature chip | 17:40 |
jonathanc | or even more common | 17:40 |
jonathanc | lets say math library | 17:40 |
@fenn | sure but all of that is implemented in hardware these days | 17:40 |
jonathanc | ha wrong | 17:40 |
yashgaroth | ok I'm assuming jonathanc is jcline | 17:41 |
@kanzure | he is a very good approximation of jonathan cline | 17:41 |
jonathanc | do you work for a law enforcement agency? | 17:41 |
@fenn | you'd think he would just /nick jcline | 17:41 |
yashgaroth | I imagine there's very few jonathan C's that know of this channel and electronics | 17:41 |
yashgaroth | me, law enforcement? heh | 17:42 |
jonathanc | sure libraries exist | 17:42 |
jonathanc | but they are often machine specific, that means optimized, otherwise why "pay" for a library if it is not optimized? | 17:42 |
@kanzure | he was using FAT as a common software-implemented thing to argue for libraries i think | 17:42 |
@kanzure | well usually a compiler will help you optimize for instruction set quirks | 17:43 |
jonathanc | not compiler optimization. more like architecture-driver optimization | 17:43 |
@fenn | i think we're optimizing for coder time | 17:43 |
jonathanc | for example FAT on a microcontroller is dumb becuase why are directories needed? they are not | 17:43 |
-!- yorick [~yorick@oftn/member/yorick] has quit [Remote host closed the connection] | 17:43 | |
jonathanc | hence they are not implemented | 17:43 |
@fenn | but someone already wrote the code so why not | 17:44 |
jonathanc | yea tell me how many you sell by "optimizing for coder time" | 17:44 |
jonathanc | that's not how you sell product | 17:44 |
@fenn | who's selling anything | 17:44 |
@fenn | i thought we were making custom lab equipment | 17:44 |
jonathanc | engineering time is cheap if you're producing a million units | 17:44 |
jonathanc | nre cost vs manufacturing cost | 17:45 |
jonathanc | i thought we were making LOW COST equipment for the masses | 17:45 |
@kanzure | nah.. mostly just for me. others are welcome to do it if they like too. | 17:45 |
@kanzure | *if they would like to too | 17:45 |
-!- EnLilaSko [EnLilaSko@unaffiliated/enlilasko] has quit [Remote host closed the connection] | 17:45 | |
jonathanc | tell me why you want to optimize for coder time if your goal is to make widely-used equipment? | 17:45 |
@fenn | so, if the difference is using a $0.50 microcontroller instead of a $3 microcontroller, i'm not sure it's worth optimizing | 17:46 |
@kanzure | because i can't be bothered to write custom libraries for the 100s of machines i work with | 17:46 |
yashgaroth | jonathanc: well, assuming it is you, this is max from the carlsbad lab | 17:46 |
jonathanc | If your device must sell for $3 then it sure the hell is worth optimizing | 17:46 |
jonathanc | how are you going to sell it for $3 if it costs you $3 | 17:46 |
jonathanc | hey max | 17:47 |
@kanzure | $0.50 vs. $3 is nothing compared to the actual lab equipment costs | 17:47 |
@kanzure | s/costs/prices/ | 17:47 |
jonathanc | everything adds into FOB cost | 17:47 |
jonathanc | go ahead and choose your $3 microcontroller | 17:47 |
jonathanc | then add in the pcb area to add it in | 17:47 |
jonathanc | that's where feature creep starts | 17:48 |
@kanzure | so you believe there are *no* contexts in which libraries make sense for my microcontroller software adventures? | 17:48 |
jonathanc | how much do you want to pay for openpcr, $599 ? | 17:48 |
@fenn | what you call feature creep, i call progress | 17:48 |
jonathanc | what you call progress, I call socialism | 17:48 |
@kanzure | jonathanc: we both know that openpcr was terribly overpriced for what it was anyway | 17:50 |
jonathanc | max what are you up to? | 17:50 |
@kanzure | for its marketing, it was probably priced okay | 17:50 |
jonathanc | openpcr profit margin was what, 30%? Work backwards from there to see how it's costs added up. every piece. | 17:50 |
@kanzure | but it's no where near an example of lowest cost, compared to the other thermocyclers we've seen out of the community | 17:50 |
@kanzure | like wiremound-pcr | 17:50 |
yashgaroth | right now? not much, just got home | 17:50 |
@kanzure | jonathanc: for one, the chasis made no sense. why was it laser cut? who cares? | 17:50 |
@fenn | i liked the lightbulb in a pvc pipe thingy | 17:51 |
@fenn | how much did that cost, $10? | 17:51 |
jonathanc | wiremound-pcr, are you kidding. $85???? | 17:51 |
jonathanc | why so high??? | 17:51 |
@kanzure | jonathanc: are you trolling? | 17:51 |
jonathanc | no seriously | 17:51 |
@kanzure | jonathanc: is it so hard to believe that someone can manage 15 engineers inefficiently? | 17:52 |
@kanzure | well, the other indicator was the socialism jibe | 17:52 |
@fenn | wiremound is a heatsink with two holes drilled in it and power resistors attached? | 17:53 |
jonathanc | arduino mini | 17:53 |
@kanzure | yeah, stacey made some suspicious choices on that one | 17:54 |
@fenn | ffs is there any official documentation on this | 17:54 |
@fenn | i was confused, the heatsink is openpcr | 17:54 |
@kanzure | well, no, there was an instructables thing and then i put some of the copy-pasted source code somewhere https://github.com/kanzure/wiremound-pcr | 17:54 |
@kanzure | (i was no-ing the documentation question) | 17:55 |
@fenn | okay, the box with the gigantic dna helix picture laser cut on the side, what is that called | 17:55 |
jonathanc | $85, so arduino mini costs $15 of that | 17:55 |
jonathanc | I dunno should be able to do it in a PIC12F508 which costs $1.00 or so | 17:59 |
jonathanc | optimize that down and there is no library | 17:59 |
jonathanc | there are no functions, there is only main() | 17:59 |
@fenn | 2 resistors ($10), arctic silver epoxy ($14), solid state relay ($8.50), MAX31885 breakout board ($17.50), thermocouple wire ($10) and various hardware | 17:59 |
jonathanc | there is no stack | 17:59 |
@kanzure | no stack? | 17:59 |
@kanzure | hm | 18:00 |
-!- klafka [~klafka@204.101.190.178] has quit [Remote host closed the connection] | 18:00 | |
jonathanc | 2 stack levels, in hardware not software | 18:00 |
jonathanc | call depth more than 2 causes exception reset to save the processor | 18:00 |
jonathanc | that's optimization | 18:00 |
@fenn | jonathanc: people use the arduino because it's pre-built and has a usb bootloader | 18:00 |
@fenn | the idea is to make it something people can build without a full electronics lab with board production capability | 18:01 |
@fenn | you start optimizing from the beginning and there's no room for improvement | 18:02 |
jonathanc | well an 8 pin processor which only requires 1 resistor and 1 capacitor to complete does not really need board production... | 18:02 |
@fenn | yeah well how do you program it | 18:02 |
jonathanc | with a programmer of course, common equip | 18:02 |
@fenn | so, call me spoiled but i'd like at least some kind of status indicator showing how much time is left in the PCR cycle | 18:04 |
jonathanc | save resources by saving $$ | 18:04 |
@kanzure | also why would it be taking 15 engineers to make a thermocycler? | 18:04 |
@kanzure | i'm probably grumpy. i've been writing a caching server all day. | 18:05 |
jonathanc | the socialism comment relates to "the public paying" rather than allowing price (capitalism) to dictate success | 18:06 |
jonathanc | "well let's just be lazy and optimize for my time, and let the end user pay for my lack of optimization skills" | 18:06 |
jonathanc | to paraphrase | 18:06 |
@fenn | price isn't the only variable; if price were the only thing that mattered, people would still be doing PCR by hand with water baths | 18:06 |
@kanzure | no, in that scenario i am my own user | 18:06 |
@fenn | anyway i stick to my position of optimizing for coder/hacker time | 18:07 |
@kanzure | in a stackless architecture i think you could still possibly use libraries if your compiler is good | 18:08 |
jonathanc | what is the library going to do? there's only 1024 bytes of memory. | 18:08 |
jonathanc | there is only 1 .c file | 18:08 |
jonathanc | there is no linker | 18:08 |
jonathanc | there is no .ld file | 18:08 |
@fenn | what kind of broken system are you using | 18:08 |
jonathanc | haha | 18:08 |
@fenn | ffs man join this millennium | 18:08 |
@fenn | why even bother with C if there's no linker and you can't port your code | 18:09 |
jonathanc | the code is portable to a chip with similar architecture | 18:09 |
@fenn | just do it in raw bytecode, it's the most optimal | 18:09 |
@kanzure | yeah we should hire 100 engineers to stare at hex editors all day | 18:10 |
@kanzure | anyone who can't write in raw hex is fired | 18:10 |
jonathanc | it depends on how much % you save | 18:10 |
jonathanc | assembly vs C only saves 1-3% in most cases with this stuff | 18:10 |
jonathanc | that's why risc is so good | 18:10 |
@fenn | so i think we've firmly established that optimizing the controller will at best save 20% of the project cost for our example project (wiremound-pcr) | 18:12 |
jonathanc | there's more that could be saved I think | 18:13 |
jonathanc | thermocoupler dealio could be cut as well | 18:13 |
@fenn | but there's still all sorts of costs unaccounted for: mechanical fabrication assembly and testing | 18:14 |
@fenn | parts procurement, "marketing" | 18:14 |
jonathanc | max31885 why? | 18:14 |
jonathanc | mechanical, you guys would have to comment on | 18:15 |
@fenn | something about bandgaps changing with ambient temperature.. i didnt design it | 18:15 |
jonathanc | lemme guess you want to use the pid c++ library next | 18:16 |
@kanzure | which library? | 18:16 |
jonathanc | cheese ball library previously discussed on the group | 18:17 |
jonathanc | used by openpcr I think | 18:17 |
@kanzure | how much of openpcr was adobe air? | 18:17 |
jonathanc | again with same rationalization "trying to save developer time" | 18:17 |
@fenn | probably this http://playground.arduino.cc/Code/PIDLibrary | 18:17 |
@kanzure | adobe air was not rationalized by "saving developer time".. it was "saving time so i don't have to become a developer". | 18:17 |
@kanzure | actually wait that doesn't make sense, josh can write code | 18:18 |
@kanzure | wtf | 18:18 |
@kanzure | why did josh let openpcr use adobe air? | 18:18 |
@fenn | why would anyone use adobe air | 18:18 |
@fenn | you still have to write code, right? | 18:19 |
@fenn | The Adobe AIR runtime enables developers to deploy standalone applications built with HTML, JavaScript, ActionScript, Flex, Adobe Flash Professional, yadda yadda yadda | 18:19 |
@kanzure | and then you add this proprietary pile of crap in between you and your code too | 18:19 |
@kanzure | and then your users have to use it too because you're sadistic or something | 18:19 |
@kanzure | "To bring the age of the biological computer to a much speedier reality, Endy and his team have contributed all of BIL gates to the public domain so that others can immediately harness and improve upon the tools." | 18:20 |
@fenn | why not just implement a web server on the device and have it serve up all that fancy html | 18:20 |
@kanzure | paperbot: http://www.sciencemag.org/content/early/2013/03/27/science.1232758 | 18:21 |
paperbot | no translator available, raw dump: http://diyhpl.us/~bryan/papers2/paperbot/9eeb0dfaf790e08cc2bcd8433b656816.txt | 18:21 |
@kanzure | oh ffff | 18:21 |
@kanzure | fuck aaas | 18:21 |
-!- klafka [~klafka@204.101.190.178] has joined ##hplusroadmap | 19:10 | |
-!- klafka [~klafka@204.101.190.178] has quit [Ping timeout: 245 seconds] | 19:15 | |
@kanzure | east bay diybio meeting notes http://ec2-107-21-158-78.compute-1.amazonaws.com/wiki/index.php/General_Meeting_3/29/13 | 19:18 |
@kanzure | http://docs.google.com/document/d/1lDLnSkcqjfKTrSMu1SE2RdejCbxehzo786py3WbDo8c/edit | 19:18 |
@kanzure | stalk: Patrik D'haeseleer, Shanee Stopnitzky, Ryan Bethencourt, Matt Harbowy, Jason Euren, Javier, Romy Ilano, Ron Shigeta, Marc Juul | 19:18 |
jonathanc | the thing about an $85 fan-based thermocycler is this: at any hardware store for $25 there's a space heater with 1500 BTU and heating element and fan in a great case with great user controls. only $25. | 19:22 |
jonathanc | bio kids have $25 not $85. | 19:23 |
jonathanc | $25 including manufacturing and even UL approval. | 19:23 |
jonathanc | and marketing and box and retail stickers. | 19:23 |
jonathanc | If the overhead is removed let's say the product retail price drops $5-7 or so | 19:24 |
jonathanc | do direct sale and drop another $4-5 | 19:24 |
jonathanc | here's a much better one. Heater Project by Mike Pearce | 19:27 |
jonathanc | From readme.txt: The Heater 3 project is used to control waterbath heaters and other specialist temperature controlled equipment within the Chemistry Department. | 19:27 |
jonathanc | This project combines the use of the 1-wire routines, serial routines, a P.I.D (Proportional, Integral, Derivative) calculation, ADC, and a interrupt driven burst mode heater control. | 19:27 |
jonathanc | Entire archive. Download (158KB). | 19:27 |
jonathanc | http://www.microchipc.com/sourcecode/PIC_Hi-Tech_C_Mike_Pearces_heater_project.zip | 19:28 |
jonathanc | the "libraries" are .c files | 19:28 |
@kanzure | yes, what else would they be? magic? | 19:29 |
jonathanc | .a | 19:29 |
jonathanc | with linker file | 19:29 |
@kanzure | ok good point | 19:30 |
jonathanc | here, it is monolithic | 19:30 |
@kanzure | there are many other environments where libraries are pulled in with source | 19:30 |
jonathanc | you can see from the source that it is not appropriate to make these into libraries | 19:30 |
* kanzure looks | 19:30 | |
jonathanc | for optimization, the code uses registers etc, and the delay loop for example, requires specific architecture for clock cycles | 19:31 |
@kanzure | checksum.c seems fairly standard, but i haven't finished reading its preamble | 19:31 |
jonathanc | the 1-wire protocol for example is relatively portable except maybe 20 lines would need modification, for example you do not abstract the i/o pins in real embedded code, you reference them directly | 19:31 |
jonathanc | i.e. linux would create some structure blah blah to set an io pin shadow register which later the lower layer driver then pushes to the hardware blah blah overhead | 19:32 |
jonathanc | in fact during a code review this week we will restructure code to remove abstraction purposely because the flash is 92% full | 19:33 |
@kanzure | oh look someone is bothering to do code reviews. great. | 19:33 |
jonathanc | to switch to another micro with bigger memory is the next step up in chip cost which ultimately impacts end user cost | 19:33 |
jonathanc | in a comp sci project these abstraction layers would be added not subtracted | 19:34 |
jonathanc | because resources are infinite. | 19:34 |
@kanzure | nobody said it's a comp sci project | 19:34 |
@kanzure | you don't need comp sci to see that checksum.c is pretty standard | 19:34 |
jonathanc | date time | 19:35 |
-!- jonathanc [~jonathanc@wsip-98-174-141-194.sd.sd.cox.net] has left ##hplusroadmap [] | 19:35 | |
@kanzure | did i fail his test? | 19:35 |
@kanzure | oh he means it's time for a date, not... okay. | 19:36 |
@kanzure | the clock/timing issue is legitimate, because you need to be careful about which functions you either call/include that might block for too long, but that's always been the case. | 19:37 |
@kanzure | and perhaps you could say that such a library manager shouldn't be used for stackless architectures.. but there's still lots of microcontroller environments that have more resources available. | 19:39 |
-!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Remote host closed the connection] | 19:40 | |
-!- Juul [~Juul@50-0-83-116.dsl.static.sonic.net] has joined ##hplusroadmap | 19:47 | |
-!- Juul [~Juul@50-0-83-116.dsl.static.sonic.net] has quit [Client Quit] | 19:47 | |
@kanzure | http://sudoroom.org/wiki/Biohacking | 19:49 |
@kanzure | http://sudoroom.org/wiki/Projects#Biology_Projects | 19:49 |
@kanzure | http://ec2-107-21-158-78.compute-1.amazonaws.com/wiki/index.php/Main_Page | 19:49 |
@kanzure | hah there's a bit of an obsessive list of hackerspaces that tried crowdfunding here http://ec2-107-21-158-78.compute-1.amazonaws.com/wiki/index.php/Fundraising_Ideas | 19:50 |
@kanzure | "unallocated space".. heh. | 19:51 |
-!- Juul [~Juul@50-0-83-116.dsl.static.sonic.net] has joined ##hplusroadmap | 19:54 | |
@kanzure | Juul: meeting update? | 19:57 |
Juul | kanzure, right now talking about what kind of space we want and whether we _need_ bsl2 | 19:58 |
@kanzure | i think you can't get to bsl2 without bsl1, not sure | 19:58 |
Juul | yeah | 19:58 |
Juul | i mean, it depends on the county/city implementation | 20:01 |
Juul | but BSL1 requirements as I've seen them are a subset of BSL2 requirements | 20:02 |
@kanzure | just gotta map out the zoning laws | 20:04 |
Juul | hmm | 20:04 |
Juul | we have a couple of people working on that | 20:04 |
Juul | i think the result will be that we should go for oakland | 20:09 |
Juul | it seems like they may have little/no regulation | 20:09 |
brownies | oakland eh? | 20:10 |
Juul | yep | 20:12 |
@kanzure | it's weird that this was from 2004 http://www.futureofwipo.org/futureofwipodeclaration.html and it mentions wikipedia. wikipedia wasn't that big in 2004. it wasn't exactly a stellar example. | 20:37 |
@kanzure | it would be a more relevant example since maybe 2007. | 20:37 |
brownies | i was using wikipedia in '04 | 20:43 |
brownies | ...before it went mainstream... | 20:43 |
@kanzure | we all were.. but it wasn't exactly something worth mentioning to the World Intellectual Property Organization if you wanted them to understand wtf you're talking about? | 20:44 |
@kanzure | in 2004. | 20:44 |
ParahSail1n | paperbot: http://www.sciencedirect.com/science/article/pii/S0019103513000791 | 20:47 |
paperbot | no translator available, raw dump: http://diyhpl.us/~bryan/papers2/paperbot/The%20Wow%21%20signal%20of%20the%20terrestrial%20genetic%20code.pdf | 20:47 |
ParahSail1n | aliens in ur dnas | 20:47 |
ParahSail1n | i think the ! in the file name broke the python | 20:48 |
ParahSail1n | ah, nevermind, it just took a while to get pdfparanoia'd | 20:48 |
@kanzure | wow i just realized your github username | 20:57 |
@kanzure | i don't know why that wasn't more obvious | 20:57 |
@kanzure | it would be more useful if i could make these realizations while i am reading your bug reports | 20:59 |
@kanzure | you even used your name.. how did i miss that? | 21:00 |
ParahSail1n | yeah, that's more of a "professional" username | 21:07 |
@kanzure | heh | 21:09 |
-!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 21:16 | |
-!- DMXRoid_ [~schiros@oberth.invisihosting.com] has joined ##hplusroadmap | 21:23 | |
-!- DMXRoid [~schiros@oberth.invisihosting.com] has quit [Read error: Connection reset by peer] | 21:24 | |
-!- Cat4D [433456da@gateway/web/freenode/ip.67.52.86.218] has joined ##hplusroadmap | 21:39 | |
-!- klafka [~klafka@204.101.190.178] has joined ##hplusroadmap | 21:48 | |
-!- DMXRoid_ is now known as DMXRoid | 21:49 | |
-!- Juul [~Juul@50-0-83-116.dsl.static.sonic.net] has quit [Ping timeout: 260 seconds] | 21:59 | |
-!- paperbot [~paperbot@131.252.130.248] has quit [Remote host closed the connection] | 22:02 | |
-!- paperbot [~paperbot@131.252.130.248] has joined ##hplusroadmap | 22:03 | |
@kanzure | paperbot: http://users.ece.cmu.edu/~aavgerin/papers/mayhem-oakland-12.pdf | 22:05 |
paperbot | no translator available, raw dump: http://diyhpl.us/~bryan/papers2/paperbot/e35c98abc8251f971cb1952d115c656b.txt | 22:05 |
@kanzure | paperbot: http://users.ece.cmu.edu/~ejschwar/papers/usenix11.pdf | 22:05 |
paperbot | no translator available, raw dump: http://diyhpl.us/~bryan/papers2/paperbot/d1033e87cf00119c59ec698d23a2eac3.txt | 22:05 |
-!- jonathanc [~jonathanc@wsip-98-174-141-194.sd.sd.cox.net] has joined ##hplusroadmap | 22:09 | |
jonathanc | microcontroller libraries http://mbed.org/explore/ | 22:09 |
@kanzure | http://moflow.org/#%5B%5BReference%20Library%5D%5D | 22:09 |
@kanzure | .title | 22:09 |
yoleaux | .: moflow :. - software security workflow | 22:09 |
-!- diginet [~kevin@107-213-84-107.lightspeed.hstntx.sbcglobal.net] has joined ##hplusroadmap | 22:09 | |
diginet | did anyone mirror the NTRS before it got taken down? | 22:10 |
@kanzure | link? | 22:10 |
diginet | the NASA technical report service | 22:10 |
@kanzure | the whole thing was taken down? those guys have balls. they are really trying to make a statement huh. | 22:10 |
diginet | they got paranoid because they realized people were actually reading things they put on the internet | 22:11 |
diginet | I'm still trying to figure out a way that makes any sense | 22:11 |
@kanzure | no it's because of sequestration. nasa has been canceling all sorts of public programs to make people angry. | 22:11 |
diginet | are you sure? I read several news reports that it got shut down as a response to the fact that some Chinese national had been mass downloading docs from the server | 22:12 |
@kanzure | that's... interesting. | 22:12 |
@kanzure | oh no somebody might be actually using our site, pull the plug | 22:12 |
diginet | yeah, it's mind-numbingly stupid | 22:12 |
@kanzure | jonathanc: is this a commercial product? what's going on here? | 22:12 |
@kanzure | hmm http://mbed.org/handbook/Homepage | 22:13 |
@kanzure | what are the target architectures? | 22:13 |
@kanzure | "Prototype to hardware - How to move from your mbed prototype to your custom PCB." | 22:13 |
diginet | "we fear people might be reading stuff we put on the internet for people to read" | 22:14 |
@kanzure | are they selling only their own chips here? | 22:14 |
diginet | mbed is ARM | 22:14 |
jonathanc | target soc is arm-based | 22:14 |
@kanzure | i see. well that's nice. | 22:14 |
jonathanc | arm cortex is popular | 22:14 |
jonathanc | 32-bit | 22:14 |
@kanzure | if they are a business, what is their model? selling devkits? support? | 22:14 |
jonathanc | low power, reduced instruction set | 22:14 |
diginet | reduced instruction set? | 22:14 |
jonathanc | soc vendors fund sw dev | 22:14 |
diginet | in name only | 22:14 |
jonathanc | i have a freescale board in front of me | 22:14 |
diginet | the RISC/CISC distinction is all but meaningless these days | 22:15 |
jonathanc | no, the cortex really is reduced instruction set | 22:15 |
@kanzure | soc vendors fund sw development because "maybe the developers will use our socs" ? | 22:15 |
jonathanc | yes | 22:15 |
@kanzure | neat | 22:15 |
@kanzure | assuming this works. i should play with this. | 22:15 |
@kanzure | this looks better than that bugslabs bullshit. | 22:15 |
jonathanc | dude soc vendors funded major eclipse dev effort | 22:15 |
@kanzure | eclipse is like pure evil | 22:15 |
jonathanc | it is what it is, at least it works | 22:16 |
@kanzure | heh | 22:16 |
diginet | eclipse is like "let's see just how shitty we can make our dev environment, you know, for fun" | 22:16 |
jonathanc | it is a common rallying point | 22:16 |
diginet | RISC has not been a relevant argument in at least 15 years | 22:16 |
@kanzure | true, eclipse is very good at getting attention | 22:16 |
jonathanc | its goal was to run on linux eh | 22:16 |
diginet | RISC predates Linux | 22:17 |
jonathanc | arduino killer http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=FRDM-KL25Z&tid=vanFRDM-KL25Z | 22:17 |
diginet | it was designed with the fact that developers were abandoning directly coding in asm for "high-level" (for the time) languages like C | 22:17 |
jonathanc | $13 for the board, includes onboard jtag debugger | 22:18 |
diginet | or you could, you know, just use the chip | 22:18 |
jonathanc | freescale also supplies rtos | 22:18 |
diginet | plenty of open source ones | 22:18 |
jonathanc | important feature is usb otg | 22:19 |
jonathanc | with enough flash to actually embed a usb stack | 22:19 |
diginet | ugh, I'm so sick of the arduino bullshit | 22:19 |
jonathanc | although i'd have to see proof of that | 22:19 |
diginet | just use an AVR | 22:19 |
diginet | it's simple | 22:20 |
diginet | and much cheaper | 22:20 |
jonathanc | or dump avr and get a real chip you mean | 22:20 |
diginet | I don't care for AVR, but pick your poison | 22:20 |
diginet | I'm just sick of overpriced, glorifed dev boards | 22:20 |
jonathanc | whatever. you wont be able to make it for under $10 | 22:21 |
diginet | AVRs cost like $1 | 22:21 |
jonathanc | i'm talking about kinetis board, not an avr breakout board | 22:22 |
diginet | oh | 22:22 |
jonathanc | adafruit obviously has a huge profit margin | 22:23 |
jonathanc | like lego | 22:23 |
jonathanc | infineon has a similar setup now too, also reduced footprint arm | 22:29 |
jonathanc | they have these funky boards which are hexagonal and plug in together, rather than stackable | 22:31 |
jonathanc | this is it http://www.freertos.org/FreeRTOS-for-Infineon-XMC4000-Cortex-M4.html | 22:32 |
jonathanc | infineon's mini stick shaped boards and their hex boards http://www.infineon.com/cms/en/product/microcontrollers/32-bit-xmc4000-industrial-microcontrollers-arm®-cortex™-m4/xmc4000-development-tools,-software-and-kits/channel.html?channel=db3a30433580b3710135a07902883872 | 22:36 |
@kanzure | if anyone cares i've dumped some more security things into here http://diyhpl.us/~bryan/papers2/security/ | 22:36 |
jonathanc | dude what a crazy url, it has the copyright symbol and tm mark in the url. that's whack | 22:36 |
@kanzure | they just want to make sure your browser is utf8 compliant, haha | 22:37 |
jonathanc | google XMC Development Tools, Software & Kits | 22:37 |
@kanzure | you knew about mbed earlier tonight? | 22:37 |
jonathanc | i've had this board for a while | 22:37 |
jonathanc | did an rtos evaluation a couple months back | 22:38 |
-!- diginet [~kevin@107-213-84-107.lightspeed.hstntx.sbcglobal.net] has left ##hplusroadmap ["Leaving"] | 22:38 | |
@kanzure | any complaints? | 22:38 |
jonathanc | problem with mbed is it is online development | 22:38 |
@kanzure | what why | 22:38 |
jonathanc | well i told you about the toolchain problem | 22:39 |
jonathanc | how are you going to compile a library if you can't get the toolchain without a $3000-per-seat license? | 22:39 |
jonathanc | the "free" toolchain is online, i.e. virtual license supported by the vendors | 22:40 |
jonathanc | who knows how long it will last | 22:40 |
jonathanc | intel had a similar system when the 80960 came out | 22:40 |
jonathanc | log in, compile embedded code for 80960j I think it was, get the static binary, then burn to local board | 22:40 |
jonathanc | they don't want to kill the 3rd party dev companies which make the toolchains, then they will be s.o.l. | 22:41 |
jonathanc | another limitation with mbed is it is nxp which I'll never use again | 22:41 |
jonathanc | also, while freescale does give away the rtos, and it is a really good rtos, it is only to be used on their own soc, not on others | 22:43 |
-!- Cat4D [433456da@gateway/web/freenode/ip.67.52.86.218] has quit [Ping timeout: 245 seconds] | 22:44 | |
jonathanc | check this one http://www.kickstarter.com/projects/paulstoffregen/teensy-30-32-bit-arm-cortex-m4-usable-in-arduino-a | 22:44 |
jonathanc | same freescale chip. freescale is solid but tries to lock in via offering software for their chip only, same story as always | 22:45 |
jonathanc | funny quote "...the Galago might just be the perfect ARM board for tinkerers weaning themselves off the Arduino." "...the Galago might just be the perfect ARM board for tinkerers weaning themselves off the Arduino." | 22:47 |
jonathanc | galago uses the same nxp chip as the mbed platform http://www.kickstarter.com/projects/kuy/galago-make-things-better | 22:47 |
@kanzure | the arm toolchains are $3k/seat? | 22:48 |
jonathanc | depends on the jtag | 22:48 |
@kanzure | doesn't llvm target arm just fine? | 22:49 |
jonathanc | i don't know if it does cortex instruction set or not | 22:53 |
jonathanc | this is a different isa | 22:53 |
jonathanc | isa=instruction set architectur | 22:54 |
jonathanc | isa=instruction set architecture | 22:54 |
@kanzure | arm cortex is not normal arm isa? | 22:54 |
@kanzure | i should probably look into this huh | 22:54 |
jonathanc | cortex is reduced set like I said | 22:55 |
jonathanc | either way it basically kills avr so why bother with that I dunno | 22:56 |
jonathanc | more libraries for you http://www.pjrc.com/teensy/td_libs.html | 23:02 |
jonathanc | "Linaro is a not-for-profit engineering organization consolidating and optimizing open source Linux software and tools for the ARM architecture." http://www.linaro.org/ | 23:04 |
@kanzure | that's quite a string of words | 23:05 |
@kanzure | oh i see i can read it now | 23:05 |
@kanzure | seems to be an industry consortium thing | 23:06 |
jonathanc | this is the toolchain for the arm cortex apparently | 23:06 |
jonathanc | https://launchpadlibrarian.net/126639247/readme.txt | 23:06 |
jonathanc | -mthumb -mcpu=cortex-m3 | 23:06 |
jonathanc | -mthumb -mcpu=cortex-m4 | 23:06 |
jonathanc | link https://launchpad.net/gcc-arm-embedded/4.7/4.7-2012-q4-major | 23:07 |
jonathanc | I am spoiled because I haven't used gcc in years now, thank darwin | 23:07 |
jonathanc | there's something to be said for buying a commercial toolchain and just building, vs all kinds of bootstrap junk with gnu | 23:08 |
jonathanc | this is a pretty sweet cortex board too, teensy 3.0 https://www.sparkfun.com/products/11780 | 23:09 |
jonathanc | I played with teensy 2.0 when it came out (avr) but a little too slow and limited chip. a great starter for beginners though, drag-and-drop firmware download interface is neat | 23:10 |
jonathanc | toolchain install is a gnu p.o.s. though | 23:10 |
jonathanc | mbed nxp soc board from sparkfun https://www.sparkfun.com/products/9564 | 23:12 |
jonathanc | $60, that aint cheap | 23:13 |
jonathanc | in that case, UBW32 is a better option | 23:14 |
@kanzure | maybe the open source options wouldn't suck so much if companies were actually trying to make something usable that doesn't cost $50k/seat | 23:16 |
jonathanc | what do you mean | 23:17 |
jonathanc | dudes get a lot done with the iar toolchain here | 23:21 |
jonathanc | tons done | 23:21 |
jonathanc | actually work on the project instead of working on makefile hell | 23:22 |
jonathanc | change linker params in a couple minutes compared to .ld nonsense | 23:22 |
jonathanc | for the cost of 2 weeks employment, gain more than that back in productivity | 23:23 |
jonathanc | plus a much better optimizer for sure | 23:24 |
@kanzure | :fear: http://www.libcrack.so/2012/10/13/hacking-the-ar-drone-parrot/ | 23:24 |
@kanzure | someone is offering a rom dump sometime tomorrow. can't wait to start picking at it. | 23:26 |
-!- OldCoder_ [~OldCoder_@209.237.22.146] has quit [Remote host closed the connection] | 23:34 | |
jonathanc | I found out from the director of sweng at the office that another hw/sw development office of ours was using UBW boards like crazy for development, same as our office | 23:35 |
@kanzure | i am being distracted by security things at the moment. this stuff is like candy to me. :( | 23:35 |
-!- yashgaroth [~ffffff@cpe-66-27-118-94.san.res.rr.com] has quit [Quit: Leaving] | 23:51 | |
--- Log closed Sat Mar 30 00:00:42 2013 |
Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!