On Fri, Jan 17, 2014 at 10:15:40AM +0100, Mike Hearn wrote:
> I must say, this shed is mighty fine looking. It'd be a great place to
> store our bikes. But, what colour should we paint it?

I think we should paint it this colour:

    They had uncovered what seemed to be the side of a large coloured
    globule embedded in the substance. The colour, which resembled some
    of the bands in the meteor's strange spectrum, was almost impossible
    to describe; and it was only by analogy that they called it colour
    at all.  Its texture was glossy, and upon tapping it appeared to
    promise both brittle ness and hollowness. One of the professors gave
    it a smart blow with a hammer, and it burst with a nervous little
    pop. Nothing was emitted, and all trace of the thing vanished with
    the puncturing. It left behind a hollow spherical space about three
    inches across, and all thought it probable that others would be
    discovered as the enclosing substance wasted away.

I think it really gets to the core of my feelings about this naming
discussion.

> How about we split the difference and go with "privacy address"? As Peter
> notes, that's what people actually like and want. The problem with stealth
> is it's got strong connotations with American military hardware and perhaps
> thieves sneaking around in the night:
> 
>    https://www.google.com/search?tbm=isch&q=stealth

WOW! AWESOME KICK-ASS PICS!

Come to think of it, I could have called it "incognito addresses" - a
term nice enough that Google and Firefox use it in their browsers - but
what's done is done and any further discussion about this is just going
to confuse the public. Remember that in the long run all this stuff will
be hidden behind payment protocols anyway, and users *won't even know*
that under the hood a stealth address is being used, making the name
just a technical detail. For now though, lets use the good PR and get
some early adopters on board.

However, the term 'incognito' probably would be a good one to use within
wallet software itself to describe what it's doing when the user clicks
the "I want my transactions to be private" setting - there are after all
fundemental bandwidth-privacy trade-offs in the threat model supposed by
both prefix and bloom filters. In this instance the term isn't going to
go away.


Anyway, back to work: For the actual address format I strongly think we
need to ensure that it can be upgrading in a backwards compatible way.
This means we have to be able to add new fields - for instance if
Gregory's ideas for different ways of doing the SPV-bait came to
fruition. Given that "addresses" aren't something that should stay
user-visible forever, thoughts on just making the actual data a protocol
buffers object?

Second question: Any performance figures yet on how efficient scanning
the blockchain for matching transactions actually is? I'd like to get an
idea soon for both desktop and smartphone wallets so we can figure out
what kind of trade-offs users might be forced into in terms of prefix
length.

-- 
'peter'[:-1]@petertodd.org
0000000000000001c9b372ed519ecc6d41c10b42a7457d1ca5acd560a535596b