Fixpoint

2024-10-04

#jwrd Logs for Oct 2024

Filed under: #jwrd logs, Logs — Jacob Welsh @ 06:02
Day changed to 2024-10-04
[06:02] jfw: whaack: any progress on the crypto-maths research? hopefully you didn't get too bogged down in the riveting internals of the HAC
[06:02] sourcerer: 2024-01-14 18:40:47 (#jwrd) whaack: jfw: I've been enjoying going through the handbook, only ~550 pages to go until I get to Montgomery forms!
[06:07] jfw: I've been looking at Montgomery 1987, "Speeding the Pollard and Elliptic Curve Methods of Factorization"; I suspect the core of the idea is in there but most of the detail isn't relevant to our goal at hand
[15:16] whaack: howdy
[15:21] whaack: i've let the study of / leading up to Montgomery forms fall to the wayside, been mostly invested in working on a business with billymg.
[15:22] whaack: thanks for the reminder though, it does slightly bother me that i'm not doing anything bitcoin related at all
[15:33] jfw: I see; perhaps the math requires more concentrated focus than you're presently able to muster; but it's probably less than you think, recall the "flashrom" trouble that looked like a brick wall until found to be just painted tissue paper
[15:34] jfw: and how's that insulation & sealing business going?
[15:35] whaack: jfw: we shifted instead to creating a market place for construction/home repair services in Gaunacaste, Costa Rica
[15:37] jfw: like a website or app or something?
[15:37] whaack: yup, webapp
[15:37] whaack: the good news is i've been programming more
[15:38] whaack: i still am sore everyday but have kind of just came to accept it
[15:38] whaack: gtg, bbl
[15:38] jfw: haha, so even brick-and-mortar ideas turn into piles of tech now
[15:38] jfw: sorry to hear about the hands :/
[16:09] whaack: eh, my life is so good otherwise, i needed some hardship to balance it out
[16:11] whaack: jfw: and yeah, can't escape the sedentary lifestyle
[16:15] jfw: well as they say, decrepity is the brother of intention
[16:19] whaack: heheh
[16:20] whaack: if you like maxims as well as aphorisms, i found a nice translation of an old spanish text with 300 of 'em https://sacred-texts.com/eso/aww/aww10.htm
[16:26] whaack: how's https://jwrd.net/ going? Is cruciform a student now?
[16:26] whaack: !e view-height
[16:26] btcexplorer: block_height: 864145
[16:26] btcexplorer: mins_since_last_block: -30
[16:26] whaack: !E view-height
[16:26] nzbtcexplorer: block_height: 864145
[16:26] nzbtcexplorer: mins_since_last_block: 3
[16:26] whaack: looks like the reconnect bug was indeed fixed once and for all
[16:29] jfw: he finished up the advanced module (mostly; we tailored it a bit). then he went to do some traveling.
[16:30] jfw: he vaguely wants to start a blog but seems to need some support in the process
[16:53] jfw: http://jfxpt.com/2023/jwrd-logs-for-Apr-2023/#6912 - it's looking like I had things a bit conflated at this point. There's Montgomery form which pertains to optimizing repeated modular reductions of integers, by allowing the intermediate ones to use a different denominator, which can be a power of two to take advantage of bit shifting. This comes possibly better from Montgomery's 1985 paper, and
[16:53] sourcerer: 2023-04-18 23:19:48 (#jwrd) jfw: the current code if I recall the terminology would be said to be using 'affine form' coordinates ie plain x & y that directly satisfy the curve equation, the straightforward implementation of the algorithms shown in the standards, but openssl & others all went with fancier versions for performance.
[16:53] jfw: may or may not be of use here. Then there are the projective coordinate forms of EC points, of which there's at least standard and Jacobian variants, which allow deferring the modular inversions (even more costly than reductions) to a single operation the end. This sounds the most promising.
[16:54] jfw: Then there are lesser optimizations like the 'binary' extended gcd algorithm, as well as 'cheats' exploiting properties of the particular curve parameters, which were chosen so as to enable same.
[16:56] jfw: warez since http://www.cacr.math.uwaterloo.ca/ecc/ was killed
Day changed to 2024-10-06
[00:48] dorion: whaack, fyi, you've quit/joined like 25 times in the last 30h.
[20:03] jfw: whaack: mind fixing that flapping connection or leaving it off for now?
Day changed to 2024-10-16
[23:15] jfw: big news coming on the gales scheme front, hopefully within the next couple days.
Day changed to 2024-10-17
[22:45] jwm2: Dont make me upgrade jacob, i can barely move around my directories!
[22:45] jwm2: jk - happy there is progress!
Day changed to 2024-10-18
[03:32] jfw: jwm2: this is on the scheme interpreter; it doesn't involve much mkdir and cding around directories, but it does involve consing up and cdring down lists
[03:39] jfw: while at the hosting plant, it took some more time to pick up a portly UPS, finalize the power and LAN cabling, and install internal FUCKGOATS TRNG, but the box is ready to go at the physical level now.
[03:39] sourcerer: 2024-09-26 01:23:32 (#jwrd) jfw: http://jfxpt.com/2024/jwrd-logs-for-Sep-2024/#12432 - after some excavation, not quite street but sidewalk at least, further confusing communications and another 5 days delay... the signal is finally live!
[03:40] jfw: I've now got its busybox rebuilt for the tar fix, rngnet feeding /dev/urandom, bitcoind built and a sync kicked off.
[03:41] jfw: it's on spinning HDDs so we'll see how that goes, get some data at least.
[03:42] jfw: I'm probably going to redo its storage one way or another before opening it up for rental, because the drives aren't even that big by present standards.
[03:55] jfw: 38.191.74.234 for who wants to "addnode" it.
Day changed to 2024-10-31
[19:27] dorion: so I tried creating a twatter account today, want to see if there's anyone who can be brought from there to here/e2. doesn't work in firefox 'strict' mode. doesn't work in 2 different versions of chrome/chromium either. receive their verification email, enter it. they throw a vague error that says they failed to create the account, but not why. such high tech.
[20:03] jfw: you don't look enough like a robot
[20:05] jfw: hadn't heard of that firefox "strict mode", it sounds a bit suspect
[20:06] jfw: like, allow scripts & cookies but block certain "bad" kinds based on content analysis? sounds rather "antivirus"
[20:07] jfw: like they grew up in a cave with respect to theory of computation and nobody ever told them about the halting problem
[21:52] dorion: jfw, here's more info about it : https://support.mozilla.org/en-US/kb/enhanced-tracking-protection-firefox-desktop ; i'm already using noscript and ublock origin, perhaps it's overkill. anyways, didn't think it'd hurt anything to run with it and haven't noticed it break any other sites.
[21:58] jfw: yes that's the one I saw.
[21:59] jfw: my point is that since what it does isn't well defined, netiher is what it will break.
[22:00] jfw: though since it's mozilla, one would expect them to err on the side of not breaking the major platforms or advertisers.

No Comments »

No comments yet.

RSS feed for comments on this post. TrackBack URL

Leave a comment

Powered by MP-WP. Copyright Jacob Welsh.