←2023-02-25 2023-02-26 2023-02-27→ ↑2023 ↑all
00:56:17 -!- TheFloatingPixel has quit (Quit: Client closed).
01:15:19 -!- fungot has joined.
01:24:11 -!- sam has quit (Quit: leaving).
01:24:37 -!- sam has joined.
01:24:37 -!- sam has quit (Client Quit).
02:00:59 -!- sam has joined.
02:18:19 -!- Lord_of_Life has quit (Ping timeout: 246 seconds).
02:18:21 -!- Lord_of_Life_ has joined.
02:19:42 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
02:20:59 -!- razetime has joined.
02:59:31 -!- razetime1 has joined.
02:59:43 -!- razetime has quit (Ping timeout: 248 seconds).
02:59:44 -!- razetime1 has changed nick to razetime.
03:23:25 -!- razetime has quit (Ping timeout: 268 seconds).
03:30:50 -!- FreeFull has quit.
04:23:11 -!- razetime has joined.
04:24:31 -!- sprout has quit (Ping timeout: 248 seconds).
04:25:39 -!- sprout has joined.
04:30:16 -!- sprout has quit (Ping timeout: 246 seconds).
04:31:11 -!- sprout has joined.
04:36:14 -!- sprout has quit (Ping timeout: 255 seconds).
04:57:20 -!- razetime has quit (Read error: Connection reset by peer).
04:59:53 -!- razetime has joined.
05:04:55 -!- sprout has joined.
05:10:07 -!- sprout has quit (Ping timeout: 252 seconds).
05:37:44 -!- sprout has joined.
05:39:57 -!- razetime has quit (Remote host closed the connection).
05:43:00 -!- sprout has quit (Ping timeout: 260 seconds).
05:46:41 -!- SGautam has joined.
06:10:32 -!- sprout has joined.
06:15:28 -!- sprout has quit (Ping timeout: 248 seconds).
06:34:23 -!- sprout has joined.
06:39:35 -!- sprout has quit (Ping timeout: 260 seconds).
06:41:22 -!- craigo has quit (Ping timeout: 268 seconds).
07:08:34 -!- sprout has joined.
07:12:22 <esolangs> [[Deadfish]] https://esolangs.org/w/index.php?diff=107047&oldid=106988 * Arctenik * (+2466) /* Implementations */ Add Minecraft commands implementation
07:13:52 -!- sprout has quit (Ping timeout: 252 seconds).
07:28:01 -!- sprout has joined.
07:32:47 -!- sprout has quit (Ping timeout: 248 seconds).
07:46:35 -!- sprout has joined.
07:51:31 -!- sprout has quit (Ping timeout: 246 seconds).
08:03:31 -!- sprout has joined.
08:08:15 -!- sprout has quit (Ping timeout: 260 seconds).
08:20:49 -!- sprout has joined.
08:25:03 -!- sprout has quit (Ping timeout: 248 seconds).
08:32:59 <int-e> `? hyperloo
08:33:02 <HackEso> hyperloo? ¯\(°​_o)/¯
08:34:13 <int-e> . o O ( It features supersonic flushing. )
08:34:42 <int-e> . o O ( But it's also connected to a number of mysterious disappearances. )
08:56:34 -!- sprout has joined.
09:01:31 -!- sprout has quit (Ping timeout: 248 seconds).
09:13:37 -!- sprout has joined.
09:18:19 -!- sprout has quit (Ping timeout: 252 seconds).
09:30:52 -!- sprout has joined.
09:31:54 -!- SGautam has quit (Quit: Connection closed for inactivity).
09:35:28 -!- sprout has quit (Ping timeout: 246 seconds).
09:42:52 <shachaf> I got a nice answer here framing it the same way int-e did: https://math.stackexchange.com/a/4642312
09:43:12 <shachaf> Though I'd still like to know about the core of the question, on general D+P>q bounds.
09:57:35 <int-e> That indeed looks eerily familiar.
09:58:37 -!- sprout has joined.
09:59:33 <int-e> And I think they got about as far as I did? I don't quite recall where exactly I gave up.
10:00:45 <int-e> It's not of huge practical importance in that you can always make your shards larger relatively cheaply.
10:02:49 <int-e> Heck I don't even remember *when* we discussed this here... half a year ago? More? Less?
10:39:27 <shachaf> Presumably less.
10:39:52 <shachaf> I'm not sure I knew anything about erasure codes half a year ago.
11:01:54 <b_jonas> no, we were talking about normal error-correcting codes and secret sharing instead
11:04:25 <b_jonas> yeah, looks like we weren't discussing them ebfore 2022-10
11:04:41 <int-e> Hmm. Well, the latter is really close to erasure codes, except that you always erase the original n shards, and p is at least n.
11:04:50 <b_jonas> (or we used other words or codes to mention them so grep doesn't find it)
11:05:18 -!- Sgeo has quit (Read error: Connection reset by peer).
11:05:54 <b_jonas> int-e: kind of, but secret sharing has more strict requirements so that you can never recover parts of the cleartext even probabilistically
11:05:55 <int-e> Basically, the design question is the same if you focus on systematic linear codes.
11:06:23 <int-e> b_jonas: we have yesterday's counting argument that ensures the secrets are actually obscured if you have too few shards.
11:07:43 <shachaf> int-e: I think in most secret sharing schemes the secret is just a single shard?
11:07:44 <int-e> And in particular, the same "matrix with all square submatrices being invertible" characterization also occurs in the secret sharing case.
11:07:50 <shachaf> You add randomness with the other shards.
11:08:15 * int-e shrugs.
11:08:32 <int-e> Maybe I'll read old logs later now that b_jonas has narrowed it down a bit.
11:08:45 <shachaf> Is the "matrix with all suqare submatrices being invertible" characterization only applicable to systematic codes?
11:10:11 <int-e> As yes, because it interacts with the fixed identity matrix.
11:10:34 <int-e> Hmm.
11:10:55 <shachaf> The general characterization is "NxK matrix where any KxK submatrix is invertible", presumably.
11:10:56 <b_jonas> I admit I don't feel like thinking too much about this. secret sharing is practically solved to my satisfaction; you need erasure code for raid arrays of hard disks or SSDs but you only need small ones with few shards and for large data and you don't mind a little bit of overhead so it's practically solved too; I feel like the actually important stuff is error-correcting codes rather than erasure codes
11:10:58 <int-e> So no, this doesn't apply to secret sharing. But I'm pretty sure *I* ventured into erasure codes last year for some reason.
11:11:21 <int-e> For secret sharing you only care about the maximal square submatrices.
11:11:29 <shachaf> b_jonas: Yes, error-correcting codes seem a lot more interesting and tricky.
11:11:43 <shachaf> I don't even know how to do them efficiently.
11:12:43 <b_jonas> admittedly you may also want larger erasure codes for long distance radio communication where you transmit shards one after another and there can be radio noise erasing multiple shards, and then you use error-correcting codes inside shards in case parts of them are erased obvisouly, and that combo might work better than a generic error-correction code combination
11:12:55 <int-e> I mean the connection is still there in the sense that if you have a systematic erasure code with sufficiently many parity shards, you can use those parity shards for secret sharing.
11:13:14 <int-e> But I'm less sure now that this is how I got there.
11:14:26 <b_jonas> but in this radio example too you don't need a perfect code, you can accept a small (compared to your data) overhead and that apparently makes the problem somewhat easier (in some cases at least)
11:16:10 <b_jonas> but also, in the radio case your code unit (or shards) will want to be large, so it's not hard to use a linear scheme based on polynomials in a finite field, right?
11:16:35 <b_jonas> it's not the most computationally efficient but not too bad either
11:17:05 <b_jonas> especially since if you need it at large scale you can make custom hardware for it and then it will be as efficient as any
11:25:04 <b_jonas> I mean at that point you're using all sorts of funny combinations of error-correcting codes and checksums too so the erasure code isn't really worse
11:54:58 <shachaf> int-e: Presumably you have to use randomness for secret sharing.
11:56:00 <int-e> ?
11:56:21 <b_jonas> yes. why is that a problem?
11:56:23 <int-e> Shamir's secret sharing doesn't need any randomness; it's an information-theoretic argument.
11:56:47 <int-e> I mean, obviously, if the secret is non-random (read: guessable) then you have a problem.
11:56:54 <shachaf> How can a deterministic algorithm split a secret?
11:58:00 <int-e> err
11:58:15 <int-e> Never mind, I'm still in erasure code mode I think.
11:58:40 <int-e> You need those extra polynomial coefficients and yes, those should better be random.
12:01:32 <b_jonas> you can make your secret random with some symmetric cryptography if that helps, that can be worth if the secret is large
12:02:04 <shachaf> I assume secret sharing usually operates on GF(256) or something.
12:02:07 <b_jonas> but it's not forced by some cryptanalysis nonsense like how you make the data that you sign with RSA random
12:02:50 <b_jonas> shachaf: in IOCCC yes, but you can use a different field
12:05:28 -!- tromp has joined.
12:10:47 <int-e> hmm when did debian add non-free-firmware
12:16:38 <b_jonas> int-e: they have had a non-free section that's like not officially part of debian. I have it added in the sources.list but then all packages from it filtered out by default in aptitude so I can install the few GFDL licensed manuals that they somehow consider non-free when they consider gnuplot with its similarly restrictive license free
12:16:45 <int-e> tromp: Hmm is any of the alpha (psuedo-)combinator stuff published?
12:17:29 <int-e> b_jonas: I know about that one. But see https://packages.debian.org/sid/firmware-nvidia-gsp
12:17:53 <int-e> I had to add non-free-firmware as well for that.
12:17:57 <tromp> not published, just mentioned in some forums
12:18:24 <b_jonas> int-e: isn't that also in the non-free section?
12:18:58 <int-e> Nope. Note that this is unstable (sid). I'm sure it's a recent change.
12:19:02 <tromp> e.g. in https://math.stackexchange.com/questions/153163/what-is-the-shortest-function-of-lambda-calculus-that-generates-all-functions-of
12:19:06 <shachaf> So I quite like this thing about computing Fibonacci numbers using the quotient ring Z[p]/[p^2 = p + 1]
12:19:30 <shachaf> And I can see how to make it work for any linear recurrence, and how to show that it works, but I feel like I'm missing some context.
12:26:07 -!- bgs has joined.
12:31:58 -!- __monty__ has joined.
12:32:33 -!- TheFloatingPixel has joined.
12:33:46 <int-e> b_jonas: Ah, this seems relevant: https://www.debian.org/vote/2022/vote_003 ...having a distinct section name helps with managing what goes into the installer.
12:41:49 <b_jonas> int-e: so the non-free firmware remains in a separate section so it's easy to tell which packages are non-free, but they're included on the install DVD? I don't particularly care about what goes to the full DVD because the debian DVDs don't seem to work well in practice anyway
12:42:34 <int-e> b_jonas: as far as I can see they've split non-free into firmware and other stuff (though I don't know what other stuff there is)
12:43:44 <int-e> anyway, I don't care beyond the fact that I can now update this system again.
12:44:19 <int-e> after adding non-free-firmware to sources.list
12:44:20 <b_jonas> I see. and that applies starting from debian 12 but not to 11, right?
12:44:52 <int-e> I assume so.
12:45:01 <int-e> It would be crazy to break stable like this.
12:45:54 <b_jonas> yeah, it looks like bullseye doesn't have a non-free-firmware section but bookworm does
12:46:42 <b_jonas> where are the names for the debian versions listed officially? there used to be a nice link from the debian homepage
12:46:45 <int-e> And who knows, maybe they'll make this addition automatic for dist upgrades to the next stable release for installations that include any non-free firmware packages.
12:46:47 <b_jonas> but they like hid it now
12:47:24 <b_jonas> https://www.debian.org/releases/index.en.html
12:48:23 <b_jonas> so 9 is stretch, 10 is buster aka oldstable, 11 is bullseye aka stable, and bookworm is the next one
12:48:26 <int-e> yeah that's the only source I know
12:49:01 <b_jonas> the page is there, they just hid the link from the frontpage
12:49:47 <b_jonas> https://www.debian.org/doc/manuals/debian-faq/ftparchives#sourceforcodenames goes back farther, but even https://www.debian.org/releases/index.en.html goes to before I first used Debian
12:49:59 <int-e> . o O ( what's a front page... I usually bypass those with a search engine )
12:51:31 <int-e> ah, the FAQ is way better
12:52:00 <b_jonas> hmm, https://www.debian.org/doc/manuals/debian-faq/ftparchives#sourceforcodenames doesn't acknowledge that "sid" is also an acronym standing for "still in development"
12:52:21 <b_jonas> it does tell what sid is and who Sid in Toy Story is, but not the dual meaning
12:52:29 <b_jonas> s/meaning/etymology/
12:53:38 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
12:54:37 <b_jonas> hmm
12:54:46 <b_jonas> they are going to run out of good codenames in a decade
12:55:21 <b_jonas> "rocky" is still usable, and "spell" probably, and there's an RC car called "RC" but that would be a really bad name for a release because everyone will read it as meaning "release candidate"
12:56:05 <b_jonas> but after that you'll have to use toys from the sequels or something
12:56:48 <b_jonas> and Toy Story 2 adds very few new toys
12:58:33 <int-e> . o O ( Video title: The Death of Debian ...and then talk about Toy Story characters. )
12:59:18 <int-e> FWIW I didn't even know that that's where they take their names from.
13:00:47 <b_jonas> "Debian invests donations into flop sequel Toy Story 5"
13:02:02 <int-e> While critics lament the confusing cast of new characters with no purpose for the story, Debian went on record stating that it was money well spent.
13:02:35 <b_jonas> or they could just go: stretch, buster, bullseye, bookworm, trixie, rocky, 15, 16, 17, 18, 19, 20, 21, 22
13:03:10 <int-e> TRUE, FALSE, FILE_NOT_FOUND
13:16:21 <b_jonas> perhaps they could cut costs if they skip the Hollywood movie that nobody would want to watch anyway. release some tie-in merchandise books and cheap toys, write a bunch of reviews of the first half of the advance screening by reviewers who walked out the room in the middle, document the characters on Wikia and TvTropes, post some social media posts on whatever people use at that time about how the
13:16:27 <b_jonas> movie is so bad that cinemas don't even bother to show it – no wait, cinemas won't exist anymore, make that streaming providers don't even host the movie
13:21:59 <b_jonas> int-e: isn't it FALSE = 0, TRUE = 1, ENOENT = 2 ?
13:23:11 <int-e> No. The reversal of True and False was almost as hurtful as the extra value: https://thedailywtf.com/articles/What_Is_Truth_0x3f_
14:01:04 -!- FreeFull has joined.
14:37:08 -!- tromp has joined.
16:13:04 <esolangs> [[Special:Log/newusers]] create * Richard565 * New user account
16:19:29 -!- razetime has joined.
16:26:41 <esolangs> [[Esolang:Introduce yourself]] https://esolangs.org/w/index.php?diff=107048&oldid=107039 * Richard565 * (+213)
16:59:13 -!- razetime has quit (Ping timeout: 252 seconds).
17:01:23 -!- razetime has joined.
17:14:39 -!- razetime has quit (Ping timeout: 248 seconds).
17:35:17 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
17:39:59 -!- tromp has joined.
17:45:07 <esolangs> [[Cellarg]] N https://esolangs.org/w/index.php?oldid=107049 * TheFloatingPixel * (+5875) Created page with "{{infobox proglang |name=CellArg |paradigms=imperative |author=[[TheFloatingPixel]] |year=[[:Category:2023|2023]] |memsys=[[:Category:Cell-based|Cell-based]] |dimensions=one-dimensional |refimpl=[https://thefloatingpixel.github.io/CellArg/playground] |influence
17:46:37 <esolangs> [[Special:Log/move]] move * TheFloatingPixel * moved [[Cellarg]] to [[CellArg]]: I mistakenly created the original page with improper capitalisation
17:49:24 <esolangs> [[Truth-machine]] https://esolangs.org/w/index.php?diff=107052&oldid=106496 * TheFloatingPixel * (+349) add CellArg Truth machine implementation
17:50:22 <esolangs> [[Truth-machine]] https://esolangs.org/w/index.php?diff=107053&oldid=107052 * TheFloatingPixel * (-112) it actually isn't
18:13:30 <esolangs> [[CellArg]] M https://esolangs.org/w/index.php?diff=107054&oldid=107050 * Gears * (+54) Make example programs wrap
18:21:23 -!- craigo has joined.
18:21:51 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
18:29:45 -!- zzo38 has quit (Ping timeout: 255 seconds).
18:47:40 -!- tromp has joined.
19:15:24 -!- zzo38 has joined.
20:36:53 -!- Sgeo has joined.
20:58:07 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
21:00:02 -!- tromp has joined.
21:03:14 <esolangs> [[Timers]] M https://esolangs.org/w/index.php?diff=107055&oldid=105101 * Rphii * (-11) /* Truth machine */
21:21:15 -!- joast has joined.
21:33:51 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
21:38:47 -!- tromp has joined.
21:40:46 <esolangs> [[K-on Fuck]] M https://esolangs.org/w/index.php?diff=107056&oldid=60105 * Richard565 * (+129) Added the show origin. Fixed some capitalization and spelling.
21:59:25 -!- simcop2387 has quit (Quit: ZNC 1.8.2+deb3+b4 - https://znc.in).
21:59:25 -!- perlbot has quit (Quit: ZNC 1.8.2+deb3+b4 - https://znc.in).
22:02:23 <esolangs> [[Timers]] M https://esolangs.org/w/index.php?diff=107057&oldid=107055 * Rphii * (-287) /* Commands */
22:03:52 <esolangs> [[Timers]] M https://esolangs.org/w/index.php?diff=107058&oldid=107057 * Rphii * (+0) /* Fibonacci */
22:07:08 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
22:10:55 <esolangs> [[Timers]] M https://esolangs.org/w/index.php?diff=107059&oldid=107058 * Rphii * (-128) /* Examples */ remove fizz buzz (for now)
23:11:04 -!- simcop2387 has joined.
23:12:04 -!- perlbot has joined.
23:38:51 <esolangs> [[Tenshi C]] N https://esolangs.org/w/index.php?oldid=107060 * Richard565 * (+2506) Created page with "{{Stub}} '''Tenshi C''' or '''TenC''' is an esolang created by [[User:Richard565]] This language is an implementation of a small subset of C to work on a custom runtime(A move architecture instruction set) he wrote in Javascript. He named it after the song he was
23:41:48 <esolangs> [[Language list]] M https://esolangs.org/w/index.php?diff=107061&oldid=107037 * Richard565 * (+15) added Tenshi C
23:49:24 -!- bgs has quit (Remote host closed the connection).
23:50:03 -!- __monty__ has quit (Quit: leaving).
←2023-02-25 2023-02-26 2023-02-27→ ↑2023 ↑all