00:00:33 <tswett> hppavilion[1]: so I think theoretically a neural net should work pretty well.
00:00:42 -!- jaboja has joined.
00:00:58 <tswett> How do you feel about reimplementing more complicated versions of state-of-the-art techniques?
00:01:15 <tswett> If you do that, you'll be golden. Golden, I tell you.
00:02:53 <hppavilion[1]> tswett: Why would I need to do that, out of curiousity? 0.0
00:03:29 <tswett> Well, there are lots of different flavors of neural nets.
00:03:42 <tswett> The basic flavor just takes a tuple of numbers as input and spits out a tuple of numbers as output.
00:03:58 -!- tromp has joined.
00:04:01 <\oren\> fizzie: 0b for binary works in gcc.
00:04:47 <tswett> Such a neural net is practically incapable of doing certain tasks, such as remembering information for any amount of time, or comprehending multiple instances of a class at once.
00:05:02 <hppavilion[1]> tswett: Of course, I want to deal with strings in some cases, so..
00:05:28 <tswett> Yup. And the basic way to deal with strings is to use a recurrent neural net.
00:05:38 -!- v^ has quit (Read error: Connection reset by peer).
00:05:52 <tswett> Which is the sort of thing that Karpathy's char-rnn program implements.
00:06:30 <fizzie> \oren\: Yes, but not in C.
00:06:39 <tswett> Well, lemme look up PyBrain.
00:06:41 <hppavilion[1]> (Pybrain supports recurrent neural nets, I believe)
00:07:13 <tswett> From the website: "PyBrain is a modular Machine Learning Library for Python." That right away makes me think the answer is probably yes.
00:07:14 <fizzie> I'm contractually obligated to bring up TensorFlow at this juncture. (Okay, not really.)
00:08:07 <fizzie> https://www.tensorflow.org/
00:08:57 <hppavilion[1]> Immediately: Oh god, category theory is nagging me again
00:09:09 <fizzie> It's pretty new, and also pretty generic, so it might not be quite as... how should one put it, task-oriented, than some other machine learning libraries.
00:09:29 -!- ^v has quit (Ping timeout: 250 seconds).
00:11:30 <fizzie> I don't have an opinion on that, because I'm pretty sure it depends on what you'll be doing. I wouldn't be surprised if it was (possibly a lot) better performance-wise, but maybe you're okay with smaller models and/or have less data.
00:12:46 <tswett> I'm not sure if TF is really a feasible thing to use here.
00:12:59 <fizzie> They've got more stuff on the tutorials section than I was expecting at this point in time.
00:13:06 <tswett> Like... does TF have neural nets as a feature?
00:13:44 <fizzie> Well, I mean. Two of the tutorials are called "Convolutional Neural Networks" and "Recurrent Neural Networks".
00:13:45 <tswett> I guess it does have a whole lot of NN-related features.
00:14:00 <fizzie> Whether you call them a "feature" depends on the definition.
00:15:12 <fizzie> I mean, the general TF framework itself is not inherently specific to neural networks, but it's certainly supposed to be convenient for them (because that's what it's used for), and hopefully there will be "standard" recipes for them.
00:16:13 <fizzie> But the generality does mean using it might involve more than just specifying some inputs and outputs, yes.
00:16:55 <fizzie> (Possibly in the form of stealing from existing examples, though.)
00:22:58 <oerjan> <boily> bellonder|. have I asked you your approximate geographic coördinates and body weigh? <-- i'm going to go out on a limb and say that portmantelloing newbies is not the best way of ensuring a response hth
00:23:54 -!- J_Arcane has joined.
00:24:07 -!- tromp has quit (Remote host closed the connection).
00:24:15 <hppavilion[1]> I've seen him around here before, so I don't think he was a newbie
00:24:32 <oerjan> but did bender| even notice he was addressed, there?
00:25:12 <fizzie> Well, that might be a bit of much to ask, given the sort of thing it is. There's the Docker-based approach. Just as a note of caution, though, TF might really only make sense if you actually do expect to do big things.
00:27:23 <fizzie> (And Docker on Windows does involve running everything in a VM, which you might not prefer either.)
00:34:01 <\oren\> I managed to generate one hangul!
00:34:26 <\oren\> (yes, 1 character. better than zero)
00:43:41 <boily> @tell bender| sorry for the porthello >_>'...
00:44:39 <FireFly> you're sorry for the https://en.wikipedia.org/wiki/Portello_(soft_drink) ?
00:45:53 <FireFly> hm, actually looks like that does not refer to the same drink as https://sv.wikipedia.org/wiki/Portello
00:56:50 <HackEgo> Headnotes are boring. Let's go fly kites!
00:57:33 <oerjan> shachaf: protip: `revert doesn't work to uncreate a file hth
00:58:19 <tswett> Hey everyone! What's a real-world, practical example of an inner join which returns more rows than either of its input tables?
01:04:31 -!- jaboja has quit (Ping timeout: 250 seconds).
01:07:22 -!- tromp has joined.
01:10:18 <boily> Firellofly. the porthello, with a hello in it.
01:11:00 -!- hppavilion[1] has quit (Ping timeout: 272 seconds).
01:11:33 <\oren\> tswett: any join between two tables, on a foreign key belonging to a third table
01:12:57 -!- MDude has joined.
01:15:18 <\oren\> For example, on a hockey_player table, you could join it to itself on "team" to form a list of all pairs of teammates
01:16:04 <\oren\> tswett: is that an ok example?
01:16:08 -!- hppavilion[1] has joined.
01:17:05 -!- mihow has quit (Quit: mihow).
01:19:58 -!- kragniz has quit (Ping timeout: 260 seconds).
01:23:07 -!- boily has quit (Quit: DISGUSTING CHICKEN).
01:24:10 -!- hppavilion[1] has quit (Ping timeout: 256 seconds).
02:12:52 -!- hppavilion[1] has joined.
02:21:50 <\oren\> Man, 28 final consonant clusters is a lot!
02:22:37 <\oren\> (techinally, 27, but I'm counting "none" as one for simplicity of programming)
02:25:20 -!- tromp has quit (Remote host closed the connection).
02:55:19 -!- Melvar has quit (Quit: system upgrade).
03:16:46 -!- Phantom_Hoover has quit (Read error: Connection reset by peer).
03:26:16 -!- tromp has joined.
03:30:17 <hppavilion[1]> Probably... one page, prefix operators. That sounds good. Second argument ignored if not filled out
03:30:26 -!- tromp has quit (Ping timeout: 240 seconds).
03:44:06 -!- bender| has joined.
03:48:13 <\oren\> I've generated 28 hangul!
03:52:27 -!- andrew has joined.
04:08:28 -!- andrew has quit (Remote host closed the connection).
04:11:20 -!- mauris has quit (Ping timeout: 256 seconds).
04:11:30 -!- andrew has joined.
04:18:10 -!- kragniz has joined.
04:22:54 -!- bender| has quit (Quit: Ping Pong Fuckout).
04:30:07 -!- mauris has joined.
04:30:20 -!- mauris has quit (Read error: Connection reset by peer).
04:36:39 -!- andrew has quit (Remote host closed the connection).
04:50:55 <\oren\> I have invented the filename extension .dat.c for a C source file containing nothing but constant value declarations
04:58:44 <\oren\> Each initial needs 5 sizes, each medial 2, and the finals only one
05:02:03 -!- tromp has joined.
05:02:35 -!- kragniz has quit (Quit: leaving).
05:02:38 <\oren\> hmm, or maybe the finals should be squashed down wiht horizontal vowels?
05:04:59 -!- kragniz has joined.
05:06:08 <\oren\> Eh, i'll see how it looks then decide
05:24:13 <hppavilion[1]> I'm trying to make it so that when a form is submitted, it POSTs the data to /calc then replaces the content of a textarea in the form to the response of that post
05:26:17 <Sgeo> You could ... do something to prevent the form from truly submitting, then use XMLHttpRequest
05:26:30 <hppavilion[1]> I've got everything except one part working, it seems
05:26:37 <hppavilion[1]> It appears Jquery's $.post isn't calling the function as its third argument
05:26:43 <hppavilion[1]> I'm not getting any errors in the JS console though
05:26:55 <Sgeo> May I take a look?
05:28:42 <hppavilion[1]> Note that I'm doing copy-paste programming because this is currently more about the backend than the frontend
05:29:15 <Sgeo> Does Submitted ever get printed to the console?
05:31:05 <hppavilion[1]> So it's either the backend or some weird bug on the frontend
05:31:35 <hppavilion[1]> Sgeo: POST requests /can/ return responses, correct?
05:31:54 <Sgeo> POST requests can return data just like GET requests, I believe
05:31:59 <hppavilion[1]> Because the library I was using refused to work if a post method /didn't/ return data
05:32:41 <hppavilion[1]> OK, on the backend there's a line of python (which DEFINITELY gets executed because the print immediately preceeding it works) that says "return web.Response(text='Done!')"
05:32:53 <hppavilion[1]> Which you probably don't need to understand aiohttp to get
05:33:05 <Sgeo> Check the network tab
05:33:08 <Sgeo> Does a response come in?
05:33:36 <Sgeo> There should be a POST, which will have both the request and the response... or just the request if it failed
05:33:54 <hppavilion[1]> So my issue is that the function that is the third (2th) argument to $.post isn't getting called
05:34:13 <Sgeo> It won't get called if there is no response
05:34:52 <Sgeo> Do you see the POST request in the Network tab, at least?
05:35:50 <hppavilion[1]> I was in Net under console (or whatever ctrl+shift+k's default tab is)
05:36:28 <Sgeo> Also you're specifying it's json, I don't know what jQuery does if the result isn't json
05:36:34 <MDude> http://fav.me/d9jtxyb This fractal art thing I've been playing aroudn with in paint.net, does it resemble any more well known fractals?
05:36:50 <MDude> Since it seems pretty simple to make, so i would think it would already be made at some point.
05:37:14 <Sgeo> hppavilion[1], what's the type that the thing is returning? Just "Done!"? Then text
05:37:15 <hppavilion[1]> But the POST goes through, nothing comes back it appears
05:38:40 <hppavilion[1]> Sgeo: So the $('#commands').html(response); line isn't doing what I expected of it
05:44:14 -!- tromp has quit (Remote host closed the connection).
05:47:26 -!- hppavilion[1] has quit (Ping timeout: 240 seconds).
05:47:47 -!- hppavilion[1] has joined.
05:51:35 <hppavilion[1]> Just have to hook up some Neural Nets and Traineres
06:05:48 -!- hppavilion[2] has joined.
06:07:46 -!- hppavilion[1] has quit (Ping timeout: 240 seconds).
06:08:48 -!- andrew has joined.
06:15:53 <MDude> WHat are you working on?
06:22:08 -!- Melvar has joined.
06:24:54 -!- andrew has quit (Remote host closed the connection).
06:25:20 -!- andrew has joined.
06:43:30 -!- clog has quit (Ping timeout: 272 seconds).
06:45:11 -!- tromp has joined.
06:49:06 -!- oerjan has quit (Quit: leaving).
06:49:50 -!- tromp has quit (Ping timeout: 272 seconds).
06:53:37 <\oren\> http://www.orenwatson.be/allchars.htm now has some hangul (after the kanji, there's 224 of them)
06:54:42 <\oren\> I think my lines are too rounded maybe
06:55:20 -!- Welo has joined.
06:57:05 <zzo38> How to adjust the kerning of the fonts used in Firefox on Ubuntu?
06:58:25 <\oren\> CSS letter-spacing property?
06:59:35 -!- clog has joined.
07:00:21 <zzo38> No, I want to modify the font metric files globally across all programs that use them (although I don't have any other program that uses them as far as I know)
07:00:41 <izabera> i just broke weechat by trying to paste one of those charactersूृ
07:03:23 <\oren\> does that break everyone's chat?
07:04:05 <\oren\> zzo38: I suppose you'd have to use fontforge to do that
07:05:17 <\oren\> allow problems to be stated as stoies
07:06:01 <hppavilion[2]> Preferably, I would have no normal operators at all
07:06:33 <\oren\> ok how about "A right triangle has two right-angle sides as lengths 4 and 5. what is the triangle's area?"
07:07:08 <hppavilion[2]> \oren\: That still seems incredibly difficult to parse. Roughly on the level of "Solving english"
07:07:19 -!- MDude has changed nick to MDream.
07:08:17 <zzo38> Shouldn't there be a SQLite extension to edit font metrics?
07:08:31 <\oren\> how about "calculate the monthy payment for a 30 year mortgage on 1 million with an interest rate of 0.5 percent"
07:08:59 <hppavilion[2]> \oren\: I think you know the difference between reasonably solvable and borderline impossible problems.
07:08:59 <zzo38> There is virtual tables support.
07:09:24 <\oren\> hppavilion[2]: joke's on you! that particular one works in Hound!
07:09:36 -!- hppavilion[2] has changed nick to hppavilion[1].
07:09:49 -!- clog has quit (Ping timeout: 250 seconds).
07:10:19 <hppavilion[1]> Perhaps I should develop a simple DB software in python and document its development in a webpage
07:15:55 <zzo38> AmigaMML:XISYNTH::TeX:METAFONT
07:18:33 <\oren\> http://postimg.org/image/oxprvhumr/
07:32:26 <zzo38> \oren\: Whatever message you posted it does not break my chat, although, I am using xterm.
07:33:01 <zzo38> (And the IRC client knows nothing about Unicode; it just passes the data as is)
07:33:36 <hppavilion[1]> And a Javascript line $('.output-0.01').val(response);
07:34:37 <hppavilion[1]> I know it gets executed. I know that response is equal to "Done!". It doesn't update the input
07:37:28 <HackEgo> [wiki] [[Talk:Silly Emplosions]] http://esolangs.org/w/index.php?diff=45842&oldid=45800 * Zzo38 * (+97)
07:38:48 <HackEgo> [wiki] [[Talk:Silly Emplosions]] http://esolangs.org/w/index.php?diff=45843&oldid=45842 * Zzo38 * (+77)
07:40:23 <Sgeo> hppavilion[1], an <input>?
07:40:33 <Sgeo> I think that requires an attribute to be changed?
07:40:51 <Sgeo> Oh .val() is meant to work with <input> too
07:42:36 <hppavilion[1]> That's the code that isn't working. Almost identical to the code I sent earlier, but being adjusted so multiple interpreters can be used based on the id of the form
07:44:16 <Sgeo> Can you add a console.log($('#output-'+tn))
07:44:25 <Sgeo> To make sure it's retrieving the element properly?
07:46:19 <hppavilion[1]> Object { length: 0, prevObject: Object, context: HTMLDocument → 127.0.0.1, selector: "#output-0.01" }
07:46:28 <Sgeo> That's not getting it correctly
07:46:39 <Sgeo> length should be 1, I think
07:47:46 <Sgeo> console.log(tn)
07:47:47 <hppavilion[1]> (This is another just-hack-together-some-js project where the beauty lies on the backend
07:49:02 <Sgeo> Does $("#output-0.01") work in the console?
07:49:55 <hppavilion[1]> Object { length: 0, prevObject: Object, context: HTMLDocument → 127.0.0.1, selector: "#output-0.01" }
07:50:19 <Sgeo> What does $("textarea") give for length?
07:51:00 <hppavilion[1]> And interestingly, it lists <textarea#output-0.01>
07:52:02 <hppavilion[1]> And I just tried adding textarea to before #output in the thing and it didn't work
07:52:32 <hppavilion[1]> Imma try changing the text area output-0.01 to a class
07:52:58 <Sgeo> Trying it locally
07:53:07 <Sgeo> document.getElementById works
07:54:09 <Sgeo> To use jQuery functions on it, you'll need to wrap that call in $()
07:54:24 <Sgeo> Or go native, that's what I prefer >.>
07:54:54 <hppavilion[1]> Sgeo: I've done native JS before. I use them pretty much interchangably when I do do JS, really it just varies by how I feel.
07:55:39 <Sgeo> hppavilion[1], I figured it out
07:55:44 <Sgeo> Something to do with the.... OH
07:55:55 <Sgeo> That . in the number, means it's looking for a class
08:00:20 <hppavilion[1]> Sgeo: Now I just need to make it work for multiple forms with class="cmdwin" xD
08:02:06 <\oren\> o btw, I added 菜落葉薬表要覚観負貨貯費賞起路身軍転軽輪
08:06:53 <hppavilion[1]> I have it working such that everything is in order EXCEPT it always puts the output one time in the same box
08:07:06 <hppavilion[1]> I suppose I could just have a fixed-position box that follows you as you scroll...
08:08:42 <Sgeo> Event handler on all forms, that when invoked, get's the id and passes the id of the form's textarea to the event handler
08:09:12 <Sgeo> Although I guess you're doing a one form many textarea thing, which is possibly not ideal, what if multiple textareas are filled in
08:09:29 <Sgeo> Actually I have no diea what you're doing
08:14:22 <hppavilion[1]> Each form is the software in a different stage. The forms only appear after that part of the program has been designed and implemented.
08:15:20 <hppavilion[1]> "So next we're going to implement relational operators"
08:15:50 <hppavilion[1]> Then there's a form where you can enter commands and try it out
08:21:36 <hppavilion[1]> OK, progress can be viewed at http://206.174.0.58:81/
09:05:35 -!- Frooxius has quit (Quit: *bubbles away*).
09:06:03 -!- Frooxius has joined.
09:08:34 -!- clog has joined.
09:11:06 -!- J_Arcane has quit (Ping timeout: 256 seconds).
09:14:13 <b_jonas> \oren\: I dunno, the hangul with o in them don't look good in your font to me, because the vowel is touching the final too much.
09:16:09 -!- J_Arcane has joined.
09:40:53 <zzo38> See your opinion of what are many of my new Magic: the Gathering cards, including now I made up a few split cards too. So far there seem not the place to post user comments, except possibly on the IRC but then it isn't all store together at once
09:45:52 -!- tromp has joined.
09:50:12 -!- tromp has quit (Ping timeout: 256 seconds).
09:51:42 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: f2c: not found
09:54:15 <zzo38> Have you kept track of all of the recent comprehensive rules changes for Magic: the Gathering even if you do not play the game?
09:54:50 <b_jonas> zzo38: I tried to, but it turns out that I missed some, so I'll have to re-read the bulletins
09:55:39 <b_jonas> zzo38: in particular, I got very surprised that Daru Stinger had gotten errataed back a few years ago (technically that's an Oracle change, not a comp rules change)
09:58:06 <zzo38> They have all of the changes on Yawgatog too, which is what I checked; they have Oracle changes there too. It can be used when working with puzzles; I always include a rule datestamp on puzzles for this reason. Someone else told me they made a puzzle but then it stopped working because they changed the rules, but I already thought of this ahead of time, and yet it has not happened to me.
10:00:51 <zzo38> There is good reason why Daru Stinger should be changed back though, which is why it was changed back to the old version.
10:00:56 -!- andrew has quit (Remote host closed the connection).
10:04:34 -!- MDude has joined.
10:04:51 -!- Tod-Autojoined has joined.
10:05:05 -!- hppavilion[2] has joined.
10:05:20 -!- teuchter has joined.
10:05:29 -!- Froox has joined.
10:07:19 <b_jonas> zzo38: I preferred him as a Human. The other daru are still Human.
10:08:32 <b_jonas> zzo38: Making him non-human seems like a power errata, which they claim generally try to avoid unless the card would be really broken otherwise; even though they claim it was for easier playing because the reminder text on the printed card says "Soldier")
10:09:07 -!- atehwa_ has joined.
10:09:35 -!- kragniz_ has joined.
10:09:36 -!- Lord_of_- has joined.
10:09:54 <zzo38> I can see the problem though with the powerful of the card with amplify ability
10:09:55 <b_jonas> And it's not only the Daru from onslaught that are Human, but the Daru Clerics from Legions.
10:09:56 -!- FreeFull_ has joined.
10:10:19 <b_jonas> zzo38: it may be more powerful than originally intended, but it's not really powerful compared to M:tG these days.
10:10:33 <zzo38> It could be fixed by printing that card with a different picture, is also some way
10:10:55 -!- hppavilion[1] has quit (Read error: Connection reset by peer).
10:10:55 -!- clog has quit (Ping timeout: 240 seconds).
10:10:56 -!- EgoBot has quit (Ping timeout: 240 seconds).
10:10:56 -!- \oren\ has quit (Ping timeout: 240 seconds).
10:10:56 -!- Lord_of_Life has quit (Ping timeout: 240 seconds).
10:10:56 -!- Melvar has quit (Ping timeout: 240 seconds).
10:10:56 -!- FreeFull has quit (Ping timeout: 240 seconds).
10:10:57 -!- Frooxius has quit (Ping timeout: 240 seconds).
10:10:57 -!- spiette has quit (Ping timeout: 240 seconds).
10:10:57 -!- MDream has quit (Ping timeout: 240 seconds).
10:10:57 -!- choochter has quit (Ping timeout: 240 seconds).
10:10:57 -!- atehwa has quit (Ping timeout: 240 seconds).
10:10:58 -!- kragniz has quit (Ping timeout: 240 seconds).
10:10:58 -!- Todpunk has quit (Ping timeout: 240 seconds).
10:10:58 -!- EgoBot has joined.
10:10:58 -!- \oren\ has joined.
10:11:09 -!- Melvar` has joined.
10:11:25 <zzo38> (So that no subtype other than "Soldier" would fit then?)
10:12:20 <b_jonas> zzo38: it's still called Daru, and it would be strange anyway to have a white card to which no racial subtype fits
10:13:15 <zzo38> Isn't there one in one of the more modern sets though? (I think the set is Phyrexia?)
10:14:03 <zzo38> A white creature card to which no racial subtype fits.
10:14:20 <b_jonas> yes, there are, but they're rare
10:14:45 <b_jonas> there's the old black Nameless Race of course, which lost its subtypes at Lorwyn
10:15:08 -!- hppavilion[2] has quit (Ping timeout: 256 seconds).
10:15:14 -!- ocharles__ has quit (Ping timeout: 260 seconds).
10:15:34 <zzo38> Yes I know, that one now has no subtype at all, but it is an old card and is black; I meant a more modern card.
10:15:48 <b_jonas> then there are black zombies and skeletons, some of which have a race subtype, and some don't, in what seems to me like an inconsistent way
10:17:06 -!- \oren\_ has joined.
10:17:11 <zzo38> If the bones do not form any other creature then the other one does not seem applicable, can be one thing, but that is clearly not always the case
10:18:12 -!- ocharles__ has joined.
10:18:18 <b_jonas> zzo38: but look at Legions of Lim-Dul, who by the flavor text are corpses animated by Lim-Dul the necromancer from the bodies of the Order of the White Shield, and the Order of the White Shield is a Human Knight.
10:18:34 <b_jonas> yet the Legions are not Human.
10:18:35 -!- EgoBot has quit (Ping timeout: 240 seconds).
10:18:36 -!- \oren\ has quit (Ping timeout: 240 seconds).
10:19:03 -!- EgoBot has joined.
10:19:21 <b_jonas> (This is the leader of the Order of the White Shield who is better known these days from Luminesce)
10:21:10 <b_jonas> There's another Order from some other plane to confuse the matters though
10:22:09 <b_jonas> Order of the Stars, who carry prominent shields on the art and flavor text.
10:25:59 <zzo38> How is the Order of the Stars relevant to that though?
10:28:34 <b_jonas> zzo38: not really, it just confused me because it's an Order with white shields
10:29:01 <zzo38> Yes, but their shields have stars on it
10:29:19 <zzo38> They are also not a knight
10:30:00 <b_jonas> yes, and they're on Ravnica, whereas the Order of the White Shield is apparently on Dominaria or wherever Ice Age plays
10:33:16 -!- J_Arcane has quit (Ping timeout: 256 seconds).
10:36:14 -!- J_Arcane has joined.
10:49:51 -!- Phantom_Hoover has joined.
11:02:13 -!- J_Arcane has quit (Quit: ChatZilla 0.9.92-rdmsoft [XULRunner 35.0.1/20150122214805]).
11:11:20 -!- J_Arcane has joined.
11:47:32 <zzo38> I have made up the calculation on the paper to figure out 1+2+4+8+...=-1 and 1+4+16+64+...=-1/3 and 1-1+1-1+1-1+...=1/2 and 1/2+1/4+1/8+1/16+1/32+...=1
11:48:38 <zzo38> Some people said the first few is wrong because it is a divergent series, but the same method is in use for the convergent series too it seem like?
11:54:20 -!- AnotherTest has joined.
12:03:44 -!- clog has joined.
12:12:10 -!- Melvar` has changed nick to Melvar.
12:12:46 <int-e> Well, people are right; one usually relies on reordering terms in a series, which only works for absolutely convergent series... Otoh, f(x) = 1/(1-x) is the analytic continuation of f(x) = 1 + x + x^2 + ... beyond |x| < 1, so under some circumstances, you may get correct answers that way.
12:13:20 <lifthrasiir> in general, the sum of divergent series is less defined (i.e. not as definitive as 1+1) than the sum of convergent series
12:13:51 <lifthrasiir> 1+2+4+8+...=-1 is one way to assign the "sum" to such divergent series
12:23:02 <zzo38> It seem to me that in order to reorder terms it must be done like: a+b+c+d+...=(b+c+d+...)+a=(c+d+...)+a+b=(a+c+d+...)+b=b+a+c+d+... but this won't let you to reorder or regroup infinite sets of terms as the rules for addition don't do stuff like that.
12:25:24 <izabera> is there a smart way to solve lights out?
12:26:38 <myname> well, there are boring algorithms that depend on knowing information for specific fiekd sizes
12:27:14 <int-e> izabera: sure; at its heart it's a system of linear equations over Z_2.
12:31:28 <int-e> Oh and the system is described by a band matrix (since no row (or column) depends only on decisions in the neighbouring rows (or columns), so for an n x m rectangle, rather than working with matrices of dimension nm x nm, one can quickly reduce that to n x n (or m x m, whichever is smaller).
12:32:18 <int-e> (for humans: once you've fixed the choices for the bottom row, all other choices are forced)
12:41:19 -!- Patashu has quit (Ping timeout: 250 seconds).
12:51:09 <izabera> i've read what to do on a 5x5 board, but what if it's a NxM
12:51:17 <izabera> and what if i need an optimal solution
13:08:17 <int-e> damn I had some code for this, somewhere... but probably gave it a meaningful filename like "t.py" (and I don't even remember whether it was python or some other programming language)... and quite possibly lost it
13:13:44 -!- Lord_of_- has quit (Excess Flood).
13:15:06 -!- Lord_of_- has joined.
13:17:42 -!- mauris has joined.
13:56:21 <myname> does finding the solution has to be efficient?
14:00:06 <myname> basic idea: start toggling from line 2 on exactly where there is light on the line above it
14:00:14 <myname> remember where you toggled
14:00:38 <myname> that will result in a board with only light in the lowest line
14:02:59 <myname> now, for each light in the bottom row make a temporary board with only that light, drag it up the same way, you will end up with toggles you can do from up to down in the original board
14:03:23 <myname> add every toggle per position mod 2
14:04:34 -!- boily has joined.
14:16:08 <boily> first time running my washing machine on extra small, cold water, delicate cycle.
14:16:15 <boily> it makes weird noises.
14:17:55 -!- PinealGlandOptic has quit (Ping timeout: 260 seconds).
14:19:30 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45844&oldid=45841 * Luis Mendo * (+54) /* Specification */
14:30:00 -!- J_Arcane has quit (Ping timeout: 250 seconds).
14:49:14 -!- PinealGlandOptic has joined.
15:00:51 <quintopia> myname: just solve the linear equation, in the case the transition matrix is invertible. In the case it is only semi-invertible, you can solve the independent subspace
15:23:46 -!- AnotherTest has quit (Ping timeout: 240 seconds).
15:28:30 -!- boily has quit (Quit: IMPRECISE CHICKEN).
15:38:28 -!- J_Arcane has joined.
15:38:49 -!- AnotherTest has joined.
15:41:01 -!- gniourf has quit (Quit: Leaving).
15:43:32 -!- gniourf has joined.
15:55:14 <b_jonas> `learn Jander was killed, or deactivated permanently, depending on which side you ask.
15:55:18 <HackEgo> Learned 'jander': Jander was killed, or deactivated permanently, depending on which side you ask.
15:56:21 -!- kragniz_ has changed nick to kragniz.
15:56:40 <b_jonas> `learn Jander was murdered, or deactivated permanently, depending on which side you ask.
15:56:44 <HackEgo> Learned 'jander': Jander was murdered, or deactivated permanently, depending on which side you ask.
16:00:29 <b_jonas> `learn Loodun Antyok demonstrates how to play a lawful character the exact opposite way from the lawful stupid archetype.
16:00:33 <HackEgo> Learned 'loodun': Loodun Antyok demonstrates how to play a lawful character the exact opposite way from the lawful stupid archetype.
16:00:41 <HackEgo> Loodun Antyok demonstrates how to play a lawful character the exact opposite way from the lawful stupid archetype.
16:01:42 <int-e> Oh no, do Hamiltonian circuits have anything to to with the Formula One series?
16:02:02 <b_jonas> int-e: and quantum mechanics, and quaternions.
16:02:14 <b_jonas> int-e: they're all named of the same brilliant mathematician
16:03:19 <b_jonas> `learn Hari wears an identical suit every day. Or the same suit? The latter glitch would be a sign of adjustments on the Matrix.
16:03:22 <HackEgo> Learned 'hari': Hari wears an identical suit every day. Or the same suit? The latter glitch would be a sign of adjustments on the Matrix.
16:03:52 <b_jonas> `learn Eto is the ageless laughing first minister.
16:03:56 <HackEgo> Learned 'eto': Eto is the ageless laughing first minister.
16:05:10 <int-e> b_jonas: I think you missed the bad joke.
16:08:29 <int-e> Unless I'm missing a connection between William Rowan and Lewis Carl Davidson... hmm.
16:11:54 -!- tromp has joined.
16:30:42 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45845&oldid=45844 * Luis Mendo * (-3343) /* Specification */
16:32:12 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45846&oldid=45845 * Luis Mendo * (-415) /* Compiler */
16:33:53 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45847&oldid=45846 * Luis Mendo * (+1) /* Compiler */
16:34:34 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45848&oldid=45847 * Luis Mendo * (+11) /* Specification */
16:34:48 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45849&oldid=45848 * Luis Mendo * (+9) /* Compiler */
16:35:50 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45850&oldid=45849 * Luis Mendo * (+6) /* Specification */
16:36:04 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45851&oldid=45850 * Luis Mendo * (+0) /* Compiler */
16:36:21 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45852&oldid=45851 * Luis Mendo * (+0) /* Compiler */
16:37:34 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45853&oldid=45852 * Luis Mendo * (-1) /* Introduction */
16:37:52 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45854&oldid=45853 * Luis Mendo * (-6) /* Specification */
16:39:12 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45855&oldid=45854 * Luis Mendo * (+25)
16:54:47 -!- Lord_of_- has changed nick to Lord_of_Life.
16:54:52 -!- Lord_of_Life has quit (Changing host).
16:54:52 -!- Lord_of_Life has joined.
16:54:53 -!- Lord_of_Life has quit (Changing host).
16:54:53 -!- Lord_of_Life has joined.
16:57:58 <HackEgo> [wiki] [[Hello world program in esoteric languages]] http://esolangs.org/w/index.php?diff=45856&oldid=45566 * Luis Mendo * (+32) Added entry for MATL language
16:59:44 <HackEgo> [wiki] [[User:Luis Mendo]] N http://esolangs.org/w/index.php?oldid=45857 * Luis Mendo * (+101) Created page with "I have created the esolang [[MATL]] (2015), a language based on MATLAB and suitable for code golfing."
17:02:17 <HackEgo> [wiki] [[User:Luis Mendo]] http://esolangs.org/w/index.php?diff=45858&oldid=45857 * Luis Mendo * (+38)
17:03:52 -!- tromp has quit (Remote host closed the connection).
17:11:03 -!- tromp has joined.
17:11:19 -!- mauris_ has joined.
17:14:46 -!- mauris has quit (Ping timeout: 240 seconds).
17:19:54 <\oren\_> Ok, so I'm making the finals more sqiashed downward, so that the horizontal medials will fit
17:20:53 <b_jonas> \oren\_: instead, move the medials so they touch in the direction where they're smaller
17:21:40 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45859&oldid=45855 * Luis Mendo * (+29) /* Specification */
17:21:44 <b_jonas> \oren\_: that is, "o" and "yo" should be able to touch some initials, whereas "u" and "yu" should be able to touch the finals
17:22:03 <HackEgo> [wiki] [[MATL]] http://esolangs.org/w/index.php?diff=45860&oldid=45859 * Luis Mendo * (+29) /* Compiler */
17:22:31 <b_jonas> \oren\_: I also tried to comment on the Deseret letters on your font:
17:22:50 <b_jonas> \oren\_: the letter "s" (both upper and lower case) is mirrored from how it should be, I believe,
17:23:27 <b_jonas> \oren\_: and in the fontdemo.html, you should change the line breaks so that the letters take up 4 lines instead of 5, and each corresponding lower case letter is two lines below the upper case on
17:24:21 <b_jonas> \oren\_: but I could be wrong. I don't like Deseret.
17:25:16 <\oren\_> https://en.wikipedia.org/wiki/Deseret_alphabet#/media/File:Deseret_second_book.jpg <-- I think I used this amage as a reference
17:25:29 <b_jonas> `perl -ebinmode STDOUT,"encoding(utf8)"; print "\x{1041d}"
17:27:14 <b_jonas> what's with your terminal?
17:27:18 <b_jonas> `perl -ebinmode STDOUT,"encoding(utf8)"; print "\x{41d}"
17:27:31 -!- JesseH has joined.
17:27:39 <HackEgo> [U+041D CYRILLIC CAPITAL LETTER EN]
17:28:00 <\oren\_> I copied that from the deseret ess
17:28:32 <HackEgo> [U+002E FULL STOP] [U+002E FULL STOP]
17:28:57 <\oren\_> and on input it appears to replace the surrogate pairs with dots
17:29:47 <b_jonas> \oren\_: according to https://en.wikipedia.org/wiki/File:Deseret_capital_Es.svg , the letter s has the left end connected to the top loop, and the right end to the bottom loop
17:30:06 <b_jonas> \oren\_: in your font it sort of seems like the opposite, though it's hard to tell from those few pixels
17:31:43 <\oren\_> I see, I'll see what I can do
17:34:13 <\oren\_> oh its god damn tmux fucking up unicide
17:34:58 -!- Lymee has joined.
17:35:26 -!- jameseb- has joined.
17:35:38 <\oren\_> and this bug will be solved essentially between in 10 years and never
17:37:44 <b_jonas> \oren\_: use screen instead. (a recent version, not the fifteen year buggy old version that spawned the development of tmux as an alternative.)
17:40:33 -!- pikhq has quit (*.net *.split).
17:40:34 -!- Lymia has quit (*.net *.split).
17:40:34 -!- Hoolootwo has quit (*.net *.split).
17:40:34 -!- lambdabot has quit (*.net *.split).
17:40:36 -!- Taneb has quit (*.net *.split).
17:40:36 -!- jameseb has quit (*.net *.split).
17:40:37 -!- rodgort has quit (*.net *.split).
17:40:40 -!- Lymee has changed nick to Lymia.
17:40:49 -!- rodgort has joined.
17:44:17 -!- Hoolootwo has joined.
17:45:55 -!- Taneb has joined.
17:46:45 -!- lambdabot has joined.
17:55:36 -!- tromp has quit (Remote host closed the connection).
17:58:35 -!- pikhq has joined.
18:08:10 <zzo38> Unicode is especially terrible for any kind of grid-based display, but I thought of some way similar to what someone else has suggested, for "universal terminal-based character coding", although the width of a string is now a bit more complicated than just strlen, rather each byte has a width of 1 or 0 depending on the value of that byte, therefore allowing many more single-width non-ASCII characters.
18:09:48 <zzo38> The format is always a code byte followed by a bank byte, and there is no right-to-left, no complex scripts, etc. Therefore it cannot be use with Hebrew, although Japanese writing is possible.
18:10:35 -!- ais523 has joined.
18:11:14 <lifthrasiir> probably it is now time to consider ligatures...
18:13:41 -!- ais523 has changed nick to dr_ais523.
18:15:12 <dr_ais523> for anyone who's wondering what I've been up to esolang-wise, darcs clone http://nethack4.org/media/ssapmmn (not finished yet though)
18:15:48 <b_jonas> dr_ais523: zzo38 has been complaining about visual artifacts he'd been seeing in the pdf version of your thesis, with some pdf viewer
18:16:37 <dr_ais523> b_jonas: hmm, it's just standard pdflatex
18:17:26 <zzo38> dr_ais523: The lowercase Greek alphabets (and only lowercase Greek) is slanted to the left. I am using Firefox to view it.
18:17:46 <dr_ais523> zzo38: ah right, lowercase Greek is known to have issues in the default LaTeX configuration
18:26:34 -!- Welo has quit (Quit: Leaving).
18:42:49 -!- copumpkin has joined.
18:45:16 <tswett> zzo38: I dunno, I'd think you would want to do that by just adding metadata to each Unicode character specifying how many columns wide it is.
18:46:11 <dr_ais523> tswett: isn't that potentially context-dependent?
18:46:44 <tswett> Do you have an example?
18:47:20 <dr_ais523> fi is a well-known one that works in English
18:47:41 <dr_ais523> there are other languages where ligatures are much more important
18:48:14 <tswett> Well, in English, wouldn't "fi" always be two columns wide, and the ligature also two columns wide?
18:48:25 <dr_ais523> the ligature's one column wide in some monospaced fonts
18:48:55 <b_jonas> dr_ais523: lots of stupid pdfs
18:48:56 <zzo38> I think in 2600. I wanted to write a postcard to them to tell them not to do that
18:48:56 <dr_ais523> (Pokémon red/blue has a 's ligature that's one column wide in its monospaced font)
18:49:09 <b_jonas> where some software auto-translates "fi" to that ligature at render
18:49:12 <tswett> I saw a one-wide "fi" ligature in a Homestuck album notes.
18:49:31 <b_jonas> a 's ligature? interesting
18:49:52 <dr_ais523> presumably because the apostrophe looked ugly if you gave it a whole column to itself
18:50:03 <zzo38> Ligatures should be defined by font metrics instead of built-in to the software.
18:50:05 <b_jonas> dr_ais523: I thought it would be more like to save screen space
18:52:06 <int-e> hmm 160x144 pixels, not too shabby :)
18:52:20 <zzo38> Some games will put two letters on one 8x8 tile
18:56:14 -!- tromp has joined.
19:00:35 -!- tromp has quit (Ping timeout: 240 seconds).
19:08:32 -!- Tod-Autojoined has changed nick to Todpunk.
19:17:23 -!- jameseb- has changed nick to jameseb.
19:34:36 <HackEgo> [wiki] [[Pig]] http://esolangs.org/w/index.php?diff=45861&oldid=45278 * LegionMammal978 * (+958) created interpreter
19:44:16 <FireFly> the 'fi' ligature is one column wide for me
19:44:55 <FireFly> There's that widespread implementation of wcwidth that I think only uses Unicode blocks to define width, and going by that fi would be a single column I think
19:45:41 <FireFly> 'firefly' ← for your ligature width testing purposes
19:46:45 <FireFly> http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c the wcwdith implementation I meant
19:50:13 -!- FreeFull_ has changed nick to FreeFull.
19:57:53 -!- tromp has joined.
20:02:22 -!- tromp has quit (Ping timeout: 250 seconds).
20:10:35 -!- dr_ais523 has quit (Quit: dinnertime).
20:33:49 <int-e> finest fluffy waffle affinity
20:40:18 <HackEgo> [U+FB01 LATIN SMALL LIGATURE FI] [U+006E LATIN SMALL LETTER N] [U+0065 LATIN SMALL LETTER E] [U+FB06 LATIN SMALL LIGATURE ST]
20:40:40 -!- tswett has set topic: The international hub for esoteric KRF design and deployment. | finest fluffy waffle affinity | https://dl.dropboxusercontent.com/u/2023808/wisdom.pdf http://codu.org/logs/_esoteric/ http://tunes.org/~nef/logs/esoteric/?C=M;O=D | https://esolangs.org/.
20:43:00 <lifthrasiir> https://github.com/lifthrasiir/unison/commit/b7c5c43#diff-c37314f89a3481e2da775620e88abc72R526
20:54:23 <mauris_> lifthrasiir: this is a pretty font :>
20:54:52 <mauris_> i would get rid of the "off" dots in the braille pattern glyphs
21:02:56 <mauris_> lifthrasiir: fwiw, cyrillic lowercase be (U+0431 б) generally has an ascender; it looks sort of weird without one
21:12:48 <FireFly> What was the URL to the font demo, again?
21:13:03 <lifthrasiir> nowadays https://lifthrasiir.github.io/unison/live
21:14:08 -!- int-e has set topic: The international hub for esoteric KRF design and deployment. | Effi's finest fluffy waffles | https://dl.dropboxusercontent.com/u/2023808/wisdom.pdf http://codu.org/logs/_esoteric/ http://tunes.org/~nef/logs/esoteric/?C=M;O=D | https://esolangs.org/.
21:14:37 <int-e> (using a name is kind of cheating, but I like it better than "affinity")
21:18:27 <int-e> oh it's an anagram of firefly.
21:18:47 -!- dr_ais523 has joined.
21:19:06 <dr_ais523> FireFly: that's the unligatured version though
21:19:41 <dr_ais523> apparently there isn't an Fi ligature, at least on my compose key
21:20:16 <int-e> > text . sort $ "firefly" -- not sure... worth a tryt
21:20:21 <FireFly> eryfifl I'm guessing, but I can't be bothered to check if fi is actually before fl
21:27:32 <zzo38> The wcwidth function for my own format is easy; it is: x?x&32768?2:1:0 (Note that sequences that include control characters can have unpredictable widths though; we don't include those here.)
21:28:55 <zzo38> However you must convert between wide and multibyte formats; the multibyte encoding for 0-127 is one byte, for 256-65535 is two bytes in small-endian format; not all wide character numbers are in use though.
21:29:24 <FireFly> I think having full-width characters occupy two cells makes a lot of sense, considering single-character cells are often roughly 2:1 height/width
21:29:50 <FireFly> and one would want full-width CJK glyphs to be roughly square
21:30:11 <zzo38> To calculate the width of a non-control multibyte string, just count all bytes in range 32-255; each of those bytes takes up one cell while bytes 0-31 take up zero cells each.
21:30:52 <FireFly> So there's a correspondence between the length of the encoded codepoint and the rendered width of it?
21:32:46 <zzo38> Also, if the first byte of a character is 128-255 then exactly one more byte follows, if 0-127 then it is only one byte long and no other byte follows.
21:33:31 <FireFly> I don't know if I like tying the encoding and rendering like that, but it could be a nifty trick for old platforms
21:34:10 <zzo38> I think it is good for terminal encoding and grid-based displays. For typesetting it is bad though.
21:35:43 <zzo38> For typesetting, the font metric file should define all stuff about moving cursor, ligature, kerning, selecting glyphs, etc; and the character encoding should not do so.
21:39:45 -!- tromp has joined.
22:07:49 <lifthrasiir> 13147 characters so far (up from 12914 characters right after the addition of hangul syllables)
22:08:17 <lifthrasiir> probably I'll work on more latin-like characters next week
22:09:51 -!- tromp has quit (Remote host closed the connection).
22:15:15 -!- AnotherTest has quit (Quit: ZNC - http://znc.in).
22:16:15 -!- tromp has joined.
22:22:00 -!- Patashu has joined.
22:51:40 <mauris_> now i'm trying to look for a shorter phrase that has all of the ligatures in the topic in it
22:52:07 <mauris_> "stiffly fix buffi fluff"! ~hardly a phrase~
22:53:39 -!- dr_ais523 has quit.
22:54:39 -!- dr_ais523 has joined.
22:55:48 -!- jaboja has joined.
22:59:10 -!- JesseH has quit (Ping timeout: 250 seconds).
23:03:37 -!- tromp has quit (Remote host closed the connection).
23:07:18 <dr_ais523> fizzie: is the wiki sending emails correctly? I've been trying to change my email address associated with my account to ais523@nethack4.org but it isn't sending the confirmation email
23:07:35 <dr_ais523> or if it is, it's getting lost somewhere which isn't my spam folder unless the email has a very strange from address
23:10:04 <dr_ais523> huh, AWIB was apparently created for a course named "DD2464 - Bigger Advanced, Individual Course in Computer Science"
23:10:26 <dr_ais523> also the AWIB docs consistently use the "brainfuck has a lowercase b except at the start of a sentence" rule
23:11:18 <shachaf> dr_ais523: congratulations
23:11:34 <dr_ais523> did you read the thesis, btw? if not, some of my other online friends beat you to it
23:11:38 <shachaf> i didn't end up reading your thesis on time :'(
23:11:50 <dr_ais523> oh wow, and it forms the possessive of "Müller" as "Müllers", which is correct in German
23:12:04 <dr_ais523> I have a suspicion that this project was an expert-level trolling attempt
23:13:35 <dr_ais523> meanwhile, it strikes me that ssapmmn might be quite good at optimizing brainfuck too
23:13:39 -!- dr_ais523 has quit.
23:13:51 -!- dr_ais523 has joined.
23:14:01 <dr_ais523> after all, minsky machines are basically bignum brainfuck confined to balanced loops only
23:14:19 <dr_ais523> although you'd have to somehow analyze the code to convert it into balanced-loops first, which is likely nontrivial
23:15:10 <dr_ais523> although I've also been focusing a lot on "if you would produce a negative number, produce zero instead" arithmetic
23:15:22 <dr_ais523> which is a) better for implementation in most esolangs, and b) really confusing, especially if you're used to normal arithmetic
23:15:40 <dr_ais523> x-1+1 is not equivalent to x, for example
23:17:34 -!- dr_ais523 has quit (Client Quit).
23:18:35 -!- dr_ais523 has joined.
23:18:58 <dr_ais523> it figured out how to write a modulus even without me telling it that the modulus operation existed
23:19:10 <dr_ais523> (using code of the form ((x / y) * y) - x)
23:19:39 -!- dr_ais523 has quit (Client Quit).
23:19:51 -!- dr_ais523 has joined.
23:20:44 <zzo38> If such optimization if it can be worked with brainfuck can also be made to work with LLVM then it can also optimize brainfuck program compiled into a LLVM code and possibly other programs too, since some program might have a few similar features
23:22:47 <dr_ais523> zzo38: well ssapmmn uses an SSA representation internally, which would be easier to target at LLVM IR than to most other languages (including C)
23:22:55 <dr_ais523> the main problem is that LLVM doesn't straightforwardly have bignums
23:25:50 <zzo38> I know those thing, although possibly if modification can be done to work even without bignums
23:26:33 <dr_ais523> well, I'm planning to infer which counters don't actually need to be bignum because they have a fixed maximum
23:26:44 <dr_ais523> but you need some bignums to be able to do infinite storage (although often they can be replaced with stacks or tapes)
23:26:50 <quintopia> dr_ais523: i take it this indicates a successful defense?
23:30:56 <dr_ais523> quintopia: it's what I'm currently workign on, it isn't finished yet
23:31:04 <dr_ais523> but it's basically a hyper-optimized minsky machine compiler
23:31:14 <dr_ais523> http://esolangs.org/wiki/Portable_Minsky_Machine_Notation is what it takes as input
23:31:37 <dr_ais523> my goal with the output is to be able to run an Underlambda interpreter, written in PMMN, with a good level of performance
23:31:51 <quintopia> which thesis were you talking about?
23:32:16 -!- dr_ais523 has quit.
23:32:28 -!- dr_ais523 has joined.
23:32:32 <dr_ais523> quintopia: http://etheses.bham.ac.uk/6120/1/Smith15PhD.pdf
23:33:29 <quintopia> ... and what were you talking about bf and mullers?
23:34:15 <fizzie> dr_ais523: It may be broken. I have vague recollections of it being broken earlier.
23:34:20 <shachaf> Is there a name for f(x) = exp(abs(log(x)))?
23:34:31 -!- dr_ais523 has quit (Client Quit).
23:34:53 -!- dr_ais523 has joined.
23:35:02 <dr_ais523> shachaf: wow, that's quite the piecewise function
23:35:14 <dr_ais523> it's, what, f(x) = x for x>1, and f(x) = 1/x for x<1?
23:35:28 <shachaf> Hmm, I didn't think to graph it.
23:35:30 <quintopia> so many things to read while i'm away from internet
23:36:20 <dr_ais523> I'm not sure why you'd name that function
23:36:21 <shachaf> If you're e.g. measuring a ratio between two things but you don't care about which direction it's in.
23:36:35 <shachaf> Well, it's a multiplicative analogue of abs.
23:38:08 <quintopia> oh i remeber a physics problem where the answer was either 19 or 1/19 depending on whether the heavier weight was on the right
23:40:43 <quintopia> i was talking to a dude who wants to invent "the most meta language ever" and seemed to count difficulty of implementing anything more usable in it as a measuring stick. he literally applauded at 3sp and figuratively cried that it would have a BCT implementation
23:41:34 <dr_ais523> quintopia: show him The Amnesiac From Minsk level 4
23:41:51 <dr_ais523> 3SP is, sadly, easier than I thought it would be
23:42:56 <dr_ais523> huh, the esolangs.org page is now the #3 entry on duckduckgo for a search for "three star programmer"
23:43:04 <fizzie> dr_ais523: There are some email messages stuck in the mail queue for the address you mention.
23:43:16 <fizzie> One of them has been there for 58 hours.
23:43:24 <dr_ais523> what's the reason they're stuck? or isn't it clear
23:43:40 <dr_ais523> I know my mailserver's working because I received a reply to that address to an email I sent from that address a few hours ago
23:43:52 <fizzie> It's not clear from plain "mailq"; I'll see if I can get something more verbose out of it. I'm not that familiar with Exim, I'm more of a Postfix person.
23:43:55 <mauris_> nice, "The particular approach taken here was a consequence of trying to optimize a Three Star Programmer implementation ..." which was "created when, while working on another esolang, ..."
23:44:03 <mauris_> need a tree visualization here
23:44:35 <HackEgo> nooooooooodl is the correct spelling
23:45:13 <dr_ais523> mauris_: fwiw the original esolang that sparked all this off was meant to have analog memory
23:45:24 <dr_ais523> but I couldn't pin the details down
23:45:33 <dr_ais523> I ended up inventing 3SP along the way though
23:45:40 <dr_ais523> which really rekindled my esolangs interest, mostly in the minsky directoin
23:46:34 <dr_ais523> also I'm surprised that there hasn't been much work into writing minsky machine programs
23:46:41 <fizzie> dr_ais523: 2015-12-12 18:42:32 ais523@nethack4.org R=dnslookup T=remote_smtp defer (-44): SMTP error from remote mail server after RCPT TO:<ais523@nethack4.org>: host nethack4.org [185.26.125.102]: 450 4.1.8 <codu@c1105889-5122.cloudatcost.com>: Sender address rejected: Domain not found
23:47:05 <fizzie> dr_ais523: I think it's trying to send messages with an envelope-sender of codu@c1105889-5122.cloudatcost.com and that hostname doesn't resolve.
23:47:24 <fizzie> Possibly a stale name in some configuration file.
23:47:26 <dr_ais523> ah right, so my server's refusing to receive messages from a server that doesn't exist
23:47:35 <dr_ais523> "codu" strongly implies a Gregor connection, normally
23:48:13 <fizzie> Seems that /etc/mailname is set to that.
23:49:16 <fizzie> It probably used to be a working name for the host at some point in time.
23:49:47 <fizzie> Annnnnd, in fact, it's what the IP reverse-resolves to, still.
23:50:14 <fizzie> But it doesn't forward-resolve back to the IP. So I think it's a cloudatcost screw-up.
23:50:41 <dr_ais523> maybe they just don't want people sending emails
23:51:13 <fizzie> Pointing the reverse DNS entry to a non-working name would be a weird way of blocking that.
23:51:34 -!- dr_ais523 has quit.
23:51:46 -!- dr_ais523 has joined.
23:51:55 <dr_ais523> it'd be like using -fPIE in order to prevent NetHack 4 using shared libraries
23:52:01 <dr_ais523> which is a workaround that someone I know actually uses
23:52:46 <fizzie> I could change mailname to some actual DNS name that points at the right place, although I'm a bit worried some mail servers (and/or spam filtering rulesets) might treat that suboptimally, because it would be a different name than what the IP reverse-resolves to.
23:52:55 <fizzie> Then again, it could hardly work less well than the current setup.
23:53:12 <fizzie> (On the third hand, they may fix the current setup at some point too.)
23:54:01 <dr_ais523> I guess I could request you to send me a copy of the email it's trying to send via some other means, so that I could confirm my email; OTOH having a confirmed email is mostly pointless if the wiki can't send to it
23:54:31 <dr_ais523> does cloudatcost allow you to set the rDNS for the IPs yourself?
23:54:53 <fizzie> I don't know. If it does, I wouldn't have access to that.
23:55:39 -!- dr_ais523 has quit (Client Quit).
23:57:05 -!- dr_ais523 has joined.
23:58:50 <dr_ais523> it's weird having created so many languages recently that I forgot about some of them
23:59:42 <dr_ais523> nice to see that Three Star Programmer is catching on, though