00:08:01 -!- kolontaev has joined.
00:08:03 <kolontaev> Hi all. What parser is easier to write from scratch: top-down parser (LL) or LR (like what Yacc generates)?
00:12:54 <zzo38> I have done both, but, I don't know.
00:13:20 <pikhq> LL recursive descent is probably the most straightforward.
00:13:48 <kolontaev> Bjarne Stroustrup in his books writes LL parser for calculator, right?
00:14:36 <shachaf> If your compiler doesn't do tail-call elimination, how should you write a recursive descent parser?
00:14:52 <pikhq> If you're writing it entirely by hand, it really is the obvious straight-forward approach to doing a parser.
00:15:24 <pikhq> Though with some downsides, e.g. you can only recurse as much as your stack will let you.
00:15:38 <pikhq> (this is particularly a problem if you don't have TCO)
00:15:55 <shachaf> How bad is it to manually write the state machine?
00:16:02 <shachaf> I feel like it could be pretty bad but maybe it doesn't have to be.
00:16:06 <pikhq> Depends on the language, honestly.
00:16:14 <pikhq> Generally kinda unpleasant though.
00:16:40 <kolontaev> shachaf: is state machine to be implemented for both LL and LR parsers, for a simple calculator?
00:17:08 <b_jonas> write a yacc grammar, use ayacc to compile it
00:17:29 <shachaf> or just write the parser by hand
00:17:57 <pikhq> Oh, for a calculator? Yeah, a hand-rolled recursive descent parser isn't even that bad to write.
00:18:05 <pikhq> And will honestly be perfectly sufficient.
00:18:25 <zzo38> You could also use other parser generators, such as Lemon, if you don't like yacc and ayacc
00:19:05 <shachaf> You're certainly not going to get that good by using yacc.
00:19:38 <kolontaev> shachaf: what is wrong with it? Too old-school?
00:20:10 <shachaf> I mean, if you don't know how to do the thing by hand, and you want to know, you should learn how to do that first.
00:20:19 <shachaf> Even if you use a parser generator or something later.
00:30:20 -!- Sgeo_ has quit (Ping timeout: 258 seconds).
00:33:21 -!- ais523 has joined.
00:33:47 <ais523> kolontaev: IIRC LR parsers end up the same as the LL parser in cases where the grammar actually is LL
00:35:38 <b_jonas> um... what do you mean by "the same"?
00:35:58 <ais523> b_jonas: well, the basic difference is that with LR, you don't have to decide what you're parsing immediately as long as you know what comes next
00:36:16 <ais523> so you parse what's in front, then use that information to determine what the thing you just parsed was
00:36:38 <ais523> but if the grammar is LL then there's only one possibility, so the LR parser can be optimised into the LL parser via constant folding
00:37:07 <ais523> the state machine is different but if you're open-coding it as functions rather than using an interpreter, it comes to the same thing
00:38:37 <ais523> kolontaev: when you have a state machine but you just translate it into a higher-level language ignoring the fact that it's a state machine, you just write out what it does
00:38:46 <ais523> rather than using some sort of state machine interpreter
00:39:46 <ais523> ayacc, my own implementation of yacc, generates open-coded output (or at least tries to)
00:40:24 <kolontaev> ais523: is this any better than state-machine?
00:40:40 <ais523> it's a different way of writing the same thing; I think it /can/ be better, but isn't always
00:41:26 <kolontaev> ais523: maybe it can be good for architectures with expensive memory accesses... :)
00:42:23 <ais523> well, I'd expect an open-coded parser to be more friendly to the branch predictor, but potentially to cause more pressure on L1 cache
00:43:29 <b_jonas> ais523: none of which probably matters because the tree parser is almost never the bottleneck
00:44:36 <ais523> b_jonas: isn't the lexer the normal bottleneck to a non-optimising compile of a low-level language?
00:44:43 <ais523> purely because it has more data to deal with than any of the other parts of the compile
00:44:47 <ais523> (also it has to read from disk)
00:45:24 <b_jonas> ais523: reading and writing the disk may be the bottleneck, but that doesn't have much to do with the lexer and parser I think
00:45:38 <ais523> I'm increasingly thinking it makes sense to combine the lexer and parser (after all, they're both state machines)
00:45:54 <shachaf> Lexer and parser? I just met 'er!
00:45:55 <ais523> in which case, the technique used to implement the combined lexer/parser may well be very relevant
00:46:34 <kolontaev> am I right I can implement lexer using yacc/EBNF forms? And to get rid of lexer? (Which would be weird, of course.)
00:46:39 <shachaf> I'd like a language with really good coroutine support that lets you implement state machines easily and clearly.
00:46:54 <b_jonas> I don't think you can combine them. you get unbounded lookaheads if your input is the individual characters of an identifier. (or 255 tokens of lookahead if you limit your identifiers to 255 characters)
00:46:54 <ais523> kolontaev: yacc requires a lexer of some sort, but it can be as simple as getchar()
00:47:16 <kolontaev> ais523: yes, if lexer is just getchar(), I can do everything in yacc, right?
00:47:56 <ais523> b_jonas: only if you need to know what's beyond the identifier to disambiguate, otherwise you only need to look ahead the length of the longest keyword
00:48:23 <ais523> % expr: product | product '+' expr { $$ = $1 + $3 } product: atom | product '*' atom { $$ = $1 * $3 } atom: '1' { $$ = 1 } | '(' expr ')' { $$ = $2 }
00:48:39 <ais523> that's an example grammar I'm using to test out generated parsers
00:49:02 <ais523> the code ayacc generates for it looks like this: http://nethack4.org/pastebin/example-ayacc-output.c
00:49:30 <ais523> it actually contains the parser twice, once in debugging mode, once in non-debugging mode, because all the "are we in debugging mode?" checks really get in the way
00:49:57 <ais523> so look at yype_start_expr onwards to see what an open-coded parser can look like
00:50:24 <ais523> it should be possible to do better, all the faffing around with yyvalid is needed to comply with POSIX rules on lexer/parser interaction
00:50:40 <ais523> but normally you don't care about those, so that code could all be optimised out
00:50:45 <kolontaev> ais523: isn't close to how Bjarne Stroustrup his LL-calculator?
00:51:07 <kolontaev> ais523: isn't close to how Bjarne Stroustrup makes his LL-calculator?
00:51:14 <shachaf> Do you have any good examples of "solving a problem by reducing it to a seemingly-harder more general problem"?
00:51:16 <ais523> well, the generated code is pretty close to what a hand-coded LL-parser would look likee
00:51:39 <ais523> shachaf: all the relatively powerful esolangs that we implement things like The Waterfall Model in to prove them TC
00:51:51 <ais523> in that case, the more general problem probably /is/ harder, it just happens to have already been solved
00:52:42 <shachaf> ais523: Sure, but that's not the same.
00:53:37 <b_jonas> shachaf: https://mathoverflow.net/a/92259/
00:54:06 <shachaf> One good example is Buffon's needle problem.
00:55:47 <b_jonas> oh yeah, some of those reductions to prove turing-completeness may count
00:58:31 -!- FreeFull has quit.
01:11:23 <shachaf> Buffon's noodle is too good.
01:12:25 <kmc> what's that
01:12:44 <shachaf> Someone gave me another example: Can you cover a 2^n by 2^n chess board with one of the center squares removed with L-shaped tri-minoes?
01:13:16 <shachaf> kmc: http://blog.sigfpe.com/2009/10/buffons-needle-easy-way.html
01:16:58 <ais523> shachaf: well, you obviously can for a 4×4 board
01:17:07 <ais523> which bothers me a bit, that sort of question is normally only asked when the answer is "no"
01:17:54 <ais523> you even more obviously can for 2×2
01:18:07 <b_jonas> ais523: you can do it for a 8x8 too
01:19:14 <ais523> 1×1 doesn't really fit IMO, there will normally be 4 centre squares but a 1×1 board only has 1 centre square
01:19:22 <ais523> so the operationn of removing one of the centre squares works differently
01:19:53 <ais523> can you do it for a ½ by ½ board? my guess is no, regardless of what square you try to remove you'll be left with -¾ of a square, which isn't divisible by 3
01:20:31 <shachaf> Of course n is a natural number.
01:22:37 <ais523> here's a nice inductive solution: you can solve a 2×2 board with one *corner* removed; if you can solve an n×n board with one corner removed, you can solve a 2n×2n board with one corner removed (via replicating the solution four times, three times with the missing corner in a centre and once with the missing corner in a corner, then placing the L-shaped piece in the centre)
01:23:05 <ais523> then all you have to do is reverse one quadrant of the solutions generated this way to get the missing square in the centre rather than a corner
01:23:21 <shachaf> Yep, that's the solution I came up with.
01:23:28 <ais523> so it's possible for all powers of 2 from 2¹×2¹ upwards
01:23:38 <b_jonas> ais523: oh nice. that's not how I'd have done it, but it works
01:23:41 <ais523> I don't see that solution as reducing the problem to a harder one, though
01:23:48 <shachaf> The "easier more general problem" is to solve a board with an arbitrary square removed, rather than one of the center squares.
01:24:03 <ais523> the "missing corner" problem is easier because you can just induce directly, rather than needing to go via an intermediate step
01:24:53 <ais523> once you can see how to do it with a missing corner, you can then see how to do it with any specific square missing, but I didn't think of that until you pointed out the "arbitrary square removed" version
01:26:33 <shachaf> The "arbitrary square" solution is the same as the missing corner solution.
01:26:45 <ais523> yes, just the inductive hypothesis is different
01:27:53 <b_jonas> recursion on power of two sized squares suddenly reminds me of that math contest problem that I solved
01:30:22 <ais523> more fun: if you can solve an a×a board and a b×b board, you can also solve an ab×ab board
01:30:37 <ais523> (same proof, different inductive hypothesis)
01:30:46 <b_jonas> darn it, that means I'm officially supposed to be good at this. but I'm not, I was just lucky with that one]
01:30:53 <ais523> 3×3 is not solvable but 5×5 is
01:31:39 <ais523> or, hmm, you can omit the centre square in 5×5 but you'd need to prove an arbitrary square is omissible
01:32:21 <ais523> and I'm not sure the corners can be
01:32:22 <esowiki> [[Teg]] M https://esolangs.org/w/index.php?diff=65632&oldid=61553 * A * (-4) /* 99 bottles of beer */
01:45:02 -!- xkapastel has quit (Quit: Connection closed for inactivity).
01:56:02 <b_jonas> there's this style of precedence parser that's like an LL parser but with a precedence value passed around,
02:18:05 <zzo38> I found that the S4 and S3 signals on the 8086 processor are used to determine which segment register is in use.
02:19:27 <kmc> signals? like external pins?
02:55:53 <zzo38> kolontaev: Yes, although I mean specifically the S4 and S3
02:56:06 <kolontaev> zzo38: ah, googled: http://ece-research.unm.edu/jimp/310/slides/8086_chipset.html
02:59:06 <kolontaev> zzo38: heh, maybe you can connect different physical memory banks for each segment register...
02:59:37 <kolontaev> zzo38: like 1Mb for each of 4 segment registers?
03:00:34 <zzo38> Yes, that is what I thought; I don't know if any computer is actually wired like that though
03:08:08 <kolontaev> many retrocomputing geeks dwells here: http://www.classiccmp.org/cctalk.html they may know...
03:09:37 <Hooloovo0> old PC stuff is in #oldx86hardware but low level 8086 questions are probably more at home in classiccmp
03:10:13 <zzo38> I think they should do NNTP instead of a mailing list like that (or as an additional interface to the mailing list).
03:13:16 <kolontaev> zzo38: NNTP is pretty old-school. And you need a client for it. While mail is still used by many people around.
03:17:45 <zzo38> Still, it would make it easier and more convenient to manage (both posting and reading), I think.
03:18:11 <zzo38> (Also, I wrote a new NNTP client software recently, and a bit less recently, also a NNTP server software.)
03:20:07 <kolontaev> zzo38: it wasn't you who run gopher server couple of years ago?..
03:21:42 <zzo38> I still have a gopher server. (Although some other people have set up a gopher server, too.)
03:38:21 -!- adu has joined.
04:09:50 <esowiki> [[User talk:A]] M https://esolangs.org/w/index.php?diff=65633&oldid=65606 * A * (+257)
04:10:10 <esowiki> [[User talk:A]] M https://esolangs.org/w/index.php?diff=65634&oldid=65633 * A * (+10) /* Golfed "Are there any mods here"? */
04:25:59 -!- Sgeo has joined.
05:18:39 <zzo38> If you distribute Usenet archives or mailing list archives as mbox files, then please ensure that any lines with "From " at the beginning are properly quoted. Do so by: sed 's/^>*From />&/'
05:20:09 <zzo38> (I have downloaded one that does not do any such quoting. And, some email software also uses improper quoting, such as 's/^From />&/' instead of 's/^>*From />&/'.)
05:21:42 <zzo38> (Either that, or don't use mbox format.)
06:02:24 -!- kolontaev has quit (Quit: leaving).
07:25:22 -!- tromp has joined.
07:46:23 -!- rain1 has joined.
08:06:38 -!- AnotherTest has joined.
08:16:18 -!- ais523 has quit (Quit: quit).
08:33:30 -!- Lord_of_Life_ has joined.
08:34:17 -!- Lord_of_Life has quit (Ping timeout: 268 seconds).
08:36:23 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
09:12:22 -!- atslash has joined.
09:16:57 -!- atslash has quit (Ping timeout: 244 seconds).
09:17:14 -!- atslash has joined.
09:28:29 -!- atslash has quit (Quit: Leaving).
09:30:46 -!- tromp has quit (Remote host closed the connection).
09:31:06 -!- atslash has joined.
09:34:38 -!- iczero has quit (Remote host closed the connection).
09:35:36 -!- iczero has joined.
09:44:59 -!- tromp has joined.
10:01:38 -!- tromp has quit (Remote host closed the connection).
10:02:24 -!- tromp has joined.
10:25:17 -!- shikhin has quit (Quit: Quittin'.).
10:26:30 -!- shikhin has joined.
11:07:15 -!- tromp has quit (Remote host closed the connection).
11:11:25 -!- tromp has joined.
11:11:52 -!- FreeFull has joined.
11:15:56 -!- kolontaev has joined.
11:21:51 <b_jonas> ``` perl -ne '/^([a-z].*bit\b)/ and print "${1}coin\n"' share/dict/12dicts/Lemmatized/2+2+3cmn.txt
11:21:52 <HackEso> ambitcoin \ cohabitcoin \ debitcoin \ exhibitcoin \ gambitcoin \ habitcoin \ inhabitcoin \ inhibitcoin \ orbitcoin \ prohibitcoin \ rabbitcoin \ tidbitcoin \ two-bitcoin
11:30:20 -!- Phantom_Hoover has joined.
11:33:20 -!- arseniiv has joined.
11:43:39 -!- kolontaev has quit (Quit: leaving).
12:00:05 <b_jonas> I'm going to hate these two weeks
12:23:28 <esowiki> [[User talk:A]] M https://esolangs.org/w/index.php?diff=65635&oldid=65634 * A * (+265) /* Golfed "Are there any mods here"? */
12:30:45 -!- tromp has quit (Remote host closed the connection).
14:01:55 -!- xkapastel has joined.
14:19:02 -!- tromp has joined.
14:50:55 -!- tromp_ has joined.
14:53:58 -!- tromp has quit (Ping timeout: 276 seconds).
17:12:09 <int-e> . o O ( In America, the week starts on the seventh day. )
17:28:14 <shachaf> Week starts on the seventh day?
17:34:56 <zzo38> And in Gregorian chant, the scale starts on the fifth note.
17:39:22 <b_jonas> shachaf: either Sunday or Saturday, depending on which numbering you take
17:42:42 <shachaf> But it sounds like int-e thinks that (a) the week starts on Monday and (b) people in America think the week starts on Sunday.
17:46:32 <arseniiv> once I saw somewhere that week seems to start on Sunday somewhere that calendar was made for
17:47:08 <arseniiv> in retrospect it would be not so consistent with the word weekend
17:48:00 <arseniiv> it would then need to be something like weekborder or weekrim
17:50:22 <arseniiv> hm it seems in at least both English and Russian “end”/“конец” mean not only a final point of something in time but also any furthest spatial points of something
17:51:00 <shachaf> The week certainly starts on Sunday.
17:51:11 <arseniiv> so it could hypothetically be applied to the starting moment also. It should look very strange though
17:51:12 <shachaf> int-e should know that, as a German speaker.
17:51:34 <shachaf> In Hebrew, Sun-Fri are named "firstday", "secondday", etc.
17:51:54 <shachaf> And the weekend is Fri-Sat so it all works out.
17:51:56 <arseniiv> also it would be strange to say something like “beginning” to designate furthest spatial points
17:52:26 <arseniiv> a peculiar thing, I bet there is a linguistic theory to explain it
17:58:33 <zzo38> The Germanic days of week are the Germanic gods, except Saturday. But in some languages they use Lyeday instead, which is also Germanic. But, if you replace "Saturday" with "Lyeday" and then abbreviate Thursday by Thorn instead of T, then you can have different alphabets of the days of the week.
17:58:43 <zzo38> (with no duplicates)
17:59:31 <esowiki> [[Flop]] https://esolangs.org/w/index.php?diff=65636&oldid=65629 * InfiniteDonuts * (+1343)
18:01:47 <b_jonas> I still vote for if you want to abbreviate the days of the week, call them W-1, W-2, W-3, W-4, W-5, W-6, W-7; and if you want to abbreviate months, call them 01, 02, 03, 04, 05, 06, 07, 08, 09, 10, 11, 12
18:02:36 <b_jonas> and until we make those exclusive, I wish calendars show both the month number and the traditional month name, rather than just one of them, because it's the job of a calendar to track that, not mine
18:02:47 <b_jonas> I still always get confused about what the name of month 08, 09, and 10 are
18:02:50 <esowiki> [[Flop]] https://esolangs.org/w/index.php?diff=65637&oldid=65636 * InfiniteDonuts * (+126)
18:02:55 <b_jonas> I got the rest of them down more or less
18:02:59 <HackEso> 2019-08-25 18:02:59.099879507 +0000 UTC August 25 Sunday 2019-W34-7
18:03:14 <b_jonas> it's W-7 and month 08 today
18:04:42 <esowiki> [[Flop]] https://esolangs.org/w/index.php?diff=65638&oldid=65637 * InfiniteDonuts * (+3)
18:06:02 <esowiki> [[Flop]] https://esolangs.org/w/index.php?diff=65639&oldid=65638 * InfiniteDonuts * (+13)
18:06:21 <zzo38> I do often use numbers for the months, but not for day of week (which I use the three letter abbreviation).
18:07:57 <esowiki> [[Flop]] https://esolangs.org/w/index.php?diff=65640&oldid=65639 * InfiniteDonuts * (+73)
18:09:48 <esowiki> [[Flop]] https://esolangs.org/w/index.php?diff=65641&oldid=65640 * InfiniteDonuts * (+346)
18:10:50 <esowiki> [[Language list]] https://esolangs.org/w/index.php?diff=65642&oldid=65576 * InfiniteDonuts * (+11) Added Flop!
18:12:34 <esowiki> [[Hello world program in esoteric languages]] https://esolangs.org/w/index.php?diff=65643&oldid=65567 * InfiniteDonuts * (+318)
18:12:48 <shachaf> zzo38: But German for Wednesday is "Mittwoch" which I thought means "midweek".
18:12:59 <arseniiv> b_jonas: I presume Hungarian month names aren’t Roman-induced? (Russian ones are, so it’s a blessing and maybe also a curse, though I was to slow-minded to point out an irregularity between october, november, december and 8, 9, 10 when I learnt their latin roots)
18:14:23 <esowiki> [[Flop]] https://esolangs.org/w/index.php?diff=65644&oldid=65641 * InfiniteDonuts * (+1)
18:14:51 <esowiki> [[Flop]] https://esolangs.org/w/index.php?diff=65645&oldid=65644 * InfiniteDonuts * (+0)
18:15:20 <myname> we start weeks at monday, nontheless
18:15:39 <myname> it's more like the middle of working week as middle of the actual week
18:16:24 <myname> i still prefer it over the american system where the week starts of mid weekend
18:16:26 <arseniiv> shachaf: oh nice, there’s analogous среда in Russian. Contemporarily it doesn’t easily parse as середина “the middle”, that what it should have meant a time ago, though it’s not yet that opaque. Though in wordplay and other matters it’s more often conflated with среда “environment”
18:17:02 <zzo38> shachaf: Maybe in German it is; I don't know German. But, I didn't say in German, but in Germanic, which is differently.
18:19:09 <zzo38> There are also two Germanic names for Tuesday, and German uses the other one from English, it look like.
18:19:13 <arseniiv> I know that in many Slavic languages which remained their traditional month names, they are in part incompatible between different languages, likely due to different climate, so leaves fall in one country at one month and in the other one at some other month
18:19:53 <myname> in german, there are two different words for saturday
18:20:53 <myname> the second and more unuusual one is roughly translated as sun-noon
18:21:12 <zzo38> Yes, I looked on Wikipedia, and neither of them seem to match either of the Germanic names.
18:32:04 <b_jonas> arseniiv: don't we already have that somewhere in the logs? in the Hungarian names for the d-o-w, Wednesday, Thursday and Friday are derived from the slavic names which in turn come from the numbering where Wednesday is number three, and Tuesday is numbered but not slavic but finno-ugric numeral.
18:32:20 <b_jonas> I think this came up somewhere already, feel free to search the channel logs
18:32:33 <arseniiv> shachaf: me personally maybe yes. In school when we were taught different formulae to solve quadratic equation I was angry for the case of even b, now I think it’s more natural one and other ones should be forgotten. Though that time it was taught not so coherent and maybe I was angry for a reason
18:33:45 <b_jonas> arseniiv: also https://www.tondering.dk/claus/cal/week.php
18:34:18 <b_jonas> oh wait, you are asking about the Month names?
18:35:56 <b_jonas> the month names of the gregorian calendar in Hungarian are more or less the same as the English and French ones, only spelled partly saner and partly less sane
18:36:28 <b_jonas> (month 06 and 07 are the less sane part; 05 is the saner part)
18:36:33 <arseniiv> b_jonas: hm for a split second I thought it was a page about weekdays in different languages and I remembered Janko Gorenc (I don’t know if he’s known here but almost every conlanger should know him as he makes himself known very actively, or was for some time, he’s collecting numerals, and even not as consistently as could then be used by anybody — no declension, no grammatical things at all, only 1…10 etc.)
18:37:04 <arseniiv> <b_jonas> oh wait, you are asking about the Month names? => yes
18:38:01 <b_jonas> numerals? like https://www.omniglot.com/language/numbers/index.htm
18:38:53 <arseniiv> b_jonas: I see, június and július. Here’s the same: июнь июль
18:39:34 <b_jonas> but aside from those, the English and French ones are way worse in summary
18:39:38 <arseniiv> there’s even a flat joke about an immigrant asking what’s the right one, июнь or июль
18:41:16 <b_jonas> the french ones are typical french ambiguous short stuff: "mars, avril, mai, juin, juillet, août"
18:41:51 <arseniiv> <b_jonas> numerals? like https://www.omniglot.com/language/numbers/index.htm => hm, I hoped Simon wouldn’t make the same mistake
18:41:57 <b_jonas> no wonder when they made the revolution calendar, they invented entirely new and long and nice-sounding names to compensate
18:42:34 <zzo38> English doesn't use Germanic days of month anymore either (although they did in Old English).
18:42:44 <arseniiv> surely I maybe would be able to remember what 5 is in some language, but how should I construct 555?
18:43:14 <b_jonas> arseniiv: the omniglot pages do include larger numbers (for many languages at least)
18:43:36 <b_jonas> but obviously it's not a full reference
18:43:43 <b_jonas> you can look up numbers in a dictionary if you want
18:43:51 <b_jonas> omniglot isn't meant to be one
18:44:53 <b_jonas> besides, numbers in Hungarian are easy, they only have six extra forms of declensions that ordinary nouns or adjectives don't have, and there's no fancy exceptional names for 11 or 12, it's all totally decimal except for that 10 and 20 have forms unrelated to 1 and 2
18:45:56 <arseniiv> <b_jonas> omniglot isn't meant to be one => I mean it would be nice to include a relevant grammar fragment if there are basic numerals already
18:46:46 <arseniiv> <b_jonas> they only have six extra forms of declensions that ordinary nouns or adjectives don't have => hm interesting
18:48:00 <arseniiv> ah maybe I misread what it should mean
18:48:09 <b_jonas> that part is serious. "kilenc" is the base cardinal form, meaning nine, the other forms are "kilencedik", "kilenced", "kilencedike", "kilencszer", "kilencedszer", "kilencrét". there are more, but you get them from these forms by ordinary noun declension suffixes.
18:48:29 <b_jonas> "kilencedik" is the ordinal, meaning ninth
18:48:38 <arseniiv> ah maybe I got it as intended then
18:50:49 <b_jonas> any noun declension will occur, because numerals can behave as ordinary adjectives, and adjectives can get declined as a noun as long as they're not used as a prefix epiteth of a noun phrase
18:52:17 <arseniiv> Russian has inherited some mess about cardinal numerals but it is no extra cases, it’s just that some numerals decline as nouns, other decline as adjectives and there are some others also, and this is completely transparent to native speakers, it’s only partly strange when one starts to learn linguistics at large, not even at a school level
18:52:34 <b_jonas> and since "kilencedik" and "kilenced" behave as nouns, they will occur in any noun declension too, plus "kilencedike" will occur in a few of them
18:53:25 <b_jonas> and this is true for other numbers too, even apart from "kilenced" having a certain fixed noun meaning that is obsolete now, meaning a tax that commoners payed to their landlords or something like that
18:54:41 <b_jonas> yes, out of the remaining nine tenth after they payed one tenth as tithe to the church. or backwards, I can't follow which is which.
19:10:04 -!- adu has quit (Quit: adu).
19:11:04 -!- xkapastel has quit (Quit: Connection closed for inactivity).
19:32:08 <LBPHacker> as a side note, I love this channel so far
19:34:39 <arseniiv> LBPHacker: just Hungarian? there were far more than it
19:34:56 <LBPHacker> arseniiv: yeah I'm reading the logs now
19:35:15 <arseniiv> how can you not detect sweet cyrillic letters I carefully typed on my keyboard :P
19:35:24 <LBPHacker> but Hungarian is one word my eyes are *very* trained to spot
19:55:29 <b_jonas> you could use computers to spot words. they're good at it.
19:59:27 <HackEso> arvomioksesi oikseni henkohattaan luulevanamme lähestyvyydyt pysyvemmistä komentuelliksemme ahdosta soletettävälimiin plastaskemiksempaili
20:00:20 <b_jonas> fizzie: there used to be some wisdom entries with hungarian text, but I don't think any of them remained
20:00:26 <b_jonas> I think I deleted all of them
20:29:36 <HackEso> 1/1:lion//Lions are the catamorphisms of the animal world. They get eaten by poets in stone dens. \ lifthrasiir//lifthrasiir is shunned by the rest of his country for being no good at League of Legends. \ learning//Learning is disco, baby. \ ghoti//“Ghoti” is a very fishy spelling. \ at//At is a daemon for procrastinating commands.
20:29:45 <arseniiv> though if that were me, I’d deleted my own too maybe
20:30:17 <HackEso> 1/2:206) <locks> who's walter bright and why is he so bright <nddrylliog> locks: he's to D what I'm to ooc <nddrylliog> locks: guilty \ 916) <augur> DIE <augur> oh hey elliott \ 1117) <ais523> nowadays I tend not to have a physical form at all though <zzo38> Interesting.............??? \ 1095) <metasepia> `addquote \item <elliott\_> `addquote <olsner> two quotes about quotes about django <olsner> I guess the worst part is that I appear in all three ha
20:30:20 <HackEso> 2/2:ckego quotes about django \\ <olsner> elliott\_: another quote? you're not helping \texttt{:/} ← and three giraffes. \ 1142) <ion> Homeopathic encryption: add 9 parts NULs to 1 part of data, shake well into three directions, repeat the process 30 to 100 times.
20:32:13 -!- Lord_of_Life_ has joined.
20:34:12 -!- Lord_of_Life has quit (Ping timeout: 245 seconds).
20:35:06 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
20:43:35 <HackEso> b_jonas egy nagyon titokzatos személy. Hollétéről egyelőre nem ismertek. He is often too busy with appeasing the M:tG gods to make any sense.
20:43:51 <b_jonas> that one does have hungarian words, and I wasn't even the one who created it
20:44:55 <b_jonas> arseniiv: I created lots of silly wisdoms, then later deleted many of them
20:45:12 <b_jonas> wisdom/b_jonas is probably the only one that contains hungarian words that I didn't create
20:47:19 <arseniiv> <b_jonas> hungarian words that I didn't create => I parse it that way now
20:48:03 <arseniiv> it implies you created some Hungarian words, probably every other one :D
20:48:50 <b_jonas> well, it's pretty likely that I created some Hungarian words, though probably not the ones in the wisdoms
20:49:02 <b_jonas> and it may be just one wisdom, wisdom/ha
20:49:09 <b_jonas> I thought there were a few other ones, but maybe no
21:02:32 -!- arseniiv has quit (Read error: Connection reset by peer).
21:03:00 -!- arseniiv has joined.
21:43:32 -!- tromp_ has quit (Remote host closed the connection).
21:45:28 -!- AnotherTest has quit (Ping timeout: 245 seconds).
21:59:14 -!- tromp has joined.
22:04:16 -!- tromp has quit (Ping timeout: 264 seconds).
23:12:09 -!- FreeFull has quit.
23:14:53 -!- atslash has quit (Quit: This computer has gone to sleep).
23:19:33 -!- b_jonas has quit (Quit: leaving).
23:26:23 -!- tromp has joined.
23:30:42 -!- tromp has quit (Ping timeout: 252 seconds).
23:45:21 -!- MDude has joined.
23:58:21 -!- Phantom_Hoover has quit (Quit: Leaving).