« first day (1512 days earlier)      last day (3629 days later) » 

00:05
If anyone wants to help invent a language inspired by Duodyadic Tiles, come to this chat:

 Expanding Duodyadic Tiles

This is a chatroom to discuss modifying my idea of duodyadic t...
3
 
1 hour later…
01:26
Why are the stars different?
The first one is pinned
The first one's pinned? I think
... every time.
 
10 hours later…
11:08
@aditsu Has anyone ever brought up ArrayList Character b?
@Sp3000 I don't think so
Might be useful for doing, say, 'db for base 100 :)
(doing the new Minecraft question)
11:48
@Sieg That doesn't seem to follow the output format any more
The OP said it's fine.
But the output is exactly correct if you run it with -t flag
I'm not sure that's what he meant since afterwards he explicitly said the formatting is only required up to input 9. So in that case you should probably count the -t flag.
Hmm... Okay.
@MartinBüttner I know you probably just want CH's confirmation, but axe, hoe and stairs are the only asymmetric recipes and indeed reflections work
(which counts 3 bytes, btw)
11:54
3 bytes?
space, -, t
that was still in reply to Sieg
I know, but the space counts too? Last time you said -v was 2 bytes for ><>
hm, I don't think so :D
let me look it up
Oh, I thought the consensus is 1 byte per flag?
11:58
At least I have seen that be the case often.
@Sieg Yeah a lot of people don't count them correctly, but the consensus on meta is 3. That being said if you already use a flag anyway to run your code (like ruby -e "puts 42") then it is only one byte, because you can usually just add it to the flag you already have
See "special invocations" here:
14
A: On "interactive" answers and other special conditions

J BMy basic guideline would be "as the problem statement says". Which is mostly to mean it's allowed to explicitly override anything it wants from what I'm about to say next. Actions to invoke For test-case/IO programs, I like it better when the whole invocation is possible as a single line, thro...

Well, the code is run like joe.py -t -c '\AR' or in repl using joe.py -repl -t
can you do joe.py -tc '\AR'?
No.
then I guess it's 3 by that meta post. (you might want to add that feature before your next submission then ;))
12:02
So it would be 1 if -tc worked?
Maybe I should just add a function that formats the output as a table :P
If it's 3 by that meta post I've got some ><> solutions to fix up...
12:04
I'm running out of characters though.
:<
@Sp3000 wait a sec
Maybe my decision to begin all function names with an uppercase letter wasn't so smart after all.
@Sp3000 he doesn't actually mention the space
@Sieg ?
I would prefer to use easily available characters.
12:07
that is very easily available. I found that after just 1 google search
also, once you know that ASCII code, its easy to type too ALT + ASCII CODE
On Mint I can't use Alt+code :<
So is the -t 2 or 3 bytes?
@Sieg I don't know what's real any more.
I'll take the power to my hands and say it's one.
well according to that meta post it's at least 2
Either way it makes me want to finish Golfish already :P
12:13
it doesn't mention the space though
Should I keep counting it as 2 then? Or should I fix up my existings to 3?
I find 3 more consistent even if the post doesn't mention spaces explicitly
the entire paragraph argues that you should count how many characters you need to add compared to an equivalent invocation without the flag
I find 3 to be mean.
I don't. some flags are incredibly powerful and can easily save 10 or 20 (like the implicit loop flags in ruby and perl)
it's weird though
you could also argue (in the example from that post) that an equivalent invocation is perl file.pl
(note the two spaces)
ah wait
12:17
"I count those as a difference in character count to the shortest equivalent invocation without them."
So... 3?
Fixing
btw, this is why Retina does all configuration as a prefix of the code, because flags are too expensive
That doesn't make much sense in a REPL enviroment though.
12:20
you could prefix your statements as well
which would essentially amount to defining a function I guess (provided the effect can be done as post processing)
Hmm...
The language is pretty much designed for that.
So can ASCII 128-239 be thought as 1 byte?
if you don't use anything outside the 255 range, yes
if you do you'll have to use UTF-8 or something which also makes the extended ASCII range more expensive
Well, there goes my problem of limited operators
I will clarify that meta post
You know, I was going to use ASCII 128+ for the Minecraft challenge but when CH said "in bytes" the UTF-8 counter was linked :/
12:30
Actually... For what exactly do I need more than 680 builtins?
Which I would get if I limited function names to two bytes
Wait, no. I would have 1 305 unique names.
I guess that's enough.
Well, maybe I should have a couple characters that lead to 3 byte sequences. Just to make sure.
Damnit: I'll have to rewrite the whole reference after this change. And all functions. Everything.
/monologue
13:04
Well I think you should rather ask yourself how many functions can you get with the printable code points ≤ 255. if that's enough there's no need to uses 2-byte characters at all
I couldn't find an easy way to type >128 characters, so it feels bad
Get yourself an APL keyboard and encode in that :P
That'll confuse a few people
This was meant to be readable too, in the first draft at least. Now not so much, so why not :P
[A-W][ -}]|[a-z] is a fun regex to match function names
You're missing ~
Or is that a special character?
13:21
Well many of them are in that range, but it should be matched there.
~ is after }
Still thinking about a way to add more functions to 1-byte range.
I know.
@Sieg different data types + overloading? but that probably doesn't make too much sense for an APL-like language
@MartinBüttner There are characters, numerals and lists.
that should give you a lot of room for overloading
13:26
List interactions are implicit so that leaves overloading for numbers and characters.
For example, 1+2 is 3, but 1 2 3+4 5 6 is 5 7 9 and 1+2 3 4 is 3 4 5
Can't really be overloaded.
true but likewise you probably have some operations that are only meaningful on lists?
In those the argument is made a list if necessary.
Dunno if overloading would be better.
But, for example N gives the length of a list, so N1 2 3 -> 3, but N1 -> 1, even though 1 is a number
Basically you never need to explicitly make a list.
I guess that make sense
but still you could overload characters and numbers
true
For example "abc"+"efg" makes no sense, because it gives ["ae", "bf", "cg"]
And those are treated as characters.
Maybe it could give the sum of character codes? :P
14:01
0
Q: Should tags be plural or singular?

Martin BüttnerAss part of The Great Tagging Survey, I've noticed that the pluralisation of our tags is a bit inconsistent. The following lists all tags that refer to some kind of object that is dealt with in the corresponding challenges. Singular numberstringgamesequenceregular-expressiondategridfractalfunct...

embarrassing typo... my S key is a bit stuck at times
Heh.
"Ass part of the great tagging survey"
truth is, this particular typo happens to me at least twice a day, but usually I notice before submitting... I think I should clean my keyboard
Probably the wrong thing to say in this context.
14:08
your mind is dirty
Still a teenager.
Urgh, I just got the terrible feeling that I overthought for the CH's Minecraft Kolmogorov :/ What's the bet that base encoding all possible outputs does better?
{"P.SW"=}% -> "P.SW"f=
Know your f's
Oh thanks
So `f maps a function to all items of a list?
14:16
yes
with an argument
@Sieg more accurately it curries a function and maps it onto a list
list argument f function
3f+ adds 3 to each list element
Figured out that much.
Hmm, I wonder if I want Joe to support user defined conjunctions (functions which return functions)... If I do, I have to just create the parser from scratch.
I could drop dependency to arpeggio if I did that.
Hmm...
... I think a simple base convert does better D:
... disappointed
14:26
how much do you have ?
90 or so
fence can be
.PP
.SS
.PP
also ?
Can't rotate
oh
only flip ?
and only horizontally
@Sp3000 in your current code, can you save something if you subtract 3 of the index, and move the [3 3 8] to the end? (ditching the two [1 0 0]s after it)
ah no that doesn't work... never mind :/
14:38
i thought about that, but I wasn't sure it'd save much
that being said, if your array has length 24, you don't need 24% in the code
Oh? Modulo indexing?
-3 bytes from both then, thanks :P
14:53
[A-W][ -~]|[a-z\!#%=\?+*<>;|\]-][,:]*
beautiful
no need to espace the ! and ? though
No. No it's not.
Oh, okay.
I wonder if the by-rows approach works better in another language. I'm sad it doesn't actually help in CJam...
I have got 90 with by rows
(Unless one or both parts can be shortened more? I dunno)
14:56
probably can reduce to 86
@Sp3000 what do you have without the by rows approach ?
I've posted 86, but I could get 85 (but that requires recalculating everything)
@Sp3000 Absolutely worth it.
Maybe tomorrow :P
Just store the data in unary.
15:19
Huh. I just noticed that these changes made no difference, but made the code less clear.
Goddamit.
those are the best kinds of changes
rewinds 3 hours back
16:21
Okay, my final decision is to default to the table output. Convert to string to get a list i.imgur.com/XJz0HG3.png
 
6 hours later…
22:17
oh, that daft question has been bumped
22:28
My internet just came back.
I went almost the whole day without internet, cable, or a landline.
The withdrawal was real.
I wonder how @Geobits is coping.
22:48
I've been reading about constructed languages recently.
(Languages for human use)
I came across a language called Lojban, which is meant to be as regular and unambiguous as possible.
I'm not a fan
timlan is unambiguous as far as our dictionary isn't awful
What is your relationship to timlan?
co-creator
22:52
That's interesting.
actually, that's a good point, we were going to re-write the parser in the coming weeks (University holiday)
The main goals of the Lojban language family were to aid human-computer interaction and to test the Sapir-Wholf hypothesis.
(Have you ever heard of it?)
yeah, I read up on it a while back
I wasn't especially impressed
my main complaints are the weird constraints on words, the phonology, and it was effort enough to find a sensible dictionary
And the fact that it uses periods everywhere.
oh, that too
23:02
I kinda want to try my hand at conlanging, but I'm deterred by creating the vocabulary, which I think is the most time-consuming and boring part.
yeah, we had so little vocab that we would try and add 10 words every day for a while, of course that fell flat on it's face, but we got some good puns out it tim32.org/blog/post.php?id=82
Well, I just finished reading through the Language Construction Kit website. You've probably heard of it.
I have not
we take the "if we knew how to do this properly it would be no fun" approach to such things
Here's the main page: zompist.com/kit.html
Probably the most extensive article is the one on grammar: zompist.com/kitgram.html
@MartinBüttner Status report: It's terrible. Geobits out.
2
23:14
@Geobits All the best!
Lojban has no punctuation, but some of the characters normally used
in punctuation affect the way Lojban is pronounced. A full stop
(period) is a short pause to stop words running into each other.
^ That is apparently why there are so many periods. I'm pretty sure, however, that spaces already fulfill that purpose.
I can't remember if we sentence separators or not (full stop), we don't need them, but I think we allow them or something, I know at least one parser pretends to recognize them (there are good reasons why we need to write a new parser)
@PhiNotPi it's just so that the pronunciation is clear, if I recall, it's a nice one to one mapping
we allow schwas at the end of any word to make a clear distinction, but you'd never write them
23:48
@MartinBüttner Could you please check that I banned enough things for Chinese Remainder Theorem to make it non-trivial in Mathematica and the like?
5
A: Sandbox for Proposed Challenges

xnorImplement the Chinese Remainder Theorem code-golf The Chinese Remainder Theorem tells us that we can always find a number that produces any required remainders under different prime moduli. Your goal is to write code to output such a number in polynomial time. Shortest code wins. For example, ...

23:59
0
A: Sandbox for Proposed Challenges

PhiNotPiEvolution of OEIS In this challenge, he goal is the recreate the Online Encyclopedia of Integer Sequences one sequence at a time. Similar to the Evolution of Hello World, each answer depends on the previous answer. There are a few ideas that have been floating around. Starting with OEIS A000...


« first day (1512 days earlier)      last day (3629 days later) »