« first day (2734 days earlier)      last day (2123 days later) » 
00:00 - 19:0019:00 - 00:00

12:02 AM
I think I just pushed the hackiest code I've ever written in my life.
setIndex and setVariable if anyone wants to look at them.
@Downgoat What language is this?
 
12:17 AM
VSL
 
Ah.
That code (the one i linked to) is hacky, right?
 
12:29 AM
It feels weird correcting a well known programming language researcher. github.com/plfa/plfa.github.io/issues/41
 
Eta reduction!
 
@Zacharý I'd claim eta-equivalence. I'm not sure if the system eta-contracts or eta-expands.
 
@Potato44 Sorry, lambda calculus on the mind. 1 is the identity function.
 
@Zacharý I'm used to 1 being the type of the empty empty tuple, but I have seen 1 for the identity function
 
@Potato44 Empty empty tuple? WTF?
Oh, set theory.
 
12:39 AM
@Zacharý The type with one element. called () in Haskell and roughly corresponds to void in C-like languages
 
@Potato44 You'd expect a FUNCTIONAL language would do it with FUNCTIONS, but types are cruel.
 
@Zacharý Wait, when you say 1 is the identity are you thinking of church numerals?
 
@Potato44 What else?
 
I was thinking someone was using 1 as a notation for the identity function
So when I said 1 was the empty tuple I was thinking about notation
 
12:54 AM
@Downgoat what do you mean
You couldn't compile V8, the JIT wouldn't work
although it would be very funny
 
@mods can someone unfreeze this pls
@Zacharý how
@Zacharý also pls link to line next time
 
The entire thing was just a massive hack since I couldn't assign due to copy-by-value semantics.
 
1:15 AM
@quartata as in I have some v8 bindings which I'd like to emscripten
 
elaborate
 
Emscripten this: github.com/vsl-lang/llvm-node so I can run in browser
 
@quartata wouldn't work?
 
@ASCII-only JIT is x86-specific
 
@Downgoat wouldn't it be faster to transpile to js
 
1:17 AM
oh @ASCII-only i have question you are biologer right
 
@Downgoat no :| i changed to cs
 
ono :|
 
because biology course has too much biology
 
@Downgoat This won't really do what you think it does
Emscripten has a completely different ABI, when you compile something it doesn't output JavaScript functions you can call
 
@ASCII-only i have really basic biology question: are cis-trans isomers a subest of structural isomers
 
1:18 AM
Now, you can take LLVM, and make bindings to it using Emscripten's own API
 
@quartata there's embind or something like that though?
 
Right
 
hmm I don't want to have to make two seperate backends :/
 
@Downgoat this is a chemistry question
 
@Downgoat googles hmm apparently it's a stereoisomer?
 
1:19 AM
@quartata idk it's in my ap bio textboko
 
@Downgoat link
 
and to answer your question no
@Downgoat really? AP Bio has stepped up their game
 
@quartata there's a physics section in textbook too but i doubt that's on the actual exam
 
@Downgoat yep
@Downgoat well you don't really have a choice I'm afraid
Embind is very easy to work with though
 
1:24 AM
@Downgoat Sounds like the Calc textbook we used: it had Calc III & some linear algebra in it, even though it was an AP Calc textbook
 
@Zacharý This is fairly typical
 
I know.
 
If your math department is nice, they'll pick 1 textbook that can get you through to vector calc
Personally it always bugged me the BC had no multivariate calc at all
(if only so that you can get credits, since there's a 50% chance you'll see most of it anyways if your teacher is good)
 
I wish they still had Comp Sci AB,
 
Comp Sci A is a joke.
I assume most people here that would be interested have already seen this, but:
(cw: flashing images in 60 fps)
 
1:35 AM
@quartata still only 25% of people get a 5
 
Doesn't make it any good. It's practically a test of how good a linter you are, and most people aren't.
That's why we make good compilers
 
uhhhh you don't get any points off for things like spelling errors, semicolons, braces, indentation etc.
 
perhaps linter was a bad choice of words
I'm not referring to the free response section
I mean when it comes to analyzing code on the MC
 
@Downgoat Why on earth WOULD they count off for indentation?!
@quartata Yeah, 5 without the course and barely reviewing Java's OOP
 
1:50 AM
@Zacharý if you used handwritten tabs
@quartata oh my god i have PTSD from the "what does this loop" output
though that's basically 100% of HS comp sci, which is really more declarative programming rather than computer science
 
@Downgoat Like I said: wish AB was still a thing: it had O notation, as it focused a little bit more on the theoretical side (data structures & O notation)
 
@Zacharý it had low enrollment which is kinda understandable given that 90% of prospective devs today want to either be frontend React people or game devs
 
I know why it got removed.
I probably would've actually taken a class for that, or try to take it online (instead of self-study like CSA)
 
2:09 AM
@Zacharý *Try (to Take) It Online!
 
LoL
 
Computer science is the best thing in the world to learn by yourself. Who needs a class?
 
@quartata for some reason this was in my recommended so i watched it
 
@Downgoat That number almost exactly represents the difference between the Software Engineering and web design/gamedev enrolment at my uni. Apparently when they introduced the latter two, the current trend of only 10-12 software engineers started.
 
@feersum this is me >_>
 
2:10 AM
What is?
 
learning cs by myself. (well, i'm doing CS at uni now, but I already know most of it)
 
Well this is a good place to learn CS by yourself, together.
 
@feersum I still would've learned most of it beforehand. It's just that it would be challenging enough to warrant getting a somewhat free block out of)
Regardless, somehow I got a 5 on the AP Physics C: Mechanics without the course. I have no clue how tho, maybe it was that Mercedes-Benz logo rolling up a hill?
 
2:25 AM
@feersum A lot of languages have classes. Especially the OOP ones
 
2:46 AM
LOL
 
3:19 AM
Funky2 can now do multidimensional breaks. For reasons.
 
@ASCII-only It's very good if you understand what's going on
I feel like the explanation at the beginning of the video is quite bad although I like the HUD
 
@quartata there's an explanation video on the same channel
i watched it after
@ATaco breaks?
 
So if you nest loops, like for(i=0;i<10;i++){for(c=0;c<10;c++){if(c==3 && i==3){break break;}}} that will break out of both layers of loop.
 
break 2; in PHP
 
Please don't compare Funky2 to PHP, that makes me feel insecure.
 
3:31 AM
Just pointing it out
One of few languages that doesn't use labels for breaks
It's nice
 
You can also pass values to breaks.
 
I was doing some thinking about how to eliminate breaks and continues entirely in Pytek
 
So x = for(var i=0; i < 10; i++){if(i == 3){break i;}}
 
in Pytek, Jan 18 '17 at 3:36, by quartata
OK I think one instance of an unncessary break is in foreaches:
 
Oh hey, I forgot about continues.
 
3:32 AM
some messages there
C allowing assignments in the condition helps a lot IMO
I know most people don't like it
Go figured out how to make it look clean though
just make everything have a "pre-step", separate with a semicolon...
 
A lot of people don't like using assignments as normal expression and think they should be individual statements.
I say that's boring, and I'm going to allow For-loops in conditions.
 
hhahahahaha
but yeah assignments as expressions may look weird, but it also can be great, so, it;s impossible to say if its bad or not,
 
@ATaco that's not really multidimensional :/ also a lot of languages can do that
 
Well a multi-dimensional for loop can be broken out of entirely, which is what I was getting at.
 
@quartata it's great (on a side note, i'd prefer it the simple way, i.e. assignment is just a normal operator
@ATaco is multidimensional for loop a conventional term though
 
3:37 AM
¯\_(ツ)_/¯
I considered doing that for Funky2, but it opened up so many annoying implementation things.
Especially because it's possible to do a <operator>= b
 
@ATaco what's the problem here
 
If assignnment's an operator, than theoretically a == b could be double assignment, without explicit implementation to prevent that.
And for reasons it would be possible to do a +== b
 
new image is spilling over horizontally on mobile
someone who is good at the economy please help me my CSS is dying
 
@ATaco wat?
 
a (+=)= b, because += would also be an operator.
 
3:44 AM
= isn't unary?
 
I'm not implying it's unary.
 
Not a possible parse
 
<foo>= operators are separate operators...
they're all manually added in
 
Not in Funky!
 
3:44 AM
yeah y'see there's your problem
 
Have fun overloading that
 
you're already specialcasing =
 
It's a feature
 
welcome to copy semantics saturday
 
Funky being funky in implementation is literally in the name.
 
3:46 AM
@ATaco fair enough. the thing is though since you aren't aiming for it to be simple then there's no point in making = an operator
 
Yeah, I've no intention to operator-ify it.
But I did try, which is what I was saying.
 
@ATaco so yeah you can't really do this
 
It behaves similarly to an operator.
It even implements operator precedence.
 
@ATaco well that's kinda needed
 
The original Funky made no attempts at operator precedence, which was horrifying.
Worse still, it was R2L.
print(4 * 2 + 3) returns 20, against all logic.
The same program in Funky2 does return 11.
 
3:53 AM
been a while since ive written one of these tweet things. just a small question. what is "prewriting" again
also, what is a pencil
 
o.O
 
im sorry, Im sorry, it's been a while
 
I should probably implement unary operators some time soon.
 
unary operators are simple
 
@ATaco :|
you what
 
3:58 AM
I know they're simple, I've just not done it yet.
 
unary operators are just dyads, that had one of their arms guillotined of in the Revolution
implement them similarly
 
I'll probably not do that.
 
its how Clang does it
 
The good thing about Unary operators, is there's none of that "Which side's metamethod do I have to use"
 
see thats a cultural thing actually
if you read from right to left...?
 
4:00 AM
Well, Funky2 does some Funky things regarding selection.
 
just like Manga
 
By default, it will go left to right, but if the right variable has a user defined metamethod, it will prefer that.
There's also some nonsense regarding better defined metamethods.
 
what if its not commutative
 
Elaborate.
 
@ATaco ew.
just do it the c# way
have overloads specify types of both sides
 
4:03 AM
Don't worry, you can do that.
 
(or does funky already kinda do that)
 
"add[side=left,left=string,right=number]" = (l, r)=>...
And I'm sure it's more horrifyingly implemented than you think it is.
str["lt[side=left,left=string,right=string]"] = new VarFunction(dat => dat.num_args[0].asString().data.CompareTo(dat.num_args[1].asString().data) == -1 ? 1 : 0); is how the < operator is implemented internally for strings.
 
@ATaco ok wtf
 
@ASCII-only Notice the lowercase method names
 
(I've been writing too much JS)
 
4:08 AM
@Pavel hmm?
 
.asString()
 
@ASCII-only The C# code has method names starting with lowercase letters
 
camelCase is how I was taught to code, it's a hard habit to get out of.
 
@Pavel wait. it's c#?
@ATaco :|
excuse me
 
Atleast it isn't snake_case
 
4:11 AM
i use capitalization correctly in python, js and c#
 
@ATaco Technically public members should be upper to, so .data is also wrong. Also, idk if it already is, but there should be no public fields only public properties.
 
well except for _underscoreCase in which case just why does this even exist :|
 
I like _underscoreCase for very under-the-hood stuff.
 
0
A: Sandbox for Proposed Challenges

OMᗺImplementing a stack Tags: code-golf, stack, data-structures Can't believe we don't have this already.. It's one of the most important data-structures, yet still simple enough to implement it in a code-golf, if it's too easy you need to pick another language ;) Challenge Your task is to imp...

 
@ASCII-only The only language I don't know how to capitalize is C++, since all the standard library functions are named things like strncpy with no capitalization of any kind
 
4:12 AM
I have no idea how to capitalize PHP
But neither does PHP
So I think I'm in the clear.
 
@ATaco That's for private fields
 
But I welcome anyone who wishes to clean up that nightmare of code.
 
@Pavel C++ is usually more often pascal cased, although that may be changing...
 
@Pavel isn't that C
 
C++ and C are, unsurprisingly, rather similar
 
4:16 AM
Honestly how you style C++ really just tells you where that person first worked...
 
Oh, I’m using snake_case in my latest C adventure
 
pFor nInstance pIf nYou vSee nThis
:P
 
I know That style
I occasionally see it in Lua code
 
@quartata what would prepositions/pronouns be labelled
 
4:19 AM
Apologies to @Pavel's dad
 
@quartata or mVariable
(i think that's right?)
 
@ASCII-only That's System Hungarian not Apps
 
“My language isn’t very good at defining local variables” case
 
Or maybe CBackend and C_Frontend. I wonder who did that
Hm
 
@quartata hmm?
 
4:20 AM
Or alternatively if you're SGI and you want to piss_everyone_off_for_2_decades_template
 
@quartata if you mean that one, that's like the only use of hungarian in the entire style guide though, i think
which is even worse than the fact that it's hungarian
 
Most of MS doesn't use hungarian notation anymore
It's generally incredibly inconsistent, really
 
Apps Hungarian isn't supposed to be consistent
anyways point is it can say a lot about you
 
4:42 AM
@ATaco Are they just random values, or labels?
 
Breaks are as label as anything gets in Funky.
But their implementation is very similar to Return, hence they can pass a value.
 
I mean, do they break from a certain definite label, or is it relative to the current position?
 
I don't actually know how to answer that.
Because break doesn't actually jump anywhere, it asks execution nicely to stop until it reaches something that can catch the break.
So when a block is being processed, if after a statement there is a return or a break on the return stack, it stops execution and passes back the value associated with it. However, it doesn't remove it from the stack. Loops themselves are what remove it from the stack.
 
 
1 hour later…
5:54 AM
Well, @Adám, indexed assignment is now finished in RAD. It looks like it's closer to TIO status than the Python version was. Just a few more atoms, dfns/program arguments, and some other stuff then it's ready.
Oh, and reading a program from a file. That's definitely the next thing on a list.
 
6:15 AM
I'm seriously starting to think dd is broken
I just made two flash drives entirely unusable and I'm not sure why
Like they don't even show up as devices anymore
I seem to have been able to create install media by opening the ISO I was trying to dd in Dolphin and just copying all the files over to the flash drive
 
6:35 AM
0
A: Sandbox for Proposed Challenges

alephalphaRemove surrounding zeroes of a 2d array English is not my native language. Please correct me if "surrounding" is not the correct word here. This is a 2-dimensional version of this question. Given a non-empty 2-dimensional array/matrix containing only non-negative integers: [0 0 0 0 0] [0 0 0 ...

 
7:06 AM
@Pavel Well, just reformat it...
So you're trying to create install media? What's wrong with it?
 
He wrote it over his main drive first iirc
 
@user202729 I can't reformat it if it's not showing up
@user202729 IDK. When dd finished, the drive was extremely hot (like painful to touch) and it would no longer show up anywhere on the system
 
fdisk -l?
 
Nothing
They just died
 
Eject and re-insert?
 
7:11 AM
I think they actually just overheated
 
7:25 AM
(o is sort in increasing order)
Somehow, this bug only got found just today
 
What do you think about this challenge? (sandbox, not mine, link to latest revision for people hate visiting the sandbox)
 
 
2 hours later…
9:17 AM
@Zacharý :| why did you port
 
3
Q: Comments in lambda submissions

JakobIn compiled non-esoteric languages a solution is essentially always either a full program, a method, or a lambda expression. In a full program submission, the use of comments is clearly allowed. It's not clear to me, however, whether comments should be permitted in lambda submissions, which must ...

 
10:30 AM
@Neil at first glance, no, i'll have a better look soonish
@Neil ok apparently it's a stupid mistake, will fix asap
 
ta
 
11:27 AM
Man, it's the second time I get +60 and then -60 on the same question because of serial voting..
 
woo, my first Windows 10 blue screen
 
@Neil :P on the other hand i had those every time i moved my laptop too quickly until the third time i sent it in for repair :P
 
11:54 AM
0
Q: Alternating Pattern

Kevin CruijssenIn a now deleted stackoverflow question, someone posted the following: Write a program or function to print alternating patterns in * and # based on a given integer n. Some examples: Input: n=1 Output: * Input n=5 Output: *#### ###** ***## ###** *#### Input: n=8 Output: *####### ######**...

 
12:10 PM
0
A: Sandbox for Proposed Challenges

Beta DecayMake a \$\LaTeX\$ calculator code-golf math Challenge Given a \$\LaTeX\$ expression as input, evaluate it and output the result. What needs to be supported Addition: + Subtraction: - Multiplication: \cdot or \times Division: / or \div Fractions: \frac{}{} Exponentiation: {}^{} Brackets/Pare...

 
 
1 hour later…
1:19 PM
Has anyone made any progress on the Blackhat challenge? I tried to solve it using mathematica but it couldn't detect the hat.
 
Just pick a state-of-the-art CNN and overfit it on the 20 example images
 
I think there are 40 example images, 20 truthy and 20 falsy
 
That sounds hard
 
not really
 
I'd begin by searching for specific pixel coordinates where the pixel has one value for truthy and another for falsy
 
1:31 PM
@EriktheOutgolfer I would technically call that hardcoding the answer
 
and how is that an objective reason to invalidate my answer?
it's a anyway, the challenge is defined by the test cases
 
Because the challenge says that you are not allowed to hardcode the answer
(I mean it says "should" but it's in a section called "rules" which is confusing)
 
yeah, it would probably receive a lot of, well, -1s
 
What I was doing was taking the comic and a picture of the hat and getting Mathematica could detect whether the hat was in the panel. Only problem is that Mathematica only works with exact matches it seems.
Also, a hardcoded solution would fail the tiebreak
 
Anonymous
@EriktheOutgolfer Because it would almost certainly fail on the mystery test cases
 
1:35 PM
wait, did such cases get added? I thought the zip files are linked
 
See the bottom of the question
 
There are way too few examples in the zip files to get a model to generalize well
You'd have to use loads of example generation trickery like deformations, scaling, etc., to get sufficiently many examples
 
@Fatalize There's the rest of xkcd.com to use though
 
oh, apparently some "mystery" cases were edited in as of revision 4
 
Those image recognition challenges would have no point if the test cases were known in advance
 
2:16 PM
Oh another github frontpage redesign
 
@Downgoat Do you still work on TeaScript?
 
@mınxomaτ i'm cringing right now as I click opt-in
andddd it's ass
Although slightly better than the previous re-design, which isn't saying much
 
It's literally everything I don't want to know, in an order that doesn't make any sense.
 
oh cool, we're dedicating an entire third to trending repositories now
SeemsGood
 
Trending repos? Why would you want to see random repos?
 
2:24 PM
you know what's really cool
when you scroll to the bottom of a sidebar
and then if you keep scrolling it scrolls the main compartment
its good
 
2:39 PM
That sounds fantastic, how do I opt in?
I'm not seeing any real issues, actually.
Use the element zapper to remove the trending bar, and it's a fair bit nicer.
 
pics?
 
3:09 PM
@ASCII-only I'm glad I did: the Python one was becoming a massive mess: Using tuples for ;, sets for variables...
And it's not really even a port: some things are massively changed (i.e. the numbers, which are totally different, this time they work for bases up to 36)
 
3:37 PM
0
Q: White Elephant Exchange

BeefsterIt's Christmas in July, so what better way to celebrate than a virtual white elephant gift exchange! For this King of the Hill challenge, you must create a bot that plays in a White Elephant exchange simulation, trying to get the highest-valued present it can. Game Rules The game will be play...

 
3:53 PM
0
Q: 3-category C++ class template puzzle

user2023370Can you create a C++ class template, (let's call it triple) with at least one type template parameter (let's call it T), that complies with the following requirements? triple<T>::value is a constant value equal to either 1, 2 or 3, where: The properties of T correspond with the following three ...

 
4:18 PM
@JonathanAllan your algorithm doesn't seem very likely for n = 1...
 
CMP: Should the order in which arguments are passed to a RAD program be ⍺ then ⍵, or ⍵ then ⍺?
 
Why do I always have tech problems at one in the morning when no one's on IRC
 
4:37 PM
@Pavel I have tech problems all throughout the day and no one's on IRC
 
<sarcasm> Have you tried turning it off then back on again? </sarcasm>
 
@Mayube I've found that both # suse and # fedora are really helpful during reasonable hours.
 
Reading a program from a file, basic assignment, indexed assignment, program arguments (⍺ and ⍵), and function evaluation by default (like Jelly) are now implemented in RAD.
 
4:57 PM
@Zacharý TIO when?
 
@Pavel After a few more atoms, proper display, and dfn/dops.
 
5:11 PM
TIL TF2 is packaged in the OpenSUSE repos
 
wait, what?
are you sure
 
Looks like it
I haven't tried installing it
 
that doesn't look like it
just looks like an article for workarounds
 
You're right, but it's linked to from a page that lists all the Games packages in the repos
Which is weird
 
@Zacharý ⍺ then ⍵
 
5:22 PM
@FrownyFrog Already implemented :D
 
But if there's only one argument, it kinda makes sense for it to be ⍵
Anything that's not 2⎕NQ#'GetCommandLineArgs' is fine by me though
 
what happens if I pass 1 argument? it's assigned to both?
 
@Pavel I might have something like that for more than 2 arguments.
@FrownyFrog No, then will give a name error
And direct functions (not direct operators) are now implemented!
 
@Zacharý Why not ⎕ARGV or something sane?
 
@Pavel Actually, that's a much better idea.
 
5:44 PM
@EriktheOutgolfer yeah. turns out it was a dead-end method too :) codegolf.stackexchange.com/a/169240/53748
 
that...looks unbeatable, like, minimal required logic
 
6:15 PM
I just realized that '' and are the same in RAD.
 
what does rankless even mean?
 
cmq: where do you all get your electricity. where do you find it. do you have a nice place to get it from.
Im not really satisfied with my current supply, looking for inspiration...
 
@quartata Generally I approach a wall and try to find a hole. Most of the time that hole has electricity in it.
 
is it DC
 
I actually know 0 about electrics so idk
 
6:27 PM
that sounds like a very technilgically advanced hole. are there a lot of them where you live
 
I just know I plug stuff in and it works.
 
i wish the holes the ants made gave me electricity
that would be rad
 
That's where I get most of my electricity from.
 
theres not a lot of heavy metals here. Its interesting.
wait,is this a game?? this isnt a game...
i get my electricity from moving a "permanent magnet" I found. i went out to an island to find it.
But it takes a lot of work...
i gotta go induce some more. later.
 
Electricity here mostly comes from hydro
 
6:41 PM
@quartata I get most of mine by breaking down glucose
 
@LuisfelipeDejesusMunoz no
 
@Pavel Same (at least when it comes to applications)
 
@Pavel Hmm
we have a lot of water here
@trichoplax are you a robot??
no, wait, other way around...
are you a human??
 
@FrownyFrog It means that everything is done with depth: 2 2⍴1 2 3 4 and (1 2)(3 4) mean the same thing in RAD (even though ⍺⍴⍵ isn't implemented yet). RAD's arrays are not like APL's or J's, they are most similar/equivalent to K and to non-array languages.
 
@Downgoat I forgot TeaScript even existed
Does it still work?
 
00:00 - 19:0019:00 - 00:00

« first day (2734 days earlier)      last day (2123 days later) »