« first day (2640 days earlier)      last day (2205 days later) » 

12:52 AM
Anyone here good with SQL
 
A little bit
Actually no I'm leaving soon so don't ask me
 
@ASCII-only All the SQL I know is Robert');DROP TABLE Students;--
 
Darn, 6 hours no answers...
 
Anonymous
1:14 AM
@Scrooble That's the exact chest lol. Hobby Lobby?
 
Anonymous
@ASCII-only Yes
 
nvm. i appear to have found a solution (or at least partly)
now it's just figuring out how to use something with not-so-great docs
 
@Mego Dunno. 'Twas my father that bought it. I was very young.
 
2:04 AM
A prominent person in the JS community just said this:
 
communist?
 
> If you're a complete beginner starting off in JavaScript, there is nothing wrong with just learning a framework up front. Don't listen to people that say you have to master fundamentals first. You have to master them eventually but you need to have fun first.
@LeakyNun autocorrect :P
 
Anonymous
@Downgoat And that's how you get crappy "front-end devs" that can't solve simple interview questions :P
 
Anonymous
Seriously though... We've had a bunch of people apply for dev positions at my company who can't solve the interview problem
 
@Downgoat who did
 
Anonymous
2:07 AM
It would make for a great PPCG challenge, and I would love to post it, if it didn't mean getting fired
 
just make it a CMC, no one will find out :^)
 
Anonymous
As a tangent to that: we're hiring, if any skilled devs in the Dallas area are looking for a job
 
too far for me :/
 
Anonymous
It would be weird to work with someone I know from PPCG, but having a reduced workload would make up for it
 
2:20 AM
I personally don't see any harm in learning something lightweight like Backbone
not stuff like Meteor or Angular
 
@quartata nothing wrong with learning Backbone but not knowing JSON is not a programming language is a problem
I know a couple people who have subscribed to this “everyone can become a programmer in 12 weeks with this course” but thinking through a problem and finding a solution is 90% of the job. You can’t do that if you don’t understand the fundamentals
 
Anonymous
@Downgoat "just write a program to determine if this other program terminates!" /s
 
@quartata the guy is actually a member of Angular core team
@Mego I have actually encountered this you have no idea
 
@Downgoat if you can think properly that's 100% of the job
 
Anonymous
@Downgoat I have also encountered this
 
2:26 AM
hmm, can't a halting machine exist as long as there are no circular references
 
@Downgoat Angular 2 legitimately hurt my brain and I have 2 trillion IQ
I can see how easy it would be to just fall into the "google for snippets and have no idea what you're doing" cycle with it
 
> 2 trillion IQ
But you are not a goat 🤔🤔
 
did I mention angular 2 also peforms like trash even when using dart
 
@Mego Good weird or bad weird? :P
 
@Downgoat what you're seeing now is my normal state
 
2:32 AM
@quartata I have worked with someone who literally copy pasted tensorflow code into a Javascript ajax callback and used made-up function names (we were doing sentiment analysis so apparently tf.angerLevel is now a thing). The problem was googled, but there was no understanding so the solution could not be applied
 
Anonymous
@DJMcMayhem Weird weird
 
Anonymous
That being said, I would absolutely get you an interview, DJ, if you wanted to move to Dallas :P
 
Technologies?
 
pHp, the H stands for Heck
 
Anonymous
@DJMcMayhem PHP, MySQL, JS, plus Sencha for mobile apps
 
2:35 AM
oh hey, Sencha
 
Anonymous
Also lots of sarcasm and a decent amount of goofing off and trolling support people
 
that... sounds like a nice place to work actually
 
Anonymous
@ASCII-only It's pretty great. I hate PHP with great fervour, but I deal with it because of all of the other perks
 
haven't heard that one in a long time. stopped using it after they tar bombed me that one time. still get their newsletters though
do you use extjs outside of it?
 
@Mego Hypothetically, if I did move to Dallas area, could you promise board game nights with you, CrazyBaci, me, and Icedagger? :P
 
Anonymous
2:37 AM
@DJMcMayhem Duh
 
Anonymous
If you moved to Dallas, we'd be at your place with our Mimic o' Munchkin in a heartbeat
 
Anonymous
@quartata No, we only use JS for frontend on our website and with Sencha/extjs in our mobile apps.
 
@Mego I just showed this to Icedagger, she said she'd probably steal it from you guys
 
Anonymous
@DJMcMayhem :( we worked hard on it, though
 
Anonymous
We could help you guys build one :P
 
2:41 AM
That might be a better solution :P
 
what would you put in it
 
I'd be tempted to put my guitar picks in it
 
not an Amulet of Yendor?
 
Am I allowed to have one if I'm still a NH scrub?
 
Anonymous
Or a smaller mimic that you can wear on your head
 
2:43 AM
hmm good point.
 
you just get a cheap plastic imitation of the amulet
 
alternatively, you could add googly eyes and line it with teeth (and optionally motorize the hinge)
@quartata this is genius idea
 
Anonymous
@ASCII-only It has teeth. We got googly eyes, but opted not to use them, so that it just looks like a normal chest from the outside
 
@Mego Or how about one of these?
 
@Mego does it actually
 
Anonymous
2:44 AM
@DJMcMayhem Brb spending money I don't have
 
Anonymous
 
@Mego wow. 12.8/10
 
2:59 AM
@ASCII-only why 12.8
> Fundamentals change over time, frameworks don’t.
Wow...
 
@Downgoat It's 128/100, and 128 is a nice round number
@Downgoat who... said that
 
Anonymous
@Downgoat That's the exact opposite of reality :P
 
Anonymous
A new JS framework pops up and becomes the new hotness every 6 months, and is abandoned in 2 years
 
Anonymous
Also I've solved my Rubik's cube about 10 times now and I'm very satisfied
 
Anonymous
I just need to memorize the movements from the second layer onwards
 
3:06 AM
@ASCII-only some guy. NPM CEO and the same Angular core dev team liked that tweet though
 
> NPM CEO
ok that's not surprising tbh
> Angular
not very surprising either :P
 
@Mego more like 3 months at this point. Vue is growing more popular by the day it is seeming
 
Anonymous
@Downgoat I thought it was older, since I recently heard of it, and I'm usually ~5 months behind the JS framework curve
 
Isn't Vue older than 6 months
 
Anonymous
4 years, apparently
 
3:12 AM
Yeah
 
my team started u sing it at least 6 months ago
but yeah, it's popularity is recent
 
I couldn't name any js frameworks other than jQuery and React but I recognize a few others.
 
Anonymous
jQuery, React, Angular, Vue... yep that's it for me
 
Anonymous
I'm not sure if jQuery counts as a framework, though
 
hmm, would templating engines count
 
3:16 AM
depends on their language :)
I've used Knockout
Oh, I've heard of Ember, too
though I think that that is backend
nope, SPA
 
@NathanMerrill ?
 
like, jinja doesn't count :)
but something like Handlebars would count as a js framework
 
and mustache/nunjucks (js edition of jinja)/pug?
 
Anonymous
@NathanMerrill I don't think so
 
Anonymous
Frameworks dictate how you write code. Handlebars is just a library
 
Anonymous
3:21 AM
Same for mustache, nunjucks, and the rest
 
Anonymous
Also Mustache sucks
 
0
A: Sandbox for Proposed Challenges

manassehkatzRebuild My Scrabble Board code-golf game board-game string My daughter and I have an ongoing Scrabble tournament. We enjoy admiring the board at the end of a game. But recently we knocked over the board and didn't get a chance to get a good look or take a picture. Fortunately, we kept track of ...

 
you write drastically different code with Handlebars
like, your HTML looks different, and you don't write all of the Jquery to manipulate DOM
 
Anonymous
@NathanMerrill By that logic, the JSX component of React is its own framework
 
I don't know React, but if that's the templating part, then sure
 
Anonymous
3:23 AM
It is, and I wouldn't consider it to be its own framework
 
Anonymous
Flask and Django are frameworks. Jinja2 and Django's templating engine are not frameworks - they're libraries that are part of the frameworks
 
eh. We're just debating about the meaning of a word.
 
Anonymous
Debating on the internet... I'll grab the swords
2
 
5:37 AM
Which finger do you type C (in qwerty layout) with?
 
@feersum properly, or personally?
 
You as in you
 
left index, usually
sometimes, my left thumb
 
yeah, left index for me. Never my left thumb
in fact...I don't use my left thumb for anything but hotkeys
 
I do the same as you guys.
I looked at a couple of tying websites and they said you should type it with the middle finger.
That seems wrong though.
 
5:40 AM
left index
 
Since for the right hand, they all agree, the fingers should move outward
As in towards the nearest edge of the keyboard.
But using middle finger would make it move inwards which seems pretty uncomfortable to me.
 
I find it weird that which fingers I use to type is very dynamic, but when I play rhythm games I would like to play with dynamic finger movement, but can only play with static finger positions.
 
Well, do you play rhythm games every day?
 
not recently, but at one point I did.
 
Which kind of controllers?
 
5:44 AM
keyboard, specifically a Cooler Master Quick Fire TK
Also, a drawing tablet, I played osu! before my windows install borked itself.
 
5 typing websites I checked recommended the middle finger to type C, but 4 real people agree index finger should be used, weird.
 
I never claimed I type properly, actually I know I don't.
 
Oh this is a regular computer keyboard
 
If you are going for high-speed, I agree: you should use the middle finger
 
Why can't you assign letters to the notes or whatever?
 
5:48 AM
the index finger is overloaded with possibilities
 
I have no idea how to use C#'s Delegates, but I think it's exactly what I need right now.
 
@feersum huh, what do you mean? I can do that in the keyboard based rhythm games like LR2 or osu's mania mode. But I find extremely difficult to play with dynamic finger movement, but I want to learn to do high key count modes like 14 key which require dynamic movement.
 
I guess it's harder if the stuff you type doesn't spell any familiar words.
 
@ATaco they are first class functions. What are you stuck on?
 
It's alright, I've wrapped my head around it now.
Wasn't sure how defining delegate classes(?) worked.
 
5:55 AM
@feersum I guess the hard part is actually learning how to translate the onscreen patterns to patterns of movement that are actually possible without getting my fingers tangled.
 
6:13 AM
@Potato44 0/10 no rhythm game should ever need more than 2 fingers
 
@ASCII-only All rhythm games that I can think of that are playable with 2 fingers are mobile games. or Dance Dance Revolution, which is played with the feet.
 
@Potato44 ok. 0 fingers: osu :P
 
@ASCII-only How are you going to play that without fingers. You need at least one to tap, plus however many you need to hold a mouse or tablet pen.
 
@Potato44 hmm that's true.
but if you're using a pen then you don't need an extra one to tap
@Potato44 also it's possible with one: a touchscreen laptop
 
Is it possible to make call functionality for a class, like lua's __call metamethod?
 
6:21 AM
@ASCII-only true that is possible, but that would get tiring very quickly.
 
@ATaco what language...
 
C#, I should have clarified.
 
@ATaco No
C# is not a scripting language
 
Boring, but acceptable.
 
although you can probably make a class inheriting from Action/Func
which might work
 
6:24 AM
Although calling is classified as an operator, it's not considered overridable.
Which is probably for the best.
 
yeah. but you can make a function object
 
7:03 AM
@Riker Just got the message that the same duck was caught again (unfortunately they didn't provide another picture).
 
ngn
@feersum Left middle finger. I learnt to use the proper digits for C and B only after I got me an ergonomic keyboard.
 
 
1 hour later…
8:09 AM
what if C, C++ etc had Mediawiki-like preprocessor directives like #includeonly and #noinclude
 
@betseg what do those do
 
in templates, <includeonly> doesnt show its contents to its own page but shows on the page that the template was put in, <noinclude> only shows its contents on its own page but not the page that template was put in
also <onlyinclude> where only its contents will show in the page
 
 
2 hours later…
10:17 AM
Anyone have any opinions on which JavaScript based golfing language would work best for a byte-restricted KotH?
I don't know any, but am considering Japt and TeaScript.
Wondering whether to choose one, or possibly accept either.
Anything which can compile to JS is a candidate, but I'm leaning towards keeping it simple, so just one language is my current preference
 
@trichoplax Byte-restricted?
@trichoplax Pretty sure Japt is the only one still in development
 
For example, your entry must be no more than 200 bytes of Japt
Japt was what prompted the idea, but I wanted to get an idea of what other possibilities there were
Although something not being actively developed might be more stable?
 
@trichoplax I don't think any are being actively developed :P
Also: I'm not aware of any ones other than Japt, TeaScript and ESMin
Actually, what about ngn/apl and ngn/k :P
 
ESMin - thanks - will have a look
 
0
A: Sandbox for Proposed Challenges

Asone TuhidAndrey Kolmogorov turns 115 Today, 25th of April 2018, the Russian mathematician Андрей Николаевич Колмогоров, born in 1903 would have turned 115 years old. Among his achievements, he's most famous around these parts for the concept of Kolmogorov complexity which describes the shortest code nee...

 
10:25 AM
SOGL and Canvas both have interpreters in JS, though almost noone except me uses them :p
 
On a side not: so many people have left
 
Left TNB?
 
@dzaima Oh yes this. SOGL and Canvas are both very good golfing languages. Although if Canvas uses only a canvas then it might not be a viable
 
@ASCII-only Canvas doesn't have a main canvas, and if not touching the ascii-art built-ins, it's pretty much like a regular stack-based but with less built-ins (because I'm lazy)
 
@trichoplax IDK actually. I wonder how many of them are still (semi-)active on PPCG
 
10:27 AM
I need something that can handle taking an array describing the board and piece locations, and returning a move (potentially as an integer)
 
@dzaima :| less builtins
Then I'd say SOGL, ngn/apl and ngn/k probably aren't bad contenders too
 
@ASCII-only I wanted to be careful about which built-ins deserve to be 1 byte, and ended up left with way too many usused :p
 
@dzaima better than what Charcoal did
me: oh hey that's useful
<1 day later: new builtins added to Charcoal
me: oh hey i'm bored
<1 year later: too much useless crap added to Charcoal
 
I should get back to Canvas development :p
 
And I should really figure out how to switch to symengine
 
10:33 AM
How can you expect to get any submissions if you make a KOTH only allowing a single obscure golfing language?
 
I'm also leaning towards something which can be compiled to JS up front, which will be easier for me to work with and hopefully faster to run
@feersum I made a KotH allowing only JS and people learned it to enter...
 
JS is not obscure or a golfing language
 
@trichoplax learning golflangs is way harder though
 
Fair point
 
@feersum [citation needed]
 
10:35 AM
I am wondering about allowing a small fixed set of languages that compile to JS
 
well you could always make the language APL or K
 
@trichoplax Through emscripten, most languages compile to JS.
I don't think many will link OK though.
 
@wizzwizz4 but it's usually pretty slow
 
@ASCII-only ASM.JS is pretty fast.
 
I want something reasonably fast, and I want a reasonably small set of languages (or just 1)
I could just post a byte-limited KotH allowing only JS answers, but I like the idea of it being a golfing language
 
10:36 AM
@trichoplax Have some sort of generic bytecode, like JVM or Dis.
 
@wizzwizz4 IDK. I thought the overhead of basically emulating an environment would cost quite a bit, but I guess it should be comparable to JS speed
 
@wizzwizz4 Then allow entries in any language that can output that bytecode?
 
@ASCII-only Emulating an environment would be a performance penalty, but on good implementations (such as Firefox) it's only ~1.5x.
@trichoplax Yes.
 
Interesting
 
@wizzwizz4 oh, that's not bad then
 
10:38 AM
I recommend Dis over JVM because, although JVM is more popular, more things compile to Dis.
Well, more things can be made to compile to Dis.
 
Eh, I'm pretty sure wizzwizz is talking out of his you-know-where here.
 
@feersum JVM is stack based but Dis is register based, which means you can make LLVM work with it.
I think.
 
Huh? I'm referring to the various performance claims.
 
Never mind; LLVM converts to JVM too. You can choose either.
 
@wizzwizz4 Dis?
 
10:40 AM
It's the Inferno equivalent of JVM.
 
related: if an "ideal" bytecode existed, what would it need
 
@ASCII-only Turing Oracle support, for one.
And a "do what I mean" flag.
 
ngn
I don't recommend ngn/apl - too many early mistakes I cannot fix without a complete rewrite. I was young and green, and uninitiated in APL...
ngn/k is not in js
 
@wizzwizz4 *ideal within realistic expectations
 
10:42 AM
@ngn oh. oops :/
 
@ASCII-only Something that matches standard CPUs as we use today.
But also works with GPUs and Quantum Computers.
 
Hmm. Fair enough
Also >_> I guess I didn't word my question well. Or word the right question at all
 
> our preliminary benchmarks of C programs compiled to asm.js are usually within a factor of 2 slowdown over native compilation with clang
 
@trichoplax They improved that to 1.5x though.
 
Wow
 
10:44 AM
> Firefox with float32 optimizations can run all those benchmarks at around 1.5x slower than native, or better. That’s a big improvement from earlier this year, when as mentioned before things were closer to 2x slower than native.
I'd have just won the browser war argument, except asm.js is archaic now.
 
I wonder if people would be more likely to enter a KotH in Japt or asm.js
 
It is used quite a lot though
 
Well all apologies to wizzwizz4, now I blame Mozilla for being full of crap :)
 
@ASCII-only Yeah; it's still good.
@feersum They have a bar chart! How can they be lying?
> Lucent licensed Java from Sun, claiming that all Inferno devices would be made to run Java. A Java byte code to Dis byte code translator was written to facilitate that. However, Inferno still did not find customers.
If you use Dis, you automatically support (slower) JVM.
 
CMP: What would a bytecode be like if it was designed for high-level languages to compile to it? Would .NET bytecode be pretty close to ideal?
 
10:45 AM
@trichoplax I'd expect either to have very little participants :p
 
@ASCII-only Python bytecode, probably.
I don't know what .NET bytecode looks like.
 
Which languages are considered "high level" here?
 
Any language that's feasible to run on an actual computer really
 
OK, so machine code is high level?
 
@wizzwizz4 Basically put things onto a stack and call functions on the stack
 
10:47 AM
@ASCII-only Oh. So, like Java and Python then.
Yeah, that's probably what high-level looks like.
 
@feersum hmm, that's not exactly high level, but it should still be easy to compile machine code to said theoretical bytecode
 
Can you name an example of a not high level language according to your definition?
 
ngn
@ASCII-only cpu manifacturers these days consider x86 a high level language - they implement it in microcode
 
@dzaima Maybe I should make it a JS KotH with a byte limit, and then allow optionally using Japt or something too
 
ngn
@trichoplax why not plain js?
 
10:52 AM
@trichoplax I don't really see the point in involving golflangs anyway - just multiply the bytecount allowed by a bit and limit to JS?
 
If any of you can get this page working, you've got an Internet Explorer-based emulator.
 
@feersum I guess I worded it incorrectly :/ It should be easy to compile any language (as long as it's not just theoretical) to the bytecode easily
 
@ngn That would work - I just liked the idea of using a golfing language since I'm going to restrict the length of the competitor code
 
@trichoplax You could just increase the length limit. As an added bonus, you're using JS, meaning no compilation and everything's fast
 
@dzaima There's a good chance I'll end up doing that. Keeps to my rule of keeping KotHs as simple as possible without losing whatever makes them interesting
 
10:53 AM
That menu on the right shows last message pushed, not posted.
 
@ASCII-only I think you need moer specific design goals for the question to be answerable.
 
@feersum What about is not specific enough
 
@ASCII-only Actually, I like this idea. And start off with strict mode required and design with speed in mind up front
 
Anyone remember the trick to make SE chat realise that one has changed avatar?
 
10:57 AM
@Adám change chat's parent user (and change back)
 
@EriktheOutgolfer Ask a mod to do it?
 
Hmm. Actually, having a speed limit isn't that bad of an idea, since for Formic Functions each round takes forever :P
@Adám Yeah
 
@Adám Just refreshed your profile
 
@Adám or Just Wait for It™
 
@trichoplax Thanks, but wasn't mine: chat.stackexchange.com/users/271065/stormwind
 
10:58 AM
Have we reached the point where if you got to design a CPU from scratch, RISC would actually have overall inferior performance to CISC because it would bloat the binary size so much?
 
@Adám Ah I see - done now
 
@trichoplax Thanks, on his behalf.
 
@feersum Probably not. On a side note, I wonder if the smaller instruction set would reduce instruction lookup times or something like that
 
@ASCII-only A big problem with Formic Functions is that several of the players create hundreds of workers each, so they get the 15ms limit hundreds of times. The idea I'm working on now has a fixed number of pieces per game so shouldn't suffer from that. I might have to keep the time limit generous to allow for golfing techniques though
 
@ASCII-only How do you know?
 
11:02 AM
I don't actually know >_>
 
@trichoplax Is it nothing more than a button press for you?
 
@Adám Not much more - 5 clicks
 
@trichoplax Couldn't it be automated so users could themselves refresh their profiles? Possibly with a time-based quota so the system wouldn't get flooded?
 
@Adám Probably.
Post a feature request.
 
@Adám That sounds sensible
 
11:05 AM
@Adám A better idea would be to make refreshes automatic :P
 
I think it could be used as a DoS system, though.
It's probably computationally expensive at the moment, like rep recalcs are / were.
 
How? That's silly.
 
@feersum Also I don't think any amount of reduced instructions (withing reasonable limitations) would bloat binary size by a significant amount
 
Asyncronous Funky is going far too well for it's own good.
 
@ASCII-only Users may edit their profile many times in a row to see how it looks. No need to refresh (if it is resource intensive) until they are done.
 
11:07 AM
@wizzwizz4 It shouldn't be computationally expensive though...
 
If it's updating all of the transcripts, it could be.
 
@ASCII-only So you're arguing that a more complex instruction set doesn't actually make the code any shorter?
 
I don't think there's caching there, but there might be.
 
If it is just a matter of resetting one's parent user, a user script could do it…
 
@feersum It does, but e.g. 2x more instructions wouldn't make the code anywhere near 2x shorter. probably.
 
11:09 AM
@wizzwizz4 Do you think there is literally a copy of the profile image stored in the database for every transcript message?
 
@Adám I guess, but it shouldn't be that resource intensive. Maybe set a limit of once every e.g. 5 mintues (starting at the first time the user changes avatar)
 
@feersum No, but there's probably a URL.
Come to the sandbox and I'll demo it to you.
 
Supposing it is implemented as a URL, one simply changes the image served by said URL.
 
@feersum But it's the URL to the image, I think.
 
Right, I'm saying today /foo.png serves a goat, then the user changes profile, and now /foo.png gives you a fish.
 
11:11 AM
@wizzwizz4 this has the correct image
 
@feersum You could do it like that, but SE doesn't seem to internally.
 
It's much more likely that an ID would be used for the user rather than a URL.
So you don't actually need toh ave a URL that serves different images.
 
@wizzwizz4 yeah. what it appears to do instead is something like store the image URL in a DB, then use that every time a page is generated
 
I killed my profile picture :-(.
I don't know how to get it back.
Here it is, for future reference.
Never mind; I think I've fixed it.
 
0
Q: Let users refresh their chat profiles

AdámHow long does it take for chat profiles to update? says it can take up to an hour. Please add a button where one can refresh one's profile using the same technology as what the moderators can use. If it is too resource intensive to allow this freely, add in an appropriate time-based quota, like w...

 
11:25 AM
@trichoplax I realized that if you want pure functions, there's a dirty trick to "fake" require it: Memoize all functions
 
@NathanMerrill Yes that works for Formic Functions, since the ants have no memory (including no state for a RNG). It's implemented and makes a huge difference, although there are a huge number of possible inputs to memoise
It is just about manageable because they can see only the 3x3 grid around them. For what I'm working on now I want each player to see the whole board, so that won't be an option
 
@trichoplax Tbh, I think you should use WebAssembly.
It's a bit irritating for those of us who still refuse to let Internet Explorer go... *sniff* but it's probably the language that best fits your criteria.
 
Is WASM available (as a real implementation, not one that compile to JS) in the consumer versions of major browsers now?
 
> not one that compile to JS
You are confusing asm.js and WASM
 
I am not.
I read of WASM prototypes that compiled to JS.
 
11:33 AM
@wizzwizz4 I'm prepared to accept the sacrifice of irritating those using Internet Explorer if necessary
 
To answer your question: caniuse.com/#feat=wasm
 
@feersum That's for polyfills iirc.
It's ironic that caniuse.com doesn't work in very old browsers.
 
Why does Chrome skip from 49 to 63 in this caniuse table?
 
@feersum For usage percentage
 
@feersum It's showing the more popular versions.
 
11:36 AM
@feersum It doesn't. Click "show all".
 
It's bascilally saying the majority of chrome browsers are 49+ - for the reason they chose 63 (well, probably) see the percentages of 49+ and 63
 
@wizzwizz4 I like the sound of that for speed, but it would depend on how much participation the contest would lose by using WebAssembly rather than JS
 
chrome's current version is 66 but caniuse shows 65 ree
 
@trichoplax solution: write the controller in wasm, add a JS interface
 
You don't need to use an assembler.
 
11:44 AM
@betseg well it hasn't been that long
 
@ASCII-only That sounds like it would still limit me to the speed of heavily golfed JS
 
@ASCII-only 5 days is long in internet
 
@trichoplax Why the need to golf it though
 
To compile WASM you have to compile it from asm.js??
wtf
 
@ASCII-only That's the idea I'm working on - a KotH where you only have a limited number of bytes
 
11:46 AM
maybe that's what I was thinking of earlier.
 
@feersum That's ages ago.
@trichoplax Limited number of Web Assembly opcodes?
 
@wizzwizz4 No I just mean an upper limit on the length of your code
 
@trichoplax Hmm :/ I don't think you can have golfiness and performance at the same time
 
@ASCII-only Fact check: Mostly false.
 
@ASCII-only I agree - hence wanting to start off with good speed so more expensive golfing is practical
 
11:49 AM
What source language should you use if you want your WASM to be efficient?
Like I assume C is not great since pointers don't compile well to a stack machine.
 
@feersum What source language should you use if you want your executable to be efficient?
@feersum Probably Java, if WASM is stack-based.
 
Wiki says Java is not supported yet.
 
Plus wasm has no GC...
 
Rip. Ok, just go with C++.
 
That has the same situtation as C.
 
11:51 AM
@wizzwizz4 Does C not rely heavily on pointers too...
 
OK, I read an actual manual and pointers seem pretty well supported
So C is OK.
 

« first day (2640 days earlier)      last day (2205 days later) »