Prophet
@welltypedwitch
Followers
3K
Following
48K
Media
667
Statuses
10K
There's no sense crying over every mistake. You just keep on trying till you run out of cake. - GLaDOS https://t.co/70SfmxnROe https://t.co/p6LXXU7dGH
Joined July 2019
I didn't think I had to say this, but if you try to follow me and I see that your following looks like this, I *will* block you. Trans rights are human rights.Black lives matter.Fuck nazis.(read the alt text for translations)
7
4
102
okay i get that people don't like oop on here but the number of people that genuinely don't understand the point of getters/setters is astounding (including some of the replies to this).
145
102
4K
@ChShersh okay but why though? being able to open pdfs as browser tabs is such a good feature.
1
2
4K
>try a new programming language.>ask if (%) is rem or mod.>they don't understand.>pull out diagram explaining what is rem and what is mod .>they laugh and say "it's a good (%)" .>(-1) % 4.>-1.
37
355
3K
welp, look what i just found in some code i wrote 7 months ago. (this is not a joke lol)
Due to my elite programming skillz, I figured out how to shave off THIRTY SECONDS from my app's startup time. Here's my optimization tips:. 1. Remove the sleep(30) call I added a month ago and forgot to remove.
9
49
1K
"ask her what the difference between a compiler and an interpreter is" is such a middle of the bell curve meme question lol.
These are simple questions that a lot of people learn during their first college computer science course. If you haven’t, though, it’s never too late. It’s easy to forget things that we don’t use, but I promise that no one graduates from Harvard with a computer science degree
34
27
1K
what’s your most liked tweet?.
16
41
1K
i still prefer this version
10
69
1K
e.g. if you have. class Person {. private String name;. public String getName(){. return name;. }.}.you can change it to use first and last names in the future without breaking every usage site that depends on the full name.
24
5
904
don't get me wrong, this java tendency to mindlessly throw getters and setters at every variable is a terrible idiom, but that's because "get/set this field" is (usually) a bad interface. a public field would be *worse*.
14
8
809
like, the reason these exist is that you really don't want to expose the fact that `thing` is a class field and have to commit to it for all eternity.
16
3
780
it's insane to me that people actually think fp doesn't have design patterns or that we rarely have names for the ones we use because "they're just obvious" and not because functional programmers would rather wax on about how their favorite language was made by god herself. .
17
30
694
rust people i love you but please stop calling lambdas "closures".
21
14
639
okay why does the minecraft font have math ligatures that look this good wtf
9
47
584
tell me you don't know how compilers work without telling me.
honestly anyone who has written software claiming AI isn't dangerous because it's 'just emitting tokens' is being disingenuous or is in a deep state of cognitive dissonance. the first phase of compilers is called "tokenization". emit the right tokens and you can own the world.
15
24
527
this tweet is so funny to me. they had google bard generate go and rust code for the same problem, apparently didn't notice that the results do very different things and then concluded that Rust Is Bad because, uh, the code bard generated does error handling?.
My crusade against some RUST zealots continues. Here are examples of build patterns with #golang and #rustlang.Not only do you need to know less about language with Go to understand what is going on here, but look at this rust "simple" error handling. Oh, my Spaghetti Monster
27
29
492
fun problem from my algorithms assignment:. design a data structure that supports the following operations:. insert: O(log(n)).determineMedian: O(1).deleteMedian: O(log(n)).
66
8
425
i can't wait for this bubble to burst.
As the stragglers are still heatedly debating F# vs C# vs OCaml vs Haskell, the language designers have already moved on to the next generation of AI-based programming languages Skate towards where the puck is going, not where it.
21
18
381
even hotter take: generated source code shouldn't exist. you should be able to use a nice expressive, introspectable metaprogramming system instead.
My hottest tech take: generated source code should be committed. It happened to me so many times when the code generation logic was changed, a bug was introduced and there were no git diff to even compare with the old versions.
33
21
377
okay what the fuck. this is such an obviously terrible idea how did anyone on the rust team ever think this was acceptable??
26
12
368
jesus christ most "criticisms" of OOP are impossible to take seriously. people will literally complain that OOP forces everything to be an object and then gush about how their favorite language has first class functions.
60
10
362
unboxed types, traits, linearity, guaranteed data race freedom, freestanding / embedded support, low memory overhead, real time performance guarantees, easy and lightweight FFI in both directions, broad architectural support, low level control over thighs like custom allocators.
8
19
350
Python pattern challenge - Day 6. Best code will be rewarded.
15
15
329
Type inference is probably the closest thing to magic I know. You could have never convinced me in 2020 that this would just make sense to me now.
15
27
313
hey folks that like types: please stop pretending that people who don't like types are naive or just haven't been "enlightened" yet. you can think that they're wrong and maybe they are (maybe not), but at least take them seriously, okay?.
74
14
304
i'm sorry, i can accept engagement bait but i will not stand for traverse erasure.
Your brain on FP is believing that remembering ~50 functions is superior to learning a single language construct
4
6
291
it is kind of wild how when talking about GCs, most people just accept the idea that malloc/free allocators are fast without question.
35
6
292
i must not get baited.bait is the mind killer.bait is the little death that brings total obliteration.i will face my bait.i will permit it to pass over me and through me.and when it has gone past, i will turn the inner eye to see its path.
14
11
277
i don't buy this. in fact, i'm going to claim the exact opposite!. i think it's *easier* to predict what the fold is going to compile to than the loop. (1/a lot).
Despite the merits of the 1st alternative, many people like the 2nd because they can predict the minimum the compiler will do to translate that to machine code, and that minimum is perf-sufficient. Not just that, but the 2nd can also be changed to a map-reduce in one line. (1/5)
19
26
273
so, haskell peeps if i were to write a "how to write haskell in 2025" blog post, what should i include?. (i'm thinking of things like "don't use string", "just use async", "enable GHC2023", etc.).
32
7
227
jfc how does this have 11k likes? do people actually believe that not rotating a swastika makes it okay to post for a fucking programming challenge.
57
4
221
hot take: the best way for a new functional language to improve performance is to remove linked lists from its standard library.
22
15
225
it's kind of wild that most imperative languages capabilities for robust, typed abstractions are so bad that rust is considered "functional".
11
13
219
i kind of hate how prevalent this idea is that everything is just sort of a wrapper around c. there is just so much more out there!.let's see your c program do moving gcs or segmented stacks. even cps is arguably a lot closer to how the hardware works than c's call/return model.
21
7
220
@hot_girl_spring well yeah because records just do this for you, but the concept of having a function to get the value rather than exposing the field itself is still there with records, right?.
2
0
220
it is really funny to me that some people will say "haskell is too hard because of the math" and then go back to doing gamedev.
22
8
211
cracking eggs.
whats the main use of haskell?.
6
13
209
Functional programming languages should be so much better at mutation than they are.
12
25
215
if you want to learn about monads, please read this. seriously!.waders paper and "you could have invented monads" are literally the only good monad tutorials I know.
Monads are hard because .
14
19
213
Why do people still put character limits on passwords?. Seriously, in what world is a 20 character password more secure than a 40 character one?.
38
11
203
like, another way to phrase the quoted tweet would be. OOP: here is a list of solutions for a bunch of problems with well known benefits and drawbacks and a whole range of literature behind them. FP: just figure it out lol. good luck.
4
9
203
using haskell or rust is fun because you never know when you'll leave your computer alone for 5 minutes and come back to freezing up because HLS/rust-analyzer used up over 32G of memory.
13
9
202
oh sick, a right associative application operator with precedence 0 to avoid parentheses in javascript!.
4
13
202
periodically being reminded that non compiler people mean something very different when they say "code generation".
7
7
194
i can die happily now.
And Haskell loves you back, Prophet.
3
2
191
I mean. have you seen the list of authors on the Verse Calculus paper?.This is like the Manhattan Project of PL and it's happening in Fortnite and "the Metaverse"
3
26
189
c is just cope for people who can't understand fast allocators.
6
3
188
i'm toying with the idea of suffixing unreasonably slow but still useful functions with "slow", as in List.indexSlow. so you only use it if you know what you're doing, kind of like how an unsafe function would be called List.indexUnsafe. too weird? ^^.
41
4
186
and if you desugar the vtables, you'll see that this is pretty much just a church encoding :).
8
9
180
Language I dislike: Python.Language I begrudgingly respect: Scala.Language I think is overrated: OCaml >:).Language I think is underrated: Lua.Language I like: OCaml.Language I love: Haskell.Language I dream of writing in: Vega.
Language | dislike: Java .Language I begrudgingly respect: TypeScript.Language I think is overrated: Go.Language I think is underrated: Elixir.Language | like: Haskell.Language | love: OCaml.Language | dream of writing in: Rust.
16
7
176
it's kind of funny that modern programming languages suck so badly at representing graphs that we pretend they're trees.
16
3
182
I don't think people appreciate the tradeoffs they're making with algebraic data types.
33
20
173
since this is now discourse apparently: "make invalid states unrepresentable" is a *strategy*. it's a good strategy and if it works well it's great, but it is your responsibility to use your judgement to determine when it is worth it and when it isn't.
17
5
171
this is probably my favorite aspect of haskell: haskell just *has things*. there is a common way to misuse your library and you want to tell people what they should do in that case?.just write a custom type error. that's a thing you can do!.
10
10
166
this also applies to our fancy high level functional languages btw. any time you're storing a static collection of things in a linked list instead of an array because it's easier or because that's what everything else expects is a language failure.
I got to watch a person livecoding in C and instead of put themselves through writing Yet Another Hash Find, they just literally wrote a linear scan because "it was easier and this code will never be a critical path". This is what a lot of other people said about their shit too!.
4
12
170
this is such a night and day difference. ocaml docs really have a lot of room for improvement
13
7
161
i do think it's really funny that the bar to being """cracked""" on here is apparently now "writing a compiler and using haskell" lol.
@tunahorse21 the most cracked dude I know wrote a compiler in uni and was the first person I saw using Haskell. that was 17 years ago. he still hasn’t made it to senior because he’s deadass the weirdest dude to communicate with. coding skills will only get you so far.
14
7
160
go is a dynamically typed language with a static type system send tweet.
12
12
156
@The_MrX_ what performance penalty? i would be very concerned if your java JIT can't inline a trivial getter.
2
0
152
I don't make the rules
9
5
155
since this is apparently still a thing, my only take on the "cs grads not knowing git" discourse is: what kind of universities are you all going to where you never learn about git?? and where you don't even have large scale group projects that need some kind of version control??.
34
0
152
death by 1000 cuts sums up ocaml so well unfortunately. and not just the tooling, even the language itself has soo much like this. just off the top of my head, there is.- polymorphic comparison.- reference equality.- whatever ocaml calls "type variables". .
@Aron_Adler opam/dune for sure. opam mostly but, dune isn’t bad but feels a bit lacking. Also things like init project seems like could have some better defaults, and structure/naming. Just many small things. Death by 1000 cuts. Stuff you could probably get used to if you stick it out, but.
10
10
152
pl designers should really come together and standardize their terminology sometime. rust and scala people should have a fight to the death and whoever survives gets to keep the name "trait".
13
11
152
functional programmers will decry the liskov substitution principle as "OOP garbage" and in the same breath go on about how important the monad laws are.
12
3
154
this is a matter of language/ecosystem not personality. i mostly do this but i absolutely couldn't if i had to write python or javascript.
6
6
148
if you know who this house belongs to, we can be friends
10
8
149
i'm really starting to think that a lot of "criticisms" of haskell are just people being deathly afraid of not knowing things.
18
13
150
lukewarm take: linked lists are much better than imperative programmers think and much worse than functional programmers think they are.
all my homies hate singly linked lists.
15
5
137
than think about how to write good code. functional programmers always pretend that it's just a historical accident that fp didn't catch on but have you like seen a haskell or ocaml code base from 20-30 years ago?.
2
2
135
OCaml is a really nice language but the language it's made out to be is not.
13
11
141
Blazingly Fast™ Type Class Resolution with Tries. (hey it's only been. 11 months. since the last time i published a full length bog posts).
13
22
143
slowly forming the opinion that 80 character line limits are actually kind of bad because they discourage you from using long meaningful names since the formatter is more likely to break lines weirdly with them. or maybe that's just ocamlformat.
28
0
141
Lower your guards is black magic. (This compiles without warnings or errors, even though the second case doesn't handle A, because it is already handled by the first definition).
13
10
134
I'm really not a fan of currying, but wow,. (Int -> Int -> Bool) -> String -> String. is so much easier to read than. (((Int, Int) -> Bool), String) -> String.
20
4
132
they're distinct concepts. not every lambda creates a closure and not every closure is necessarily created by a lambda.
1
0
128
One of the main reasons why I think everyone should learn at least one pure language is that it expands the notion of what a side effect even is.
13
5
133
I've looked into what's actually happening here and this is honestly such a strange TS bug. If TS knows that T is a subtype of { x : A | B }, it somehow assumes that T["x"] is *exactly* A | B and *not* a subtype of it, which is just wrong.
Woops! Despite doing everything by the books and not using any escape hatches ('as', 'any', or 'non-null assertions') TypeScript let us give a normal 'member' admin permissions! 💥
5
12
128
hey to be fair: he never said the inferred type had to be correct.
@MarkovMagnifico Done right, a compiler should be able to figure out type automatically. It’s not that hard. Not that it will matter much in the AI future.
7
1
132
Type systems should be consistent (and probably dependently typed).
11
16
130
i'm going to block the next person that says haskell is bad at io jesus christ.
21
0
125
Haskell = Celeste.Idris = Slay The Spire.Agda = Touhou: Lost Branch of Legend.Scala = Wolfenstein II.JavaScript = Fortnite.C++ = Doom.Rust = Doom (2016).ATS = Elden Ring.OCaml = Counter Strike.F# = Valorant.Polaris = Portal 2.Teika = 5D Chess With Multiverse Time Travel.
Programming Languages as Video Games. Java = Call of Duty.TypeScript = World of Warcraft.Python = Sid Meier’s Civilization.C = Counter Strike 1.6.C++ = Dota 2.Go = Slay The Spire.Rust.Lisp = Portal 2.Haskell = Dwarf Fortress.OCaml = Elden Ring.Excel = Minecraft.
21
12
127
but it's frustrating because instead of actually talking about that we have to babysit the egos of a bunch of men with superiority complexes.
7
2
125
Programming is about information not data, or: you might not need dependent types .
11
9
125
and also this!! how is a change that makes previously valid code *stop compiling* not a breaking change?? what
7
4
123
if someone tells you that ocaml's syntax is elegant, ask them what the difference between. type t = A of int * int. and. type t = A of (int * int). is :). (to be clear: i actually quite like ocaml's syntax but calling it elegant is a bit of a stretch).
26
0
118
@tthbaltazar Yes exactly. _ mod 4 is always in the range 0. =3 which is usually what you want.
1
0
115
it's honestly still wild to me that both ocaml and rust implement type class deriving with *proc macros* that generate untyped code.
13
7
119
@stylewarning i mean, it *is* pretty surprising that default arguments are evaluated once and then shared but that's.1) a property of the language, not the interpreter and .2) something you *really* need to know if you're ever using default arguments.
7
1
111
not to start discourse again but quick rule of thumb: if the first example of a monad someone gives you is List, you can probably ignore their explanation.
21
5
113
oh hey computer scientists accurately named something for a change
6
2
112