01:29 <d_bot> <darrenldl> python doc has always been (much) worse to navigate even if im left with just .mli

03:37 <d_bot> <mk-fg> Fwiw, never had a problem with ocaml docs, but there's got to be a bias wrt commenting on it here, as presumably all people who had an issue parsing those will give up on trying to learn ocaml and likely won't be in any related channels :)

03:38 <Corbin> I would put up Ruby as an example which blubifies OCaml here; that is, OCaml sits between Python and Ruby in terms of reference documentation quality.

03:40 <Corbin> There's two features in particular. First, most Ruby packages are documented using centralized tools, so that if you can ask your package manager to quickly install the gem from a central repo, then you can also ask the central repo to show you the gem's docs.

03:40 <Corbin> Second, many short methods are directly inlined into the docs; you can just click a method if you want to see how it's implemented.

03:41 <Corbin> https://www.rubydoc.info/gems/thor is an example popular package, for those who haven't clicked around Ruby docs before.

03:48 <d_bot> <mk-fg> Yeah, quick check source indeed looks like a cool feature

03:50 <d_bot> <mk-fg> Didn't actually get to use much docs on third-party ocaml stuff myself, as mostly use it to implement small system/helper tools for which stdlib works perfectly fine, and integrating with various libs is trivial via a bit of C glue

03:54 <d_bot> <mk-fg> One feature that I probably like the most in ocaml and python docs, is that they're easy to download and then use by putting/replacing module name in the URL directly via custom search in the browser to open the ref page for that module instantly

03:55 <d_bot> <mk-fg> E.g. "ml Hashtbl" will just open `file:///home/.../Desktop/docs/ocaml-X.Y/libref/Hashtbl.html`

03:55 <companion_cube> Corbin: I'd say OCaml is on par with your second point

03:56 <companion_cube> one keypress tells me a function's signature and doc

03:56 <companion_cube> `gd` jumps to its definition

04:03 <Corbin> companion_cube: We're comparing Web docs to Web docs. That's a cool editor feature though.

04:18 <d_bot> <Anurag> central doc site should come in a future update to the ocaml.org website. The work-in-progress website is up and running now, an example: <https://v3.ocaml.org/p/containers/3.6.1> and docs at <https://v3.ocaml.org/p/containers/3.6.1/doc/CCParse/>

04:20 <d_bot> <mk-fg> * <https://v3.ocaml.org/p/containers/3.6.1/doc/CCParse/index.html>

04:22 <d_bot> <Anurag> Thanks for the fixed url. I missed the `index.html` at the end

04:28 <d_bot> <mk-fg> https://cdn.discordapp.com/attachments/436568060288172044/921619902475943986/docs-new-vs-old.jpg

04:28 <d_bot> <mk-fg> I like how condensed the old docs are tbh, and lacking in visual noise

04:29 <d_bot> <mk-fg> But not like it's a big deal, of course

04:34 <d_bot> <mk-fg> Also I think the brightness and comparatively low contrast is what makes the new page immediately harder to look at for me

05:24 <d_bot> <Jektrix> Maybe a little floating table of contents on the side would be nice so you could jump to definition without scrolling

06:20 <d_bot> <NULL> https://ocaml.org/api/Str.html is new-style but more condensed

12:16 <d_bot> <VPhantom> I like the new style a lot, for what it's worth, but I think the biggest improvement for legibility for me is the use of sans serif fonts.

12:17 <d_bot> <VPhantom> (Well, I'm not so sure anymore, because even that Str.html example looks great. 🤔)

12:44 <d_bot> <AZAH> Yep I also prefere the new style

14:51 <d_bot> <crab> is this discord server active?

14:52 <Fardale> Yes

14:52 <d_bot> <crab> a lot?

14:54 <Fardale> What is the definition of "a lot"?

14:56 <d_bot> <crab> im assuming yes

14:57 <d_bot> <crab> also, why did you decide to join the server as an IRC instead of a discord user

14:59 <d_bot> <arctumn> That's a user choice, they have their reasons

14:59 <d_bot> <arctumn> Only this channel is under IRC (if i'm not wrong)

14:59 <Fardale> I was on IRC before this channel was bridge to the discord server

15:26 <reynir> (re: new vs old doc) I personally prefer the higher contrast text of the old docs. After all, reading the text is what the docs is all about, no?

15:27 <d_bot> <mbacarella> Are you looking for confirmation that it's true? Or are you looking to add a citation to Wikipedia? I can anecdotally confirm it's something he said and that traders at the company did read OCaml code.

15:34 <d_bot> <Kakadu> Citation would be better because I can use it against everybody who says that FP is too complicated

15:40 <d_bot> <zakkor> The new v3 site/docs look GREAT!

15:40 <d_bot> <zakkor> Tailwind/Tailwind typography ftw

15:46 <d_bot> <mbacarella> Perhaps his "camel trading" talk has the quote

16:40 <d_bot> <VPhantom> Oy, FP _can_ be complicated (category theory), but for day to day work, I find that lots of my functions tend to melt away to half to a quarter of their initial size when I work in OCaml vs any of my previous, imperative languages. The expressiveness quickly becomes very apparent even just as a rookie fooling around with the Stdlib.

16:44 <d_bot> <Kakadu> Thanks, it looks like it is a right reference indeed

16:45 <Corbin> VPhantom: Funny you say that; my category-theory-based language's expressions are much smaller than OCaml's; lots of lambdas and name-plumbing just melts away~

16:49 <d_bot> <let Butanium = raise Not_found;;> what is this language ?

16:49 <Corbin> https://esolangs.org/wiki/Cammy

16:50 <d_bot> <mbacarella> The specific success case if you want more color is that they tried to rewrite a hodgepodge of Excel spreadsheets (aka their core prop trading infrastructure) into real trading systems with proper software engineering and the traders found the C# code that was supposed to replace it very difficult to follow because of mutability and inheritance.

16:54 <d_bot> <mbacarella> So. Nobody wanted to risk gigantic sums of their own money (that's what prop trading means) on automated trading code they couldn't understand.

16:56 <d_bot> <mbacarella> I think a big contributor to their culture of thinking about risk this way was that one of the partners worked on a nuclear submarine. Which is obviously an environment where reality doesn't care about how confident you feel about your system.

17:44 <d_bot> <let Butanium = raise Not_found;;> if you want a great challenge

17:44 <d_bot> <let Butanium = raise Not_found;;> https://adventofcode.com/2021/day/18

17:45 <d_bot> <let Butanium = raise Not_found;;> I have 4 mutually recursive functions for now

18:49 <d_bot> <Frisk> I remember someone (Scott Wlaschin maybe?) mentioning PMs somewhere being able to read f# interfaces and inferring how they'd be wired together, not realizing they were actual code

18:50 <d_bot> <Frisk> Anecdotally I have a MUCH easier time showing people who don't code reason code over say JavaScript much less c# or Java.

18:51 <d_bot> <Frisk> It's specifically people who have gotten used to OOP that have a harder time with FP concepts than people in general, I've found.