All the hype on HN seems to be towards Rust and Go (with Go having its share of haters), but Nim sounds particularly interesting as a side-step from Python that runs faster than both Go and Rust.<p>Can anyone give me some insight as to why it's not as hyped, if at all? I'm rather unfamiliar with Nim, and it's seems to have a smaller community, but it feels like it should be getting all sorts of love considering its speed of execution.<p>Thanks in advance for the help!
The release is happening today, so I posted a bit early. Some parts, like the binaries, are not online yet.<p>There are tutorials for newcomers: <a href="http://nim-lang.org/tut1.html" rel="nofollow">http://nim-lang.org/tut1.html</a>
<a href="https://nim-by-example.github.io/" rel="nofollow">https://nim-by-example.github.io/</a><p>Edit: Fixed tutorial link
Something I find quite interesting about Nim is write tracking: <a href="http://nim-lang.org/blog/writetracking.html" rel="nofollow">http://nim-lang.org/blog/writetracking.html</a>. It uses the language's effect system, and allows not only specifying for instance that a function is referentially transparent, but also specifying how an impure function accesses or modifies global state.<p>An example from the above link:<p><pre><code> var gId = 0
proc genId(): natural {.writes: [gId].} =
gId += 1
return gId
</code></pre>
To quote the article: "Here the effect systems shows its strength: Imagine there was a genId2 that writes to some other global variable; then genId and genId2 can be executed in parallel even though they are not free of side effects!"<p>The effect system can also be used for exception tracking, specifying what kind of exceptions are function may throw. It can moreover be extended to work with user-defined effects.
Nim is very exciting to me. For a long time I've wanted to develop a viable HDL as an internal DSL in a modern imperative compiled language. Until I found Nim I couldn't find anything beyond C++ that would work (Julia might actually, I haven't tried yet), but Nim is just absolutely perfect. The custom operators, static params (a weak form of dependent typing), and the macros in particular makes Nim a perfect match. It also helps that the compiler is itself written in Nim and pretty easy to work with. It's exciting when a language allows you to solve a whole new problem that was just not feasible before.
Hopefully one day the freepascal backend gets updated. There is a lot of Pascal/Delphi code out there that can benefit (rewritten and call legacy pascal) from Nim's improvements.
So far, Nim has passed the most basic test for me: it installs according to instructions and compiles its own example code. I'm always baffled by the number of language releases that don't reliably do that.<p>Nim looks like a worthwhile language, and I'm looking forward to learning it.
As a long time Python programmer who has struggled to cross the gap from dynamic scripting languages to modern statically typed languages Nim is by far the most frictionless language I have tried. Before I found Nim I longed for a language like C# where the generics "just work" and overall the language just feels like it designed at-once rather than piecemeal overtime. Everything just seem really "nice" in C# and I am able to transfer my Python experience over to it. But being locked up to .NET/Mono I never really used it outside of Unity3D.<p>I tried Golang, because it was sold to me as something I would love since I am a long time Python developer. I strongly dislike Golang. It doesn't have much in the way of letting me model my programs like I am used to. I am told "that's wrong, do it the Go way". This is too much friction. Once I am done thinking how to solve my problem algorithmically, I do not want to then figure out how to rethink my algorithm just for the sake of the maintainer's of Golang.<p>I tried Rust. I think Rust is beautiful (mostly). However, Rust has a far too fundamentalist view on memory safety. And that's not to downplay the importance of memory safety. But there's just too much friction. I want to sit down an implement my algorithm. I don't want to stop and spend just as much time thinking about the particulars that Rust demands.<p>When I found Nim I almost couldn't believe it. The core language was simple, clean and immediately absorbable. I was able to start writing basic Nim programs after just perusing the language docs for a few minutes while the compiler compiled. I read that Nim had powerful meta-programming facilities and this started to turn me off. I had heard that macros were generally a negative force in the world but only knew them from Lisp. Then I learned that Nim's macros are just functions that take and AST tree, perform modifications to that AST and return an AST tree. Wow that's pretty simple. Oh hey the generics "just work" like in C#. Woah, Nim even supports inheritance!<p>Nim is definitely the next language for me. In thinking about it, I find that I agree with one of Jonathon Blow's sentiments that we have been seeing a number of new up and coming languages but they are all Big Idea languages. Big Idea languages who's ideas have yet to be vetted and proved out over the course of a decade or two. They all incur too much <i>friction</i>.<p>Nim seems like a competent, low-level systems language with a helpful repertoire of well-implemented modern tunings to features that are all mostly established in the field. It doesn't try to revolutionize systems languages. It tries to bring us a modern refinement by bringing us a highly performant yet relentlessly productive take on what has already been shown to work.<p>Please don't be offended if you see me around evangelizing its existence.
Is the JavaScript back-end receiving any updates? Last time I looked at it it <i>looked like</i> it was lagging behind (emphasis in "looked like", I did not spend too much time with it).
>The unpopular "T" and "P" prefixes on types have been deprecated<p>I really like that change. Those always looked so archaic to me.
Anyone knows what kind of project would nim be a good choice for at the moment ? I mean both the language feature as well as available libs. I think i've read somewhere that web server coding isn't a target yet (at least regarding websockets use, which was my primary concern at the time i looked).
Maybe it's just me but I find this language pretty ugly. And regardless of any technical advantages it may offer, that's enough reason to not use it. There's too many languages out there to come out with a new one where the code isn't clean and nice.