Learning a new programming language

[zah21]

Hi guys,

I'm considering learning a new programming language this year. I'm proficient with Python & I've also coded data structures from scratch in C++ so a bit familiar with that as well.

Anyway, I've researched a bit about the languages that are worth learning and here are the choices that I came up with: Go Lang, Rust & good old C.

Let me know which one would you recommend.

Thanks.

[Ant P.]

Raku

[Irre]

I am very old (73). I have programmed, in order: Algol, Fortran, PL/1, Assembler, Rexx, C/C++

The only language I really liked was: Rexx

[NeddySeagoon]

zah21,

Choose a language that is popular is the problem field you want to work in.

[apiaio]

I am not so old as Irre. Only 66 .
I have programmed, in order: Algol, Fortran,Cobol, Assembler 8008 ... Macroassembler PDP11M, Pascal.
Nowadays LAMP, JS, SQL as hobby. I have never been professional programmer.

@zah21 - doesn't matter which language you will start learning. All of them are similar. Next year will appear
another language.

Edit:
And yes. NeddySeagoon is right.

[NeddySeagoon]

Its time to roll out the Gentoo Old Timers.
That was created in response to Gentoo Young Ones.

I'm 67 and learned Algol, BASIC, FORTRAN and a smattering of different, mostly 8 bit assemblers. 6800, 6502. Z80 ...

[The Doctor]

If you are learning for fun I recommend Haskell because it is so different. (I (also (think (lisp (might (be (fun.))))))) But the parenthesis are a bit much. Both of these languages are actually fairly popular among their users.

Learning more of the same is just boring and brings negative transfer to your programming in both languages.

[Zucca]

Learn awk and perform magic on the command line. ;)
awk is the last language I learnt. It's been quite useful for myself.

Next I'll get into go maybe... or rust. Hard to decide.

[Irre]

A language I never learned was APL (Another Programming Language!). It was very compact, but impossible (for me) to understand.

[C5ace]

Learn C and get an understand of assembler. You can then code anything you want.

[xahodo]

Well, I have a bunch of questions, before I can provide an answer:

• Which architectures do you intend to target?
• Which OSs do you plan to target?
• What do you want to achieve with your software?

Rust is actually quite fun to learn (but the borrow checker can give you hell), but it doesn't support that many platforms (yet). There's also the issue that they don't have that much stuff going for them when you want to program a gui or game.
On the other hand, bash scripting is nice to learn so you can create your own ebuilds and have a bit of fun with the console doing stuff.
Or, if you're a masochist, brainfuck is most certainly an option.

[figueroa]

I am very old (73). ...

Ouch. I read that and said to myself, wow that's old. Then I realized, me too.

I wouldn't recommend random learning of another programming language. Learn what you need to be productive, marketable.

[Muso]

Go

The speed (nearly) of C with the ease of Python.

[psycho]

This short article gives some good reasons to learn C. TLDR version:

1) It's low-level enough that it makes you more aware of what your code is actually doing (e.g. with memory etc.), with potential benefits to the way you think about programming, even in other languages.
2) It's so old and yet still so widely used that it's one of the languages most likely to be used to illustrate programming concepts (e.g. other languages will often show off their features next to C equivalents).
3) It's still the language that most of the basic stuff is written in (Linux and other kernels, most of the GNU tools, etc.).

[Banana]

I choose my language which fits the "problem" I want to solve. I find it easier to have a problem/task to work with then just learning it by a book. General books about a language still help but learning by doing works better for myself.

I recommend haskell and smalltalk to be added to the list.

[Goverp]

I'll second the vote for Smalltalk if you just want to learn more about programming. The pure OO mindset is well worth acquiring IMHO.

[szatox]

I like learning new things that are very different from what I already know.
You know, you can have 3 hammers, but you could do the same job with just one hammer most of the time, so you can replace 2 of those with a saw and a drill to make your toolbox much more versatile.

If you already know OOP, how 'bout checking out FP? Haskel, erlang, lisp? AFAIR lisp in particular allows programmers to develop very advance software in a very short time. Which results in a situation where every user has his own implementation of X solving 90% of the general problem, and it's a different 90% than all the other implementations
I've tried erlang and it was pretty convenient with it's actor model, great for building distributed, asynchronous systems.

Shell is a powerful tool too, it has some quirks, but it does a pretty good job when you want to process a data stream.

Both of these languages are actually fairly popular among their users.

a bowl of water is wet inside

[pjp]

learn C

Do you know of any resources that address modern "best practices" that avoid the "bad parts" of the language? I only seem to come across them in bits and pieces, and I'd rather not learn it incorrectly. For a lot of it, I suppose a quick reference guide of "Do use / Don't use" and "Do / Don't do" would probably be fine. I need to get back to K&R.

[The Doctor]

The thing about best practices is I don't think any such thing exists. I've had several professors fundamentally disagree about what best practices are. It seems to me that the real best practices are to adopt a maintainable style for yourself and stick to it.

Just my 2 cents.

[C5ace]

learn C

Do you know of any resources that address modern "best practices" that avoid the "bad parts" of the language? I only seem to come across them in bits and pieces, and I'd rather not learn it incorrectly. For a lot of it, I suppose a quick reference guide of "Do use / Don't use" and "Do / Don't do" would probably be fine. I need to get back to K&R.

MISRA-C
https://andrewbanks.com/wp-content/uplo ... dard_C.pdf
plus many more.

[zah21]

Hello guys,

Looks like the star of the show is C & Rust. That's good, I've just bought books from Humble Bundle few months ago covering those languages

Go is also not that bad, I mean it has concurrency builtin and language is kind of similar to Python.

It's hard to decide right now, may be I'll do a coin toss & decide later

[pjp]

The thing about best practices is I don't think any such thing exists. I've had several professors fundamentally disagree about what best practices are. It seems to me that the real best practices are to adopt a maintainable style for yourself and stick to it.

Just my 2 cents.

Yeah, I wasn't sure how to put it other than that. I'm referring to the parts that are known as "don't do" (incorrect use of scanf being the most obvious). I guess I should try to put together a list as I come across them. Also, I've read about some professors who teach Bad Things, so there's that (I can't recall specifics, but they fell into the category of the stuff I was hoping to find in a quick reference format).


@C5ace: Thanks. Firefox doesn't render it, so I'll have to find a reader. I think at some point I'd seen a reference to the JPL, but it seemed a bit unapproachable at the time. Will work some more on the K&R this afternoon.

[Hu]

I've read claims that Rust is good because it makes it hard to write certain types of bugs, and improves the review process by confining certain bug types to areas that are easier to recognize. I can't get over how incredibly bad their install/bootstrap process is, so I've never been able to take the language seriously. This may be an overgeneralization, but my perspective is that if they can't even care about having a good install process, something that everyone has to deal with and where bad practice is blatantly obvious, how can I expect them to care about doing a good job elsewhere, on the parts where I can't readily evaluate how well or poorly it was done?

[pjp]

Not being a developer, what has turned me off of rust has been the community "PR" efforts (I'll leave it at that). I still decided to give it a look, and here's what I found:

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

I can't take anything seriously that recommends that approach.

Generating a new project

Let’s write a small application with our new Rust development environment. To start, we’ll use Cargo to make a new project for us. In your terminal of choice run:

cargo new hello-rust

This will generate a new directory called hello-rust with the following files:

Seriously? For "Hello, world!"?

There was some awkward / confusing syntax too, but I suppose I'd eventually figure out that. I've read / heard that there is a somewhat steep learning curve even for experienced developers. But the first 3 issues are significantly off-putting (and that excludes other issues I've read about, so I've decided for now that it isn't for me).

[psycho]

Code: Select all

fn main() {
  println!("hello world");
}

That's literally all there is to it (well, and rustc hello.rc && ./hello). I feel as though if a language gets any simpler than that it's making too many assumptions where I'd rather tell it explicitly what to do...but that's probably just an irrational feeling. I think my first hello world was something like

Code: Select all

10 PRINT "HELLO WORLD"

(followed by "RUN") on an Apple II (no lowercase in those days). As a result I probably just feel like adding a few brackets makes it a superior language