Which programming language did you choose and why?

[PrettyDeadman]

Noone argues here that c/c++ are the only useable languages or that they are the only language for real developers, just that they are still useable in modern environment, which is a pretty level-headed and non-offensive opinion.

Oh but that's not true! Read the past 1-2 pages please In my view you got things mixed up and actually I'm the level headed guy who's saying C/C++ has their place, even though they're a bit shit Peace.

Having said that, everybody's entitled to their viewpoint

I apologize if I misinterpreted some of the posts here, haven't really been following this exchange too close.

 

[d1nolore]

So if one was to start learning C what editor/IDE or whatever would be recommended?

and does anyone know any good online courses?

 

[Monk]

I read that for non-computer science students, what was used in the past to introduce them to programming was a language like Pascal, and for non-tertiary students, it was usually some version of Basic. But the purpose is to make programs that immediately work on automating processes, like a simple traffic queuing program.

For com sci undergrads, the starting language taught was C, and very likely because they had eventually learn to understand, for example, how compilers are made.

 

[Rincewind]

For C, although not free, this book should be probably still the best (written by the creators of the language themselves):
https://www.amazon.com/Programming-Language-PROGRAMMING-LANG-_p2-ebook/dp/B009ZUZ9FW

gcc (free) and some simple text editor is all you need, you can use mingw on windows.

For LISP/Scheme, nothing beats SICP (and that's free too):
https://mitpress.mit.edu/sites/default/files/sicp/index.html

 

[Rincewind]

For com sci undergrads, the starting language taught was C, and very likely because they had eventually learn to understand, for example, how compilers are made.

Yeah that was in the very old times, even I started with Pascal first at uni. Nowadays they just start with Java or something like Python at most places. Maybe even JavaScript.

 

[Absinthe]

Mostly all wrong again, mate. Look, I'm not entering a "programmer's dick measurement contest" with you. For the record, and that's where most of your assumptions are wrong, I started with assembly on C64/Amiga/x86, then learned pure C much later, C++ even later, then gradually moved to proper high level languages, etc. And that was like 30+ years ago. I used to work on commercial C projects as a pro dev, so I know what I'm talking about. And yeah, I used to be a shortsighted "real men use only C" type of guy too, and probably past me would be pissed off of my above comments, just like you are now

I'm not a "real men only use C" kind of guy. There is such a thing as the right tool for a job and I'm not interested in a dick-waving contest with randos on the internet. I just think it makes a better starting point because it teaches you things you otherwise wouldn't bother to learn and understand, and those things can be important even when you move to higher-level languages with their conveniences because you've learned to keep into account the kinds of things that happen under the hood will write better software for it. I also think C is especially a better pick to learn if you want to write videogame engines. As you have reluctantly noted, C++ is in fact a highly popular pick for writing videogame engines.

And I think that many of your complaints are basically rooted in whining about the absence of high-level abstractions and conveniences, which is really missing the point of using C (learning how to do without them and be more directly responsible for the performance of your code, which can teach you good awareness and good habits even when you move on to other languages and learn how to do things that way).

Some words of advice: don't get too attached to programming languages, they're just tools. Your problem (which was my problem too in my "use C or go home" days), is that you're identifying with a piece of technology. They're just tools, and that's about them. They're not an extension of you. The more tools are in your toolbox, the better -- hard to argue with that. Again, this is not a contest, but based on your writing you're simply not in a position to judge the merits of different languages, because you're just not too familiar with them, otherwise you wouldn't have written what you did.

I do not recall opining on the merits of other languages, except to say that learning to rely on conveniences and abstractions teaches you bad habits. I agree with you on languages being just tools, and that learning a broad toolkit is a good idea, but that's far off. When a person wants to start programming, you can recommend one or two languages to start off with, but you don't recommend twelve.

Again, I just think C is a better starting point, especially if you're interested in writing videogame engines, where C and (especially) C++ are very much the preferred tools of choice. Even Unity with its C# fetish (btw C# is more related to Java than C, hurray marketing gimmicks) is itself written in C++.

Anyway, broadening your horizons would do you some good. So you can do that, or be the same guy I was, for whom it took 10+ years to realise there's life outside of C And when more experienced people told me essential the same advice I'm telling you know, I gave them the same bullshit as you... :D So maybe it's pointless, but some people might actually listen and learn something from this, who knows.

What are you, a Born Again Christian? No wonder you sounded like you had some kind of thing against C going on. "At first I used to love C but now I have Seen The Light and C is dark and vile and primitive, better avoided, and one day you too will realize this as I have, Brother." The grass isn't greener on the other side, dude. They're just tools, and they have their advantages and disadvantages. Just because you changed your perspective and preference doesn't mean you've been enlightened.

I'm just trying to offer a shortcut to newcomers, so they'll waste less time, overall. My advice still stands: learn C, and learn some LISP/Scheme (the order doesn't matter that much, but as I explained, I'd advise people to learn some LISP/Scheme first).

Yeah, both of them are worth learning, but they are suited for different purposes.

In closing, here's a "riddle" for you: imagine someone who is clever enough to write a LISP in pure C, which is clearly as high-level a language as it gets. Why did he do it? For one, he's really good at C, so he can't be an "idiot", can he?

Frankly, the world is full of people who manage to be clever and yet idiots at the same time. Turns out competence in one aspect is not competence in others, who knew. Also, what you're doing here is definitely a "programmer's dick measuring contest".

Peace brother.

Peace.

For C, although not free, this book should be probably still the best (written by the creators of the language themselves):
https://www.amazon.com/Programming-Language-PROGRAMMING-LANG-_p2-ebook/dp/B009ZUZ9FW

Yeah, that's the K&R book, and it's highly recommended.

gcc (free) and some simple text editor is all you need, you can use mingw on windows.

Yeah. In the interest of being a little more convenient for a Windows user and first-time programmer I'll recommend the Dev-C++ IDE (which also compiles C and uses GCC): https://www.bloodshed.net/

If you're a Windows user opting for text editor + compiler, you may as well use Notepad++ as your text editor: https://notepad-plus-plus.org/

 

[Krice]

It's better to start with C and move on to higher-level languages as desired than the other way around, because this way you start off with a better grasp of the fundamentals of programming.

These "fundamentals" (whatever they are) are not limited to low level languages. Most concepts you need to learn are in both low and high level languages, but you don't need to implement (only understand) some stuff in high level languages which saves time and allocates it into the actual game development. Just think about C and strings, then compare it to the ease of std::string.

 

[Catacombs]

For C, although not free, this book should be probably still the best (written by the creators of the language themselves):
https://www.amazon.com/Programming-Language-PROGRAMMING-LANG-_p2-ebook/dp/B009ZUZ9FW

The actual book is pricey, even on eBay. One is better off reading a PDF, which is free on z-library.

I'd also recommend Modern C. It goes at a slower pace than K&R, but it's thorough and readable.

 

[Azdul]

Also, "game engines" is a big topic. AAA rendering engines, yes, they are written in C++ (because of industry inertia and a lack of vastly better alternatives, not because those devs all really love C++). But for non-AAA games, especially if not 3D, I'd say C/C++ is overkill nowadays, you're just complicating your life for no real benefit.

In games industry, you start with high level, sane, portable code. If it too slow on target platform, you need to make it faster, using excessive force if necessary - preferably in next 2 weeks, or the whole project is scrapped.

After rolling your own containers, threads management, scheduler and memory allocators, and squeezing last few bits of the structure to fit into cache line - the source code barely resembles the language that you've started with.

There are games written in C# that use arena allocators instead of dynamic memory allocation, where garbage collector thread is disabled, and parts of VM are rewritten to be less portable and 5% faster. Or games written in Java - that are restricted to 2% of Java API. Or games where most of the logic is written in highly customized version of LISP, and then compiled. So you don't have to use C++ for professional game development - you just need to be brave (and slightly insane).

 

[ChaDargo]

The bad habits you learn in higher-level languages because of their conveniences can haunt you for a lifetime. If you want to make a game engine, you cannot afford those bad habits.

The languages that focus excessively on making programming "easy" usually make the mistake of trying to hide a lot of what code does "under the hood" and trying to use all kinds of conveniences to make it unnecessary for you to engage with bothersome aspects like careful memory management or learning how pointers really work among other things. But then you get used to all of those ridiculous crutches (and the bugs and inefficiencies that can come from relying on them) and never learn how those crutches operate or how programming works without them, which means you end up writing shitty code without even knowing what you're doing wrong, and if you want to write a game engine, efficient and bug-free code matters a lot.

This is so wrong on so many levels... Don't listen to this guy.

C is useful to learn, but it's essentially a glorified cross-platform assembler, not much else.

If anything, higher levels of abstractions, automatic memory management and safe language semantics (e.g. avoiding mutations and manual handling of concurrency) results in fewer bugs, not more.

Some higher level language constructs are almost impossible to handle correctly with manual memory management (e.g. closures). But even with complex object graphs, you will eventually end up writing your own half-assed GC algorithm that walks the object graph etc.

Ideally, you'd learn C just to have an idea about what the machine is doing, and something high level to know about the possibilities. The order doesn't really matter, but I'd say go for the high level language first, like Racket (it will be much more fun, and if you end up just learning a single language, it's better if it's a high level one).

Also, "game engines" is a big topic. AAA rendering engines, yes, they are written in C++ (because of industry inertia and a lack of vastly better alternatives, not because those devs all really love C++). But for non-AAA games, especially if not 3D, I'd say C/C++ is overkill nowadays, you're just complicating your life for no real benefit.

For C, although not free, this book should be probably still the best (written by the creators of the language themselves):
https://www.amazon.com/Programming-Language-PROGRAMMING-LANG-_p2-ebook/dp/B009ZUZ9FW

gcc (free) and some simple text editor is all you need, you can use mingw on windows.

For LISP/Scheme, nothing beats SICP (and that's free too):
https://mitpress.mit.edu/sites/default/files/sicp/index.html

So would you suggest I start learning with C or (my preference) with Racket, particularly How to Design Programs. I've already done a good amount of work studying Racket's first chapters, and a good amount of work finishing Python Crash Course. I like Python because it seems very easy to jump in and start making things, but I feel like I don't learn as much about the fundamentals as I have with Racket and HTDP. I also see the appeal of C for being closer to the hardware level and how influential it has been. I'm not really concerned with the idea that this or that language is better: I want the most efficient and helpful path to learn to start making games that aren't buggy pieces of shit.

 

[Rincewind]

And I think that many of your complaints are basically rooted in whining about the absence of high-level abstractions and conveniences, which is really missing the point of using C (learning how to do without them and be more directly responsible for the performance of your code, which can teach you good awareness and good habits even when you move on to other languages and learn how to do things that way).

I suggested to people to learn some high level language and C, remember? Those were not complaints, but statement of facts: having to do things less conveniently is one part of the issue, the other is the limited capacity of the human mind. THe mind has X amount of energy to deal with complexity; you can either spend that on getting bogged down on lots of low-level details, or you can work with higher levels of abstraction, that effectively gets "more mileage" out of your fixed amount of mental energy.

Again, I just think C is a better starting point, especially if you're interested in writing videogame engines, where C and (especially) C++ are very much the preferred tools of choice.

Yes, won't disagree with that, but I don't think this discussion was just about game programming (which is a very narrow niche of general programming), or game engine programming (which is an even narrower niche of game programming). Like I won't disagree that probably the best practical way to go about writing a video codec is C, but again, that's an extremely narrow niche.

Again, no disagreements there, I still think knowing C and understanding how the machine works at a low-level is important. But read back what you wrote, your initial reaction was a lot less balanced

"At first I used to love C but now I have Seen The Light and C is dark and vile and primitive, better avoided, and one day you too will realize this as I have, Brother."

Yeah, that's a pretty accurate representation of the sequence of events, actually.

Just because you changed your perspective and preference doesn't mean you've been enlightened.

Well, you'll only know if you go and walk the same path

 

[ChaDargo]

So learn C and Racket. The problem I've had with that thought is exactly what you said, my limited amount of mental energy/time lol

Seems like the best thing to do for me would be to focus on Racket, while learning bits of C. Thanks for the advice, everyone

 

[ChaDargo]

Just curious, Rincewind: why do you suggest Racket or LISP over Python? I think I have a good idea as to why, but it would be interesting to hear a more expert opinion.

 

[Absinthe]

You said you want to learn how to make your own game engine, so that's why I told you to just learn C, and I stand by that. Videogame engines are overwhelmingly written in C. Even the C++ engines tend to be written C-style with C++ features on top.

Learning Racket can do you good, but that's usually not a good pick for game engines.

 

[Rincewind]

So would you suggest I start learning with C or (my preference) with Racket, particularly How to Design Programs.

Haven't read HTPD, just had a quick look a while back but it seemed like good and fun introduction to Racket.

I'm not really concerned with the idea that this or that language is better: I want the most efficient and helpful path to learn to start making games that aren't buggy pieces of shit.

That's a good attitude

Just curious, Rincewind: why do you suggest Racket or LISP over Python? I think I have a good idea as to why, but it would be interesting to hear a more expert opinion.

I suggest LISP/Scheme over anything else because of the sheer mind-reprogramming it will do to you, for the better. Hard to explain really, but if you dutifully work your way through the exercises in SICP, it will change your outlook about programming forever and it will make you a better programmer in any language. When I first read that book, I was doing Java full time, and I was quite surprised how my Java style had changed as if by magic just after working through the exercises in the first few chapters. I know, it sounds weird and a bit esoteric, but maybe I got you interested enough to give the book a go and see it for yourself People used the word "enlightened" a bit pejoratively earlier in this thread, but that is what learning LISP/Scheme for the first time is: an enlightening experience about computation.

Python is okay I guess, it's good for everyday scripting and it has tons of useful libraries (one of the best things about is in fact the ecosystem), but the core language is a far cry from the sheer power of a LISP/Scheme. Having said that, I kinda like Python for what it does.

Some highlights of things unique to LISPs that are missing from Python:

- working on a program while it's running (that's right, changing data/functions/whatever on the fly)
- ability to write expressive DSLs (invent your own mini-programming language that describes the task you're solving much better than a generic language)
- macros (programs generating programs)
- no distinction between code and data (!); they're the same thing
- TIME TRAVEL!

-----

One more thing about writing game engines: if you're 100% sure that's what you want to do, that's of course fine and by all means go deep in C/C++. But probably there's hundreds (if not thousands) of non-engine developers to every single engine developer nowadays in gamedev (I don't work in gamedev, so I'm just making numbers up... maybe some working gamedev folks want to chime in on that. But there's a similar trend in practically every area of computing; for example, there are very few people working on actual web servers, but tons of developers writing web apps using those web servers etc. Similarly, how many people work of video/audio codecs in the world today? Very very few, but lots of programmers use those codec libraries, which is easy.)

If I personally wanted to write a game, I'd just use some existing engine. The thing with engine development is that it is very technical, low-level and math heavy, and you'll spend years just learning about the theory and tech involved. It's a very narrow, and all-consuming area, so if you're that kind of super technical and very much math-focused person, then maybe it is for you. But chances are that you're a person who just wants to create games and you are more interested in design, gameplay and such, in which case you'd be better off using some existing engine. I might be wrong, but engine development seems to be something that doesn't leave room for much else, and the barrier entry is quite high. I don't think it's a good idea for most solo devs, except for geniuses who only need 4 hours of sleep or something

Ah, and all that is assuming we're talking about 3D engines. Non-3D game engines can be definitely much simpler, but pretty much everything uses 3D nowadays anyway, unless it's 100% pre-rendered. So probably you'd be better off using some 3D engine anyway, but in theory you could write your own much simpler 2D engine from scratch.

 

[Rincewind]

Also, check out the comic on this page :D

http://landoflisp.com/

Click on the blinking links!

 

[ChaDargo]

So would you suggest I start learning with C or (my preference) with Racket, particularly How to Design Programs.

Haven't read HTPD, just had a quick look a while back but it seemed like good and fun introduction to Racket.

I'm not really concerned with the idea that this or that language is better: I want the most efficient and helpful path to learn to start making games that aren't buggy pieces of shit.

That's a good attitude

Just curious, Rincewind: why do you suggest Racket or LISP over Python? I think I have a good idea as to why, but it would be interesting to hear a more expert opinion.

I suggest LISP/Scheme over anything else because of the sheer mind-reprogramming it will do to you, for the better. Hard to explain really, but if you dutifully work your way through the exercises in SICP, it will change your outlook about programming forever and it will make you a better programmer in any language. When I first read that book, I was doing Java full time, and I was quite surprised how my Java style had changed as if by magic just after working through the exercises in the first few chapters. I know, it sounds weird and a bit esoteric, but maybe I got you interested enough to give the book a go and see it for yourself People used the word "enlightened" a bit pejoratively earlier in this thread, but that is what learning LISP/Scheme for the first time is: an enlightening experience about computation.

Python is okay I guess, it's good for everyday scripting and it has tons of useful libraries (one of the best things about is in fact the ecosystem), but the core language is a far cry from the sheer power of a LISP/Scheme. Having said that, I kinda like Python for what it does.

Some highlights of things unique to LISPs that are missing from Python:

- working on a program while it's running (that's right, changing data/functions/whatever on the fly)
- ability to write expressive DSLs (invent your own mini-programming language that describes the task you're solving much better than a generic language)
- macros (programs generating programs)
- no distinction between code and data (!); they're the same thing

-----

One more thing about writing game engines: if you're 100% sure that's what you want to do, that's of course fine and by all means go deep in C/C++. But probably there's hundreds (if not thousands) of non-engine developers to every single engine developer nowadays in gamedev (I don't work in gamedev, so I'm just making numbers up... maybe some working gamedev folks want to chime in on that. But there's a similar trend in practically every area of computing; for example, there are very few people working on actual web servers, but tons of developers writing web apps using those web servers etc. Similarly, how many people work of video/audio codecs in the world today? Very very few, but lots of programmers use those codec libraries, which is easy.)

If I personally wanted to write a game, I'd just use some existing engine. The thing with engine development is that it is very technical, low-level and math heavy, and you'll spend years just learning about the theory and tech involved. It's a very narrow, and all-consuming area, so if you're that kind of super technical and very much math-focused person, then maybe it is for you. But chances are that you're a person who just wants to create games and you are more interested in design, gameplay and such, in which case you'd be better off using some existing engine. I might be wrong, but engine development seems to be something that doesn't leave room for much else, and the barrier entry is quite high. I don't think it's a good idea for most solo devs, except for geniuses who only need 4 hours of sleep or something

Ah, and all that is assuming we're talking about 3D engines. Non-3D game engines can be definitely much simpler, but pretty much everything uses 3D nowadays anyway, unless it's 100% pre-rendered. So probably you'd be better off using some 3D engine anyway, but in theory you could write your own much simpler 2D engine from scratch.

Great points across the board that solidified my sticking with Racket (with some side effort in C). Perhaps an engine is too much to chew and not worth the effort; that's something I'll worry more about when I get more expertise in general.

I definitely noticed what you're saying about Racket. When I started, I was doing Python and although it was easy to figure out how to make some basic things work, it was a bit overwhelming. Did HTDP's first couple chapters, came back to Python, and I could easily create functions even for my own little library for a text-based thing I was trying to make. Racket seems to force you tou figure it out; Python seems to offer tons of tools/tutorials to avoid figuring it out.

I'll definitely be incorporating C into my studying, if just to be able to grasp the lower-level issues involved in programming. Thanks for the advice and the link to that comic. This thread has helped me a lot. Appreciate you guys!

 

[Rincewind]

I definitely noticed what you're saying about Racket. When I started, I was doing Python and although it was easy to figure out how to make some basic things work, it was a bit overwhelming. Did HTDP's first couple chapters, came back to Python, and I could easily create functions even for my own little library for a text-based thing I was trying to make. Racket seems to force you tou figure it out; Python seems to offer tons of tools/tutorials to avoid figuring it out.

Good to hear the magic is working for ya too!

I'll definitely be incorporating C into my studying, if just to be able to grasp the lower-level issues involved in programming.

Knowing C, Racket and Python will give you a very solid base for sure. Then you'll be able to make an informed decision on what tool is best for the job at hand.

Thanks for the advice and the link to that comic. This thread has helped me a lot. Appreciate you guys!

Glad to hear it helped, bro!

 

[Krice]

So learn C and Racket.

It doesn't matter that much if you are a beginner. You'll learn programming in any language, just pick one randomly. But maybe stay away from more esoteric languages like purely functional languages. They are better for specific tasks (science stuff etc.) and less practical in game programming.

 

[Hag]

I'll definitely be incorporating C into my studying, if just to be able to grasp the lower-level issues involved in programming.

Knowing C, Racket and Python will give you a very solid base for sure. Then you'll be able to make an informed decision on what tool is best for the job at hand.

Very good choice of languages. Racket probably is one of the best to get you started in development and just do things that works, its ecosystem looks great.
C is good when you want to go deeper and do your own low-level stuff, as well as getting some language culture since it is probably the single most important language of all times (what doesn't make it the best though).
Python is more on the meh side compared to the others but its has a "it just works" style that is really refreshing.

It doesn't matter that much if you are a beginner. You'll learn programming in any language, just pick one randomly. But maybe stay away from more esoteric languages like purely functional languages. They are better for specific tasks (science stuff etc.) and less practical in game programming.

Well, some language are better than the others, or simply less suited for studies. Java is ok for what it does but learning CS with it is pain and sadness. C++ is the quickest way of getting the worst habits. Javascript is an abomination on every level. Just to talk about ones I know about.

To learn development you need to study two things :
- Functions, algorithms and the like, which is easiest done in a lightweight language, and better even in a Lisp,
- And then implementation / low level, so you know at least how it works, and C (or assembly for the crazy ones) is your best bet.

 

[Catacombs]

but if you dutifully work your way through the exercises in SICP, it will change your outlook about programming forever and it will make you a better programmer in any language.

Did you read the entire book? I've read people say the first three chapters are good enough.

 

[Krice]

C++ is the quickest way of getting the worst habits.

I don't know why the language would do that, it's more about what the programmer is going to do. If anything C++ teaches you to avoid bad programming the hard way.

 

[Twiglard]

I suggest LISP/Scheme over anything else because of the sheer mind-reprogramming it will do to you, for the better. Hard to explain really, but if you dutifully work your way through the exercises in SICP, it will change your outlook about programming forever and it will make you a better programmer in any language.

Read Practical Common Lisp and PAIP. SICP may be too basic if you already have some programming experience. Learn Scheme only if it's a revision/implementation with record types. Otherwise there's too many problems with maintainability and bad habits.

Don't learn C as the first language as you're going to have problems knowing how to structure your programs while figuring out that you still need to zero-terminate after strncpy(3), and the difference between int(*)[3] and int**.

 

[Urthor]

Anyone here big on the Scala/Haskell train?

Decided I need to do some stuff in a language that doesn't look like my job, and hey the "objects and Java were all a giant mistake" guys seem to have it right.

 

[Rincewind]

Anyone here big on the Scala/Haskell train?

Yeah, been using Scala at work for 4+ years, quite nice. Couldn't go back to Java, really. Recently started using Kotlin, it's a bit like Scala-- but I quite like it a lot nevertheless.