32 replies to this topic

[DZee]

Ultimately it's a waste of time to spend so much time on different languages. If I had to retake my path I would have jumped into C++ and never looked back. In reality only the specialist really shine in what they do. There's no reason to try and be good at everything. My teacher once told me; sometimes you need to learn to love what you hate and embrace it.

That's what defines if you truly are willing to walk down the path you really set out for you. Don't plan things years ahead, I know plenty of things I wanted to get taken care of that never happened because my priorities changed. Just get started.

Edited by DZee, 10 August 2012 - 09:47 PM.

[ApochPiQ]

Ultimately it's a waste of time to spend so much time on different languages.

Patently untrue.

Even if you never wind up using most of the other languages, they will give you new ways to think about solving problems, and that will almost inevitably make you a better programmer.

[SharkBaitHooHaHa]

@Matt-D Thanks for the link, I'll be sure to check it out! I'm interested to hear what Carmack has to say about functional programming I read one of his articles on ray tracing and it was superb

@SimonForsman Have you seen/read the book "The Land of Lisp"? That's the book I currently own about Lisp but as I said I haven't been able to read it and right now I don't have access to it (this Saturday I fly back home to my collection of programming books <3 ). I mention this book because it teaches Lisp through game examples and I found it really interesting in that aspect. I think you might be interested in this one

@Hodgman Oh okay, so scripting languages are kind of the quick and dirty languages? That makes sense haha I'm guessing then for the functional programming
languages I just gotta jump in and see how I do it then. Thanks!

@Dynamo_Maestro I'll read about F# but I probably won't go too deep in it since I'm not that big of a fan of .NET but thanks nonetheless!

And thanks to everyone else that answered!

[SharkBaitHooHaHa]

@DZee I see your point there but I like to think of it as my Dad puts it,

"You can have large sea of knowledge, but have it only be puddle deep. Likewise you can have a puddle that is miles deep. But in the end, the most interesting sea is the one that is expansive and populated with shallows and deep ends."

I realize that I shouldn't try and be good at everything but I feel there is always something to gain from learning something new. Sometimes even just that joy of learning something new is enough to keep you going when you feel tired. I will probably specialize in a very limited field as you say, but I doubt I can keep my thirst for knowledge quenched. I wouldn't be surprised if at the end of my life I'm sitting in bed reading on the newest quantum computer or virtual reality gaming consoles(hopefully they get there by then!)

@ApochiPQ I agree completely with you on that, that's why I made this post

[3Ddreamer]

SharkBaitHooHaHa,

My philosophy for learning languages being the new learner in them which I am, is this:

Learning Languages
1) Need in the short term is survival.
2) Want is close behind because desire and keeping interest both drive you forward and inspire creativity.
3) Preparation for future demands keeps you ready for new challenges.

With these in mind I will be ordering my language learning priorities. I hope this helps.


Questions for the programmers: Would imperative programming add a strong tendancy for spontaneity in a game for the player just by nature?

1) What are the two most imperative languages?
2) What are the two most declarative languages?
3) Is functional programming independent of the previous two characterstics or not necessarily? To me this seems to depend on the language, is this not correct?

In the mid 1980s I studied Basic and COBOL, loving the imperative because I could either make something happen the same way every time or add randomness in the "Then" of the statement. This is why I believe a heavily imperative language is good for spontaneity of the player in the game while keeping order, too.

Recursion seems to be the ideal approach in many situations for complex circumstances, but would it not hurt game performance if used very much, risking FPS drop instances due to "perfect storms" of too much processing at certain moments?

It seems to me that imperative would be the best in general for game performance. Which of these mentioned (imperative, declarative, or functional) is the best for performance? ( All other things being equal.)


3Ddreamer

[DrMadolite]

"You can have large sea of knowledge, but have it only be puddle deep. Likewise you can have a puddle that is miles deep. But in the end, the most interesting sea is the one that is expansive and populated with shallows and deep ends."

Exactly. Once you know 1 language, the learning curve for the next language in the list just drops drastically. So the idea that knowing too many languages somehow gives you less understanding of each one, isn't usually the case (unless you insist on learning every single thing there is to know about 20 different languages).

====================

@ DZee:
Remember that your brain is an amazing device (in fact the most complex object known to man - a hypermassive star is less complex), it's vastly capable of storing information and there's even ways to "learn how to learn" things. I think you're making a solid point about specialization though, that it can be superior to generalization in many respects. But that's only true if we're talking about comparable time spent on either. Typically it's a transverging (? see below) relationship:

My ad lib theorem on learning:
"The more you generalize, the less time you spend on learning each new field.
The more you specialize, the more time you spend on that one field."

The golden mean of the above is when you've specialized enough so that further specialization becomes less effective than studying other fields to complement your own. Specialization is usually all about what you're actually doing. Generalization is usually about what you know. It's when you let generalization dictate your actions or specialization dictate your knowledge, that problems occur.

Edited by DrMadolite, 12 August 2012 - 03:19 PM.

[jtw]

C, C++, Assembly

In that order.

With C and C++ you can create whatever you can imagine.

With assembly added you will be a god instead of an angel

Edited by jtw, 12 August 2012 - 11:36 AM.

[hupsilardee]

Seeing as this is a game related website, I think a shader language should be on that list as well. But the way things are at the moment, the shader language that gets used is dependent on the graphics API...

[3Ddreamer]

husilardee,

That is an interest point.

If the developer wants to have his game cover two different APIs, then is there really a single shader language which will allow it?


3Ddreamer

[SimonForsman]

husilardee,

That is an interest point.

If the developer wants to have his game cover two different APIs, then is there really a single shader language which will allow it?

3Ddreamer

CG works with both OpenGL and D3D, (I'm not sure where it stands featurewise these days but nvidia is usually decent at keeping their stuff up to date)

[Telastyn]

I'd boil it down a bit:

• C
• C#
• JavaScript
• SQL
• Haskell

I'd be okay if some of these were swapped with others. C++ could go in for C, Java or Python for C#, F#, OCaml or a lisp dialect for haskell. Scala for... Scala's not a good replacement really; it kinda satisfies the modern OO and functional language areas but is 'odd' enough to not translate super well to each.

And you should know enough HTML, CSS, and regexes to make stuff work. These will put you in a solid, well-rounded future. Assembly and Prolog are cute but you can get away without them. You can even get away without 1-2 of the items in this list and still be a successful, productive programmer.

[3Ddreamer]

husilardee,

That is an interest point.

If the developer wants to have his game cover two different APIs, then is there really a single shader language which will allow it?

3Ddreamer

CG works with both OpenGL and D3D, (I'm not sure where it stands featurewise these days but nvidia is usually decent at keeping their stuff up to date)

After a little research, am I correct on this: CG (or Cg?) requires a compiler with an OpenGL or DirectX shader program? The vertex or pixel shading might be very important for certain things, huh? ...indispensible perhaps?. Would creating a laser beam and other types of light beams for a game be something which could use this? Somebody I know is very experianced in graphics and reminded me a few times to learn about vertex shaders. I am still trying to understand the applications of it.


3Ddreamer

[Hodgman]

That's a good point about the importance of 'shader languages' as a category.
Everything that happens on a GPU is controlled by shaders (vertex/pixel/compute/etc), so to learn the GPU side of graphics programming, you have to learn one.
Vertex shaders are typically used to convert 3D points to 2D points on a screen, which the GPU then uses to draw triangles. Pixel shaders are used to calculate the colour of each pixel covered by a triangle.

nVidia Cg can be used in both D3D and GL, but nVidia Cg is almost exactly the same as Microsoft HLSL (they created the language in cooperation, then both used different names for their implementations of that language). On every cross-platform game that I've worked on, we've written the shaders once and compiled that code as HLSL for D3D, and also compiled it as Cg for GL - with a few #ifdef's here and there for D3D/GL differences. That's how simmilar the two languages are.