It's a whole hell of a lot easier compiling dependencies in Linux. 99% of the time they are going to be using the make/automake/autoconf toolchain, so it's as easy as:

./configure && make && make install

Repeat this for each dependency. It seems like a hassle, but is nowhere near the hell I experienced trying to get Windows C/C++ libs to compile and cooperate.

Here, I will sum up most peoples' timelines:

Started programming in language X when I was a kid or relatively young, or used game making tool Y. I have since moved on to a "real" programming language / environment.

Spent the next 5-20 years getting better. Along the way, I created tons of unfinished or half-finished projects. I began to wonder how people can manage to create exciting, polished projects in such a deep and difficult field. I started to get distracted by other non-gaming opportunities because I didn't feel like I could cut it.

Eventually I made something that's "complete". It may or may not be very impressive.

I still feel like I don't know anything.

Lua is best if you have a large C or C++ library which you want to wrap and don't need to leverage any libraries you don't already link to, . It is one of the fastest scripting languages out there, and really simple and no hassle at all to learn. Drawbacks are that it doesn't have "batteries included". It's a perfect fit for most games (in the big leagues at least), because they are going to be coded in C++ by default so the code you want to wrap is going to be sufficiently rich that you don't need a lot of the extra power or external libraries the other languages give you.

Python gives you a huge number of libraries and a more powerful set of operations in your scripting layer, including a very flexible OO model. Drawbacks: it's slooooooooooooooow, and while it's easy to learn the basics it can take a long time to get the hang of using some of the more powerful features, like 'special names' in classes.

Ruby, I have no experience with. Sorry.

Yes, there are many alternatives to SDL that aren't as low-level, as you've just mentioned. I applaud your insistence to not reinvent the wheel, but don't worry too much about finding the "perfect" library before earnestly starting your next project, there's no real silver bullet out there. Just find one that seems cool and easy to use after reading the FAQ and a tutorial or two.

I use Code::Blocks for C++ and for when I occasionally try out some D. I prefer SPE for python, and I'm always willing to use gvim for editting anything if I don't feel I need code completion. Wings3d for modelling, Gimp for 2D, nothing really fascinating.

Actually, it would be a great example of a positive-sum game (i.e. cooperative, as opposed to zero-sum or competitive) if you had to somehow get all these warring / competing factions to cooperatively create sufficient technology to fight / meet the alien race, perhaps restricted by some sort of time limit.

According to the IGDA site, they are in my area: http://www.igda.org/nctriangle/

I can go over and smack their mailbox with a baseball bat or something.

The demakes compo is such a good idea that it should probably be repeated every year. So long as companies keep making games, we will have source material.

I punched a horse in the face.

Right in it's face.

Fuck yeah. :guitar solo:

Wow, there was a demoparty in San Jose? I would have never guessed.

Sucks that I live in North Carolina...

edit:
It's cool how much more awesome demos are on a big projector... I especially like this one:

A model-viewer-controller setup is probably the best way to go.

The more I program, the more I realize that separation of concerns is probably one of the hallmarks of good code.

It is certainly tempting to want to centralize everything into a God Object or some big bad Manager class, but as your code gets more complicated and you want to do more things, you will realize that keeping all your parts separate is a huge boon.

As for the original topic, I usually choose the path of "has" (containment). As for worrying about loading the same image 20 times for 20 instances of the same enemy, I think it's best to code your resource manager to cleverly pass references/pointers to an already-loaded image when you ask it to load the same image 20 times.

I mentioned a little bit about my stance on what I call "idea men" a few weeks ago on a post about procedural generation. Short version, I'm with you on this.

Wow, nice post! I pretty much already knew all the pitfalls you were talking about, but it's good to see someone state all of these in one place, for the benefit of those who don't know better.

Reminds me of Dave Dobson mentioning to me once that "sometimes it takes just as long to create content generation code as it does to just create the content yourself".

From: http://science.slashdot.org/article.pl?sid=09%2F01%2F07%2F2130209&from=rss

What we really need is an alien race to show up, blow up a major city or two, leave us exact directions on how to get to their home planet and specs on what sort of weaponry they have, and then leave us alone for about 200 years. That's about the only way I can see the military getting into manned space travel in a big way.

Sounds like it would make an interesting game, sort of a military sci-fi version of Buzz Aldrin's Race to Space. The only problem I see is that there is a whole lot of nonviolent buildup for some sort of violent climax at the end. I don't see any obvious way of fixing this pacing problem except by making some sort of ongoing conflict (violent or non) during the buildup.

Hey there! Good to see a demoscener come into the indie gaming fold. I went to Buenzli in 2006 when I was studying abroad but I haven't got the cash to finance flying all the way to Europe for demoparties. I just need to get one of those "job" things over there...

[must]

Unsuprisingly, you can. If you don't declare any shapes Box2D won't do anything, and you can declare everything as sensors there will be no physics done at all, just collision detection. Also, Box2D's broadphase is a standalone component, which can be useful on it's own for simple collision applications.

I'm coming from a Chipmunk background, where by all accounts it seems like you must have a body (physics object) to have a shape (collision object). This adds a lot of unnecessary overhead, or does stuff like push around / spin objects when I don't want them pushed around / spun. I would much rather have a physics engine from which I can remove the collision responses and just keep the collision detection. I will have to check out Box2D.

You're adding evidence to the argument that designers are dead weight.

That was a little harsh, I'm not trying to be an asshole, I'm just tired of "hey I have this cool game idea now I need people to make the game for me" posts. This is the thing that made me start programming; you can't have a game without a program of some sort, so it felt like the most direct way of implementing the game design ideas I had in my head.

Also, as you might notice, spinoff games are looked down upon in these here parts . I would strongly suggest you look into making this as a mod using an FPS engine (Source, Unreal 3, Quake 3, etc.). They are fully capable of going from first- to third-person, and by what you're saying it sounds like they would provide everything you want. There is a ton of code that you or the programmer would have to create which is already readily available in an FPS engine, you would honestly be saving yourself several years of development time.

Hey, Ars Technica interviewed Jeff Rosen (Wolfire Games, makers of Lugaru) and he pointed out some telling figures suggesting that you should make your next game cross-platform.

Indie dev suggests peers should support OS X, Linux gaming

I guess every time you make a Windows-only game, God kills a kitten.

[Innovation Award]

Innovation Award
http://www.igf.com/php-bin/entry2009.php?id=684

Hell yes.

Holy crap, this paired with some procedurally-driven IF or something could create a very interesting game.

[velocity]

Alternatively, it could be made so that the velocity of a note (which is usually one byte if I'm not mistaken)

Fixed.