next »

[powly]

Not to bitch about Unity or any middleware but you all of you who haven't already done so should definitely try writing even a simple 3D system from scratch (using only 2D drawing commands, namely horizontal/vertical lines or even dots). You won't regret it. It's 1) a wonderful experience and 2) very useful because you actually understand some stuff that's going on under the hood (projection, lighting, transformations) Next thing you know you'll be writing your own full-blown engine. It just is that much fun.

[Tycho Brahe]

Yeah, I started off with pure opengl and c++ and really enjoyed it

[Martin 2BAM]

I also enjoyed it a lot (pseudo-shaded points and lines in DOS mode 13h and Borland's BGI, then OpenGL).

But it depends on your target: If you want to make technology or games.
These tools allow you to skip the technicalities of making games and just make games.
If you code it yourself, in most indie cases it will never reach the point where you start coding the game, because it takes a professional to create a confortable and useful engine (I say in most cases because there are exceptions)

[nihilocrat]

Not to bitch about Unity or any middleware but you all of you who haven't already done so should definitely try writing even a simple 3D system from scratch (using only 2D drawing commands, namely horizontal/vertical lines or even dots). You won't regret it. It's 1) a wonderful experience and 2) very useful because you actually understand some stuff that's going on under the hood (projection, lighting, transformations)

Yeah, this is a pretty fun thing to do. It makes you realize that raster graphics really are just smoke and mirrors. When I wrote a 3d renderer from scratch (just wireframes) I actually rotated the world around the camera, I forget exactly why but it seemed more natural at the time. I even got a cheap "zoom" function by playing with the projection of the camera.

Next thing you know you'll be writing your own full-blown engine. It just is that much fun.

Uh, no, no... this doesn't happen. You realize you've spent months coding an engine and not a game, and an amazing feeling of emptiness overtakes you as you haven't really gotten anywhere. More importantly, your engine kind of sucks and has performance issues, can't render relatively simple scenes without dropping below 60fps, etc.

Well, if you're me. I'm so tired of writing engines but never writing games.

Now, onto another Unity annoyance: apparently it's not really built to be used with source control systems, the only problems I've experienced personally are that I have to check in all the metadata directories, but it often complains that they're already added and blocks my commit. I'm not really sure what's going on under the hood so I'm not sure why any of that is happening. I'm going to try experimenting with Bazaar instead of svn for a bit. I like svn because it's faster and TortoiseSVN is more mature and friendly than TortoiseBZR, but bzr does a better job of handling a chaotic file hierarchy with stuff moving around a lot.

[Hayden Scott-Baron]

One of the nicest things about Unity is that it is Mac and Windows compatible, and there are continued efforts to ensure that the output works on crappy netbooks as well as the latest Win 7/Vista based desktops.  On the other hand, there's no Linux support, and the presence of a Mac version of your game will only make those guys feel even more annoyed/entitled.  Being able to compile both versions from either PC or Mac is great too.

Now, onto another Unity annoyance: apparently it's not really built to be used with source control systems, the only problems I've experienced personally are that I have to check in all the metadata directories, but it often complains that they're already added and blocks my commit. I'm not really sure what's going on under the hood so I'm not sure why any of that is happening.

Unity simply doesn't work with source control. It's doing way too much stuff with the folder structure and it breaks things like SVN.  If you get Unity Pro you can get SVN support. There is also a dedicated Unity Asset Server, but this is $500 per seat(!), not that great, and I think also limited to pro users. 

[Melly]

While I personally have only partial experience with it, writing your own engine, be in 2D or 3D, can be great fun at first, as you feel a constant sense of accomplishment at seeing stuff working on screen.

then you run into the problems.

Sometimes your engine just has huge performance issues, despite working. As you study why it might be, you realize you'll have a ton of code to rewrite completely. That feeling of constant accomplishment vanishes, sometimes for weeks or months, as you rework the damn thing.

Then, like nihilocrat mentioned, comes the fact that after a long time of development, you have made no game content, and your engine is still a lot worse than most others that exist out there and can be used for free. That's when you feel you've wasted your time.

So, what I'm saying is that, unless you're NOT interested in making a game, don't bother coding the engine as well, unless you're also a badass coder that already knows a lot more than average about this stuff.

[deathtotheweird]

Not to bitch about Unity or any middleware but you all of you who haven't already done so should definitely try writing even a simple 3D system from scratch (using only 2D drawing commands, namely horizontal/vertical lines or even dots). You won't regret it. It's 1) a wonderful experience and 2) very useful because you actually understand some stuff that's going on under the hood (projection, lighting, transformations) Next thing you know you'll be writing your own full-blown engine. It just is that much fun.

May be fun for you, entirely frustrating for others. Especially if you just want to make a game.

[Tycho Brahe]

Exactly, everybody's different, I personally would rather be making an engine, than making a game!

[powly]

Ah, by a 'full-blown engine' I meant a game-specific engine, not a general one. A one that fits just your needs for that specific application. And if done so, it will actually lack all the useless bloat offered by other peoples' engines; it will be, in fact, faster, lighter, smaller and as a free bonus you can boost your ego with it ('hey, look at what I wrote!' :D)

To put this to the right context, I began programming inspired by games but quickly came more graphics/demoscene -oriented so it's just stuck in my brain that I mustn't use any excess libraries and write the most optimal and small code I can. Many of the people in TIGForums seem to be more like artists, designers or people with lives; engines are just the right decision for them as they were designed to ease out developement, not to give you the tingling sensation of writing your own matrix library.