It's probably somewhere in between.

You'll likely learn a fair bit, since something of that size is useful for design practice if anything.

That said, you're mostly likely guaranteeing that the game won't be finished for years to come.

So many times I see people post on here that they want to make a game so they set out to design an engine right off the bat. A game engine is meant to be a reusable library that can be the basis of several games. It shouldn't be a one-off deal. And making an engine when you really want to make a game is often setting yourself up for failure.

I think it's always a good idea to set out to make games rather than engines when you are just starting out. Learning the concepts behind engine design & architecture comes much more easily when you have knocked out 3 or 4 games. If you've never made any games, you have to rely on studying other engines to get an idea of how to architect your own - otherwise you're just guessing at what you need. And engine A will most likely do things differently than engine B, which is structured differently than engine C. Getting some games under your belt will show you the common systems between games, what's reusable, and give you some ideas on simplifications and generalizations of your own. It can also help you to understand the differences between other engines out there.

Architecting a good engine is no easy task. Your frame of mind is just as important in game programming as your knowledge. If you view the problem from the wrong angle, you'll wind up with a skewed solution. It's much simpler to think in terms of 'game' than 'engine' in the beginning.

Of course, to each his own, YMMV, yadda, yadda. Some people are gifted enough or persistent enough to pull off anything they set their minds to. Just make sure you understand the big picture before diving in head first.

I was interested in writing a graphics engine, and so, because I had worked with irrlicht previously, that's the one I focused on. Now, if you go to the irrlicht site he's up to version 0.12 I believe, with shaders and everything.

There is no way that I'd be able to understand all the code for all of that, so, I went to sourceforge and downloaded the source for irrlicht version 0.1

The first release, and I've found that looking at that code is not incredibly complex, but it is useful enough that I just stole the class heirachy and structure, and now I have my basic design firmly imprinted in my mind, I rarely look at the code.

Moral of the story: If you see an open source engine that looks good, find its first release and look at that code.

Oh, and if you are starting from scratch, probably forget the networking and one or two other things, because, once you have a framework, or an engine (partial or completed) that you are satisfied with, you can re-design, re-write it to incorporate networking, and you'll have a much easier time with the graphics, so you'll be able to spend most of your concentration on learning the networking part.

My opinion. [smile]

Why is everyone saying it takes years to develop an engine and make a game with it? That's bullcrap. If you're wanting to write a full blown engine with a bunch of bells and whistles, then yeah... but an engine for a first time game maker? Shouldn't take too long unless you're trying to write a Half-Life 2 clone. The current engine for my game is rather simple, and it is reusable. It didn't take me years to write (granted i've been working on the game for about a year, but maybe for a few hours every month... i'm very lazy, heh).

Write a simple engine that doesn't have real-time lighting, doesn't have bump mapping, etc etc etc... and you should be good to go. Start small.