23 replies to this topic

[fireside]

I think the jury's out on rouncer's advice (even aside from alphadog's logic point) until he actually finishes a project. I can see potential there, but I don't think I could say there's an advantage to doing it that way without anything being finished. The time frame given is probably about right, though. It would probably take a year or two to learn an engine well, but about 10 to write one. There is a lot of middleware around for c++ engines, though, so I don't know, but that's like using a lot of smaller engines. With physics, terrain, character motion, etc, it's kind of getting beyond one person anymore.

[rouncer]

alphadog said:

Including you? :)

No I listen sometimes, just I use my own brain too. Personally im on my 11th year programming for windows and direct x, I feel ive accomplished alot, but yeh... no game yet. Maybe this year? I might be wrong, but, I really think using an engine limits you and is superfluous to a small game anyway.

[TheNut]

fireside said:

it's kind of getting beyond one person anymore

I think its been like that for a couple years now, probably since SM 2.0 was announced things started to explode.

rouncer said:

I really think using an engine limits you

Depends how you look at it. Engines catered for a specific genre are obviously limited. A framework however is vital if you want to develop in a reasonable time span. Having a foundation to build your games on goes a long way to save you time, so choosing an open-ended engine that you will grow with would be extremely beneficial down the road.

[alphadog]

rouncer said:

I really think using an engine limits you

Depends on the engine. If I grab OGRE, Havok, RakNet, etc. I get complete best-of-breed stack that, with some elbow grease, integrate into something that doesn't limit me at all. And, if it somehow does (I'm hard-pressed to see how in any but edge cases), then hopefully I've integrated to interfaces, meaning I can pull out the item and swap in another in many cases.

Instead, you recommend building a renderer, a scene manager, an animator, a networking stack, a physics engine, an input/output manager, an audio stack, etc. Integrate each piece. Validate it for security and stability. And then you build a game?

rouncer said:

and is superfluous to a small game anyway.

How small is "small"?