Quote:Original post by hplus0603

Multiverse.net disagrees with you. Myself, I'm still undecided -- you'll need a lot of variation points in a generic client to make it truly useful for a variety of users, but it's possible that such a client could actually be created.

It's possible, I suppose, but I'm a bit skeptical. Different styles of games have different needs in terms of networking, etc. The way you'd write a game for a full, player modifiable persistent world is a bit different than what you'd do for something more resembling an MMOFPS, is a bit different from something you'd do for a WoW like game.

You could do a rather generic engine I suppose, that'd work for a lot of stuff, but it'd probably run into real problems at the edges. Still, it'd probably be enough for a lot of stuff that wasn't really pushing the limits.

I think that a very important point for a framework like this to be accepted is to make the components mostly independent (minus a small lib of common utilities).

Most of the frameworks I have tried have a bit of an everything-or-nothing approach. I want to be able to use my own network code if I want, or to make use of the music engine and skip the sound FX because I've already got a great wrapper around OpenAL. In the same way, I might like the network code on the server side, but reimplement from bare sockets on the client side, which means having the packet format well defined.

This is probably what I enjoyed the most with the Boost library: a set of utilities, that I can learn and use at my rate, and that don't give me the feeling that I'm getting myself onto the thin edge of the wedge.