Open Source Development

#1 My own feeling about open source is that it works best when it is standards based, evolutionary rather than revolutionary, modular, and directly useful to programmers.

By being standards based it is possible for all who work on the project to agree to the details without too much communication.

Open source works very much as an evolutionary market based system, with programmers dedicating their time resource to successful projects. This also means that the end product must be of use to the programmer themselves, as with money not being involved there is no feedback from 'customers' who do not dedicate resources.

Modular systems are good, as they cut down on the communication required. Linux distributions advance fairly quickly because their base components are just that, lots of components.

Games as a whole do not fit into most of these categories. To some extent neither do game engines, though API's such as OpenGL, OpenAL, SDL, etc do.

So if I were to decide to make a game with other people to show off our talent, the Open Source method might be applicable - but it is not intrinsically ideal.

However, a large number of groups working on individual projects based around a common, functional, core is a great way to work. GarageGames is doing this, and although it's not open source it does have many benefits for you if you wish to show off your talents.

Had I founded EnkiSoftware now, I would seriously consider using the GarageGames engine rather than building my own.

Doug EnkiSoftware Limited

#2 I kind of scare myself sometimes. Last night I really was hell bent on starting something of an open source game. Basically I was thinking of starting the design and implementation of the largest online game in the world (in terms of users) based on some of the idea's I've been thinking over for a possible book article about massively scalar multiplayer games.

I may even get a sourceforge site setup for this. I just need to find the mathematical basis for some of the game design elements (its going to rely on mathematical idea's for a lot of its elements, and possibly genetics and AI grafted into that whole mix).

Phil.

#3 The frame work for massively multi player systems, particularily ones which are more peer to peer than pure client server, would make good open source candidates in my opinion. I've got my own ideas about how to go about this, but far too little time!

Good luck to ya Phil...

#4 I love the idea of open source with any kind of software and I think with a lot of work it might be able to work.

However my opinion would be to have an open source game engine where anyone submits parts that they want in it then a team of programmers and testers review it and add or throw out that piece of code.
Then after the engine is done perhaps a team of developers could be selected from those who submitted parts to the engine and then a game could be developed and published.

just a thought =)


- ian wheat

#5 phil, doug,

You may want to take a look at www.worldforge.org. Specifically, Phil, it may have been what you were talking about.

just to let you know

#6 I'd just like to take a moment to point out that there is a difference between licensing(GNU, BSD, proprietary) and development models (bazaar, committee, lone nut coding until 5AM).

#7 Yacine: Yeah, Ive seen worldforge a while ago. Its not quite where I was heading (my idea was to use more of a peering model so that you could scale to essentially infinite proportions).

I just dont have time to do it though. Perhaps if I get to write the book article I'll try and do a very simply ascii style version of it.

ascii explorer or something.

Phil.

#8 Yacine: I've also been aware of Worldforge for a while.

Eric's correct to say that license and development models are different things, although certain styles of development benefit from certain licenses.

You could build an engine and give out the source under an open source license, as Quake and others have done, in a philanthropic measure (which also has some nice publicity with it). But this doesn't help build a game. So you'd need to have a development model such as the bazaar along with the open source license. The evidence seems to show that this is a good for some software but not for all. And games are more than just software.

I'm more interested in open source models which put together small APIs to help make games with, such as OpenAL etc. I have some ideas I'd like to work on but time a problem at the moment!