Steamworks looks nice, but I don't think that it does what the thread author suggests.
For one, Steam will never be a good marketing tool or revenue source for free games. Without ad revenue, Valve will never make any money off of free games, so it doesn't make any financial sense to give them the kind of visibility that bigger indie titles get.
One thing that's nice about GameTap is that everyone can rate every game, and it's easy to sort games by their rating. So it's easy to jump in and download the highest rated game that you haven't yet played. Still, GameTap has too much of a gatekeeper.
That leaves Greenhouse and Kongregate. Greenhouse has 1 game so far. Kongregate says that they will never support exes.
I say that we build an application to do this. The code itself wouldn't be that hard to write. I think that the difficulty lies primarily in centralizing and organizing the resources necessary for the system to operate fluidly.
So what resources would be required, and how could those requirements be met?
1. Cheap bandwidth and storage for the games and client app. The obvious solution is Amazon S3 (
http://en.wikipedia.org/wiki/Amazon_S3). The less-obvious solution is to have each developer provide their own hosting service.
2. Database of users, ratings, games, comments ... etc. I think that this could also be hosted by Amazon S3, or, if we don't want to invest in our own server space, we could possibly find a friendly site to host the db (TIGSource? I dunno). A key thing to note here is that all users will be able to submit games, and only the other user ratings will differentiate them. There will be no gatekeeper as in TIGdb.
3. An advertising solution. Could be a custom one, or something sourced to another company.
4. Programmers to write and maintain the code. These could be culled from a community like this one.
5,6 ... things I am surely forgetting.
I would gladly commit some coding and db-design time to a project like this, if other people would be willing to help solve the other problems.