Thief: The Fan Made Project  
For a game, we need a game engine.
In general, there are these options:
1 write our own engine suits our purposes best
complete control
takes a big fat lot of work and time
2 use an available engine quick project advance needs modifications
3 use and modify the Thief engine original look and feel
available now
legal problems
extremly hard to extend and modify from a certain point
will become outdated and cannot be updated without source
From the current list of people willing to contribute to the project, I see we have almost no experienced programmers. Some are willing to learn a programming language or have already started, but we would need programmers with at least three or four years of experience - and about half a dozen minimum.
Compare what resources it took LGS itself to get Thief I done... we certainly can't beat them on efficiency if this is a volunteer project with members all over the globe.
Having the Thief (II) sources would certainly help us with the third option, but I don't expect we can get them in a way we could use them without legal problems. And despite the game is called "Thief", I believe most of us adhere to local laws as long as they make sense. I, for one, do.
We want the result to be distributable, ownable and playable without risk of legal trouble. So even if some good soul could give us access to the Thief sources, we couldn't use them above finding out how certain things were done - we'd probably needed to buy the license.


For the certainly interesting and motivating option 1, we lack human resources.
For the obvious and promising option 3, we lack money and opportunity.
So we are left with option 2. Which isn't a bad one, considering most of us are well capable of fine-tuning some parameters with incredible patience, but few are able of writing code or affording large sums of money. It fits best our demoscopic structure.
I 'd like to thank Matt Thornberry for his links to the 3D Engines Collection, and others for their helpful contributions. I am still in the process of evaluating available free engines. From the commercial available engines, I would select the Quake(3) engines. It is very customizable, there is are plenty of good tools for creating models, maps etc for it, and it is technically really good, stable and efficient. Aside from the basic consideration of it being both available on several platforms and using OpenGL. So, for now, I am still evaluating...