Thief: The Fan Made Project  
I think I should make my reasons clear. I do think we need a number of missions, woven into a good story, with cutscenes and the like. However, Digital Nightfall already gathered a group of well known and qualified people to make CoSaS. I am planning further ahead.
The Thief I/II engine always was too limited, at least in a graphical way. Also, there are some artifacts from kludges LGS needed to make the baby go gold in time, I gather.
The players demands will increase with time, no matter what - there are still some who play old Amiga games, out of nostalgic feelings, or because some things were never made that good again. But the general audience moves with the time. The technical-graphical part of Thief always was not the state of the art, and this gap will widen fast.
The current Thief engines use Direct3D. I think using Direct3D is a mistake in general, because it is more trouble for everyone involved. Especially for the player - DirectX isnt even compatible to itself, let alone being portable. Remember that Thief broke when a new Version of DirectX came out? This wasnt because of Thief. LGS made a patch for us then, but if this happens again, it will be hopeless. All new games will force us to update that DirectCrap, and the Thief game will be lost.
With 3D graphics, OpenGL is the way to go. Everybody who really knows how to use both interfaces will probably agree with me there; if you don't, please mail me, and explain. For those not into graphics programming, you may want to read Carmacks statement on DirectX vs OpenGL. While not everybody likes the games id software makes, from a technically perspective they know their job damn well. The article is a bit dated, but I feel still valid.
If we want a better AI, more functionalities, different display stuff, new object types or fix some of the bugs and weirdnesses the Thief engine still has, we cannot do it without the engines' source - which we are not likely to get soon. So IMHO the only way that makes sense is to use a new engine.
Additionally, some still want Thief Multiplayer. While I personally can't imagine how this would feel, what the rules/goals should be, I think the possibility is not bad by itself.
I know that splitting weakens us, but in the long run, I see no options other than a new engine or dropping Thief completely and hoping for some wonder.
I personally don't believe in wonders, but gladly accept if they happen.
The low number of programmers volunteering almost rules out making a completely new engine. So we are checking available free engines. The fallback could be making a modification for an commercial engine, or use the GPLed source of an former commercial engine.