Disclosure: I worked on developing VehicleForge.

Matt: it's great that you are excited about FANG. It's a really cool project they've got going there.

0th: Though my team's solution didn't get picked, I have some insight into the technology, and even rapped with Marcin a bit about possible options for employment of our solution with OSE... our code still hasn't been released by DARPA, so that's kind of in a holding pattern. The parts we wanted to get in there were a distributed model (i.e. many forges), much deeper semantic search, and plugin-less realtime collaboration on design assets. I have been working on and off with the allura team, but their focus is still pretty firmly on making the best darned free software forge (i.e. sourceforge, github, etc), and big-system problems (i.e. hardware, etc) would still be underserved.

First: you basically can't participate in FANG if you are a non-US person: citizen or green card holder. That's the problem with building a swimming, jet-powered tank as the first experiment. When the challenges are done, the site may very well go away, or be commercialized... there are already some folk that are planning to do that.

Next: the bulk of the goodness in the frontend of vf is in Apache Allura:
http://sourceforge.net/p/allura/wiki/Allura Wiki/
The team that is executing for the FANG challenges has done a pretty hard fork (last year, in about March), which is probably no longer compatible with the Allura baseline... and their code has also yet to be released. The things they added were mostly shinies (allura is a bit cumbersome to skin), backend (openstack cloud hosting) and social bits (twitter-like stuff). However, there are also some STEP viewers and other goodies. The search is also better. There is a 3d collaborative viewing piece (i.e. could replace how google docs is currently being employed for the backhoe), and while the guys that built it are super cool, the 3d client (built in unity) won't be open source, and the licensing fees for the server are pretty huge.

Also: The part that isn't in Allura is the backend and fat client tools. They include technology for doing design synthesis from many parts, and telling you if a design is feasible to manufacture. It's pretty cool. However, the client doesn't run on Linux, and I think actually requires VisualStudio.

Finally: On some of the threads here:
http://forum.opensourceecology.org/discussion/845/proposal-phase-ii-for-open-source-ecology039s-online-app
I have proposed investigating allura as a solution for rolling up the core collaboration capabilities for opensourceecology.org into a single user experience: wiki, forum, tasking, repos, site-wide search, etc. Additional tools are easy to add as python libraries, a language which is well-suited to much of the tooling OSE is helping to mature (FreeCAD). I still think this approach has merit, but if you go check the thread, there is a lot of debate as to whether this actually meets the long-term goals for OSE: eventually, we need something that can be deployed to a Village with no connectivity, and still be "connected" to the global community... but a first step of consolidation of all the *.opensourceecology.org apps would go a long way to increasing project cohesion.

Would love to rap on this further! Or, if you see things you like as a FANG competitor, post them, and we can see what works for OSE!

Cheers
Nick