The only thing which really matters for ARM GPUs is how good they are against the best their rivals can put out there. If Imagination Tech, NVIDIA or Qualcomm chips have better price, performance and power requirements, then few companies will use ARM over their current chip providers (which is the case today). As for these claims; they are entirely believable and nothing special. Their competitors are all claiming similar things, ARM are just making more noise about it.

The success rates vary from year to year, but last year we (Crystal Space) had 6 successful students, all of whom had their code integrated into trunk development very shortly after GSoC ended.

I don't think you can totally judge success from how the mentoring organisation benefited however. Imo the main point of GSoC is to give students experience in working with/on open source projects. There were a few discussions on this at the GSoC mentor summit last year, I think a lot of people (not really orgs, just people looking in at the program) miss that success should be partly judged by how well the students 'grew' as developers (both technically and in how they interact with other developers), not just by whether they met their project goals and contributed something useful to the organisation.

Making a graphics engine is hard and costs a lot of man hours (thus lots of $$$). There's not many people who can just start contributing to them (compared to other OS projects). The Open Source engines will always be at least a generation behind, simply because they're always going to be slowly implementing what's already been done in the commercial engines, while companies like Crytek are busy working on their next-gen stuff.

On the plus side, the Open Source engines (Ogre and CrystalSpace anyway) are good enough for people to make decent looking games if they wish to do so. Gameplay is what counts right? I'll take TES: Oblivion quality graphics (hell, Morrowind even) if the game play is great. Unfortunately making games is as hard as making the engines that they run on...

I believe it would appear as if the distance between objects had decreased. The speed of light would also always measure the same.

Works fine for me... Usually I run it on my PS3 so I can watch stuff on my TV (so I haven't noticed it being CPU hungry).

The more money the BBC can pull in from stuff that my taxes (TV licence, which I don't object to) have already paid for, the better. I realise that the TV licence won't drop in price, nor will I get any money back, but perhaps the money will go on creating more good stuff for me to watch.

Nah, probably the BBC execs will get big bonuses.