Quote:
Has the environment changed in such a way that even if some key people wanted to stay after a few weekends (after being enticed by us owners) that they would not even have the permission to do this work - not on company time?

No car-player work after the Sonicblue takeover was done on company time. Much as I like the spirit of Rob Schofield's suggestion, I can't see it being either politically or financially feasible. The only scenarios as I see them are:
  • The most likely outcome: v3alpha8 is the final, summit release of the car-player firmware. Don't forget that this is itself a good result, as there was widespread fear that the D&M takeover would halt all firmware development (at 2.00).
  • Next most likely: a new v3 alpha is released following spare-time work by Rio employees, like the previous ones -- but this time, as a bug-fix-only release from the v3alpha8 branch. This would have to happen at the right sort of time in Rio's commercial development cycle, which is not right now, and possibly not until the autumn (or winter). That would then almost certainly be the summit release.
  • Least likely outcome: a new player is written as open-source by "the community". This is appealing to me as a car-player user, as it would put car-player development beyond the whims of a commodity flash-player company and the developers it has jaded. In a sense, a lot of the hard bits are done: madplay, tremor, vfdlib, libdaap. But you still get to rewrite the UI, the database, the cache, the USB client code (hint: 2.6 has USB mass-storage target support built-in), and, unless you can live without them, the visuals. It's a lot of work for a tiny hardware population (at least two orders of magnitude smaller than RiscOS).

But what keeps at least this jaded developer hanging around in this commodity flash-player company is the thought that one day we might get to do our H4. That will only make sense to those of you who know about John Harrison's marine chronometers (or have read Longitude), but after having built the amazing pieces of engineering that were the car-player and Rio Central, and despite none of it having the success we'd hoped, there is still honour in building the same thing all over again but this time the size of a pocket-watch. Karma could've been the H4, but never quite measured up (not least because we didn't do the 40Gb one).

If Rio succeed in giving you a car-player the size of a big pocket-watch -- and in 2005 Rio just might -- then all of a sudden you might not feel you need your bulky old marine chronometers.

Peter