That's a great question about hardware freedom. Mike may not be able to talk too much about how a specific hardware vendor does it, but it would be interesting get more examples and perspective in addition to the discussion on screen sizes.
Didn't we see in the past that there is a balancing act. The flexibility that allowed some features to come out earlier in devices like the Clie line from Sony also left some issues for application support down the line. But, personally, I'm really glad because it also propelled the whole Palm platform ahead light years!
I think the answer is for the hardware vendors to stay creative, but to provide additional information and support to developers for their particular devices to minimize the support issues. I really want features earlier, and I really don't want the devices to all look the same because of OS restrictions!
As far as Zodiac, didn't they put in some audio/video processing improvements special to Zodiac to support gaming? I doubt that they really changed much in the supplied OS from PalmSource (but that would be an interesting question also). More likely, I'd guess that they built new supplementary APIs for functionality and hardware support that just hadn't yet been released in whatever version of the OS the device was built on.
Another tradeoff, though, is probably that hardware vendors probably become very attached to their own customization and don't like to invest to revise it when it already works. (As we've seen, it's not so trivial to produce a bug-free device!)
I'm just speculating, but I wonder if that is one of the reasons Cobalt OS has not been adopted quicker. Not that there's a particular problem with Cobalt, but that PalmOne has a large investment in the tweaks and enhancements to Garnet specific to their devices, and so they may have been reluctant to abandon that investment to move forward to Cobalt right away when Garnet works just fine. PalmOne wisely wants to make the most of what they've done, both with regard to code and with regard to device form factors that work well.
One piece of supporting circumstantial evidence for that theory is that I talked to one group working on a Cobalt smartphone. They surprised me by telling me that Cobalt was actually
easier for them to build a smartphone on than Garnet because it already did a lot of the work for them that they would have had to add to Garnet!
But remember that there's still time to follow-up this interview with more questions to Michael Mace at
this thread at AllAboutPalm. I think it will end sometime next week, so don't wait too long! I hope you'll throw out some of these questions to him there.
