Disclaimer: This is not a critique or dissing of Nokia policies. It's a personal impression and personal take on the 'what would you have done differently'.

I had a bit of a catch 22 feeling with Fremantle development. As in, for your application (and, as a consequence, the developer ) to be 'famous' it had to work on an N8x0 in a demoable way so it could make it to extras and people in general. On the other hand, if your stuff really required Fremantle class hardware, it was not really demoable and thus the developer behind it did not have good chances of getting a device... Think about how Fremantle stars were selected - these were actually based on the work/popularity/feedback from N8x0 devices, so the results were a bit of a mixed bag.

As to what I would have done differently... I would have kept the Beagleboard port in much better shape. It *was* a commercially available, rough hardware equivalent, that didn't cost that much, so people could do their thing (and you could even hand out a few, it's cheaper and does not need an NDA). Didn't even Kate demo early Fremantle stuff on a Beagleboard ? I think abandoning that direction was a chance wasted for acquiring an early developer base.

That said, I'm more than happy committing some time to develop for the N900 *now* that I have something that does not include constant fighting with prerelease SDKs and guesswork how stuff would really look/feel like if there were no Xephyr/scratchbox/VMplayer overhead/influence