Well my biggest reason for worrying about Palm is that they seem to be faltering on the hardware side. The only company releasing significant numbers of Palm OS handhelds now is palmOne, and between the ho-hum Tungsten T5 and the buggy and needlessly memory-deficient Treo 650 I'm starting to seriously question whether or not palmOne can continue to succeed as a hardware vendor in the long run. The innovative days of the Palm V are long past, by the end of '05 they could easily fall to the #3 worldwide PDA manufacturer behind RIM and HP, and without any major new innovations the Treo can only do them so much good (especially with several promising Pocket PC equivalents right around the corner).
The ChinaMobileSoft acquisition was a good idea, they needed the portable Linux expertise, but I don't think it's going to help them much at penetrating the Chinese market - the big non-Chinese handset manufacturers are pretty much all in bed with Symbian or Microsoft at this point, and the domestic Chinese ones won't want to get stuck paying royalties to Palm when they can cheaply develop their own Linux derivatives. The SyncML switch is probably being forced on PalmSource by enterprise customers, but it's really a pretty minor change at this point, and after the Cobalt fiasco I don't imagine a lot of developers are going to be rushing out to re-write their conduits to support it.
I don't really see Microsoft as doing anything particularly smart here, they're just taking advantage of the fact that they're Microsoft; they have a standard, consistent API based on an OS that's been around for years and that will still be around in some form a decade from now, they have a vast network of hardware partners and ample experience at keeping them happy and releasing new products, they provide stability at a time when PalmSource can't.
PalmSource could still survive, but they're going to have to do a lot more than they've done so far - they need some big and innovative new hardware partners, they need to get Linux-based Cobalt shipping on devices in time for Christmas, they need palmOne to bring on a couple of decent industrial designers and a QA department that actually does its job, and they need to take some big steps to counter Microsoft's growing lead in digital media. And they really need to fire David Nagel and the other clowns who've gotten them into the mess they're currently in.
The OS will almost certainly stick around, though - it may have to be acquired by a bigger and better-managed company, but it's still the best-designed handheld OS out there and it's not going to be allowed to just disappear.
Don't get me wrong, I'm most definitely a Palm OS fan and would rather not see yet another software market taken over by Microsoft, I'm still rooting for Palm in spite of all of their bungling, but for business purposes it no longer makes sense for me to act like Palm OS is going to continue to be the preeminent mobile software platform for years to come.
