[cervezas]

Originally Posted by doctorow

I wonder why it'd take almost another year to release the SDKs.

Well, there's a lot more to a polished OS than just getting a couple of simple applications to run in a demo. A complete OS must support hundreds of APIs that wouldn't necessarily be exercised in any of these demos. And while many of the concepts behind the MAX framework have been in development almost since PalmSource first started working on the Linux version of their OS, the announcement at 3GSM reveals a major change of course that I think was only finalized when the acquisition was completed in November. PalmSource has now decided to shed most of their proprietary system components in favor of open source ones.

For example, until recently they'd been using a graphical user interface engine that they acquired from Be called Picasso. Picasso has been around since development began on Palm OS Cobalt, but it's gone now--replaced by GTK+. Likewise the proprietary schema database API that was supposed to be brought over from Cobalt is gone and replaced by an open source database engine called SQLite (which I think will be a very welcome change, by the way). Also gone are most of the PalmSource developers who worked on these proprietary components. Remember Dianne Hackborn? She's at Google now. PalmSource is hiring like crazy as I write this.

I believe there are still big decisions about how ALP will work that at the time of the demo had not been hammered out completely. One I'm aware of is the question of whether ALP will run X Windows like the Nokia 770 does, or should they develop a lighter weight windowing toolkit that will perform better? X Windows would make it easier to port applications from other Linux platforms and would be more in line with the new thrust away from proprietary solutions, but at the moment it's pretty heavy and slow on a resource-constrained device. That's a pretty fundamental question.

Some of the other things that take time before you can release an SDK: making sure all the APIs you're creating are things you believe you can stick with well into the future, getting the developer tools ready, writing all the documentation, writing the MAX native applications that will serve not only as part of the ROM image delivered to customers, but as sample code for developers to use.

I know very few of the details, of course, but my guess is there's a lot left to do. Even though everyone is egging them on to hurry, my hope is that they really take their time and get it right. It's only because of tremendous forethought that the original Palm OS held up as well as it did over generations of new devices. They need to put the same kind of forethought into ALP if they want the platform to succeed the same way.