For Palm Inc., it’s time to turn that portrait of Palm’s creator, Jeff Hawkins, to the wall.

Palm split its Palm OS engineers into a separate internal company that will license the software not only to third parties, such as Sony Corp. and Handspring Inc., but to Palm’s own Solutions Group, which now is, essentially, a hardware manufacturer.

The Palm OS outfit will be headed by recently hired CEO David Nagel. One can only hope Nagel thrives on a challenge.

The split is probably long overdue. Palm got its early traction in the market with the Palm Pilot handheld, which was innovative in terms of both hardware and software. For quite a while, Palm was the only game in town.

Meanwhile, Microsoft Corp. was mocked and criticized for failing in a business it had pioneered: creating a dominant client operating system, this time for handheld computers. The company looked astonishingly unheroic. Successive new releases of Windows CE were sniffed at, if not derided. But Microsoft kept plodding ahead. By the time the PocketPC 2002 software was released in 2001, based on Windows CE, Microsoft was gaining support from hardware makers, applications vendors, value-added resellers and integrators focusing on specific vertical markets, and users.

Splitting off the Palm OS group can be seen as a bold act…or a desperate one. It clearly pits the two companies against each other as rival operating system vendors, and therein lies Palm’s challenge – and its opportunity. The Palm OS group will be battling for the hearts and minds of software developers, who will have to be convinced that the Palm software delivers something more than what they can get from Microsoft.

That’s the kind of battle Microsoft has waged since its beginning – and always successfully. The company has invested, and continues to invest, huge amounts in Windows CE and its kin, which now form a set of modular, adaptable operating systems for embedded systems and cellular handsets, as well as handheld computers. Palm can’t even begin to come close to those financial resources.

Palm is also hampered by the fact it took, and is taking, so long to actually shift its operating system from the Motorola Inc. DragonBall microprocessor to those based on the far more powerful and capable 32-bit ARM and StrongARM architecture. Handheld devices are becoming vastly more powerful, driven by the many of the same hardware dynamics that have driven the PC and laptop markets: more power, more RAM, less cost. Compared to Microsoft, Palm is late getting its software on that sweet curve. There’s a chance it might even be too late.

But Palm does have a lot of vital resources. The company still seems to treasure the original insight, bequeathed by Hawkins and Co., that smaller can be better, but that the really important thing is that smaller is different – different resources, different usability issues, different application and user interface designs. With the separate Palm OS focus, this “vision thing” is an asset that not only can hold loyalists but attract new ones, especially from the enterprise. But it’s an asset with limited shelf life.

Second, the OS group could, finally, actually do something with the Java platform for which Palm OS became a reference platform some two years ago. Java 2 Micro Edition, with its various device profiles, is being widely adopted by device makers around the world, because it allows them to turn these gadgets into programmable computers. As far as I can see, Palm has simply squandered that opportunity, while Microsoft has been assiduously cultivating its own developer community and getting attention, and a growing allegiance, to its .Net scheme for Web applications written to such standards as XML and SOAP.

With J2ME, and the growing activity around Java standards for Web services, Palm could position itself as a bold, even daring, innovator surfing the ever-changing wave of handheld development.

The third resource is wireless communications, where Palm has seemed like a confused follower, rather than a dynamic leader. Palm devices are mobile, but most of them are still appear to be used as extensions of a user’s Windows PC, tethered to it by the HotSync cradle. No one knows exactly how the “wireless market” will develop. But that’s the point. For operating system developers, the key is flexibility. What does the OS need to be able to exploit wireless connections? If that question is asked, it will lead to a continuing, vigorous stream of innovation tied to making Palm handhelds more effective for corporate end users.

Jeff Hawkins is now, with the other Palm alumni who joined him at Handspring, discovering what it’s like to be a hardware builder who’s staked his future on an operating system other than Microsoft’s. The Palm Solutions Group is about to discover the same thing.

Palm’s future lies in its operating system. The question Palm CEO David Nagel needs to answer is not “how do I maintain market share?” but “how can I re-invent Palm?”