Thursday, December 09, 2004

PalmSource goes Linux -- shades of Apple's Pink

PalmInfocenter.com: An open letter to the Linux community from PalmSource
Today we at PalmSource announced we're going to extend Palm OS® to run on top of Linux. We've written this letter to explain what we're doing and not doing, why we're doing it, and how we're doing it. We'll also answer some likely questions.

... We intend to offer future versions of Palm OS Cobalt as a software layer on top of Linux (specifically, on the Linux kernel plus selected Linux services appropriate to mobile devices). The Palm OS software layer will include our well-known UI as well as a set of middleware and applications that encompass the best of Palm OS. We intend that properly written Palm OS 68k applications will run unchanged on Palm OS for Linux, and that Palm OS® Cobalt native applications using the Palm OS Protein APIs will port with a simple recompile. In addition, Palm OS for Linux will be able to run many third party Linux applications and services (GUI applications will need to use the Palm OS APIs).

... We're not open sourcing Palm OS®; we're going to implement it as a software layer that runs on top of Linux. Our business model will be licensing that layer, with hardware companies that use the layer in a device paying us royalties. We don't charge developers a license fee to create software that is compatible with Palm OS. Our development tools are also free; they are built on Eclipse, and we are a member of the Eclipse Foundation.

While we're not open sourcing all of Palm OS, we do expect to open source some of our code, and will actively seek to invest in the open source community through code contributions and other means.

... We think the Linux platform will become a leading operating system for mobile devices [jf. note they didn't say PalmOS mobile devices only], and we believe the endorsement and support of PalmSource for that platform will greatly accelerate that process. We think the combination of Palm OS and Linux can attract more mobile licensees and developers, create more new devices, and bring in more users than either could on its own.

... The Palm OS layer written for use on Linux will be designed to be portable to any suitable mobile Linux distribution, and we'll expose Linux APIs under the Palm OS layer.

... Together, we'll have the technological and market critical mass to challenge -- and, we believe, beat -- even the biggest proprietary operating system companies in the mobile market. [jf. of course they're still proprietary, but what the heck -- they're not Microsoft]

... We are acquiring China MobileSoft, a leading Chinese mobile phone software company. CMS has been developing a version of Linux with optimizations designed for smart mobile devices, especially around battery management and fast boot time. We will be using that technology as the foundation of Palm OS for Linux (although we will also support other Linux distributions).

The original PalmOS was based on an embedded device OS. Something out of Canada if memory serves. I'd love to know what when wrong with Palm's OS development process.

Apple spent years on OS development during the Classic era. They tried Pink (hence the title of this posting), Taligent (with IBM), the OS-like OpenDoc project (I really liked that one), and the beloved NewtonOS. All more or less failed. Finally they bought NextStep with its Unix based OS. That became OS X.

Now Palm revisits those days. Indeed, I think they've watched Apple and decided Apple's mixed (some would say parasitic) open/closed source strategy is a good one. I thought at first Cobalt was dead, but if they're really following Apple's example PalmSource may keep enough of the Cobalt API to make it worth pursuing -- though I doubt there will be many apps released for Cobalt directly. Sounds like CMS has done a lot of grunt work -- Linux is not known for power management.

It's interesting that they're using Eclipse. They intend to capture developers by allowing them to produce PalmOS and even non-PalmOS (write directly to Linux API) apps that may be repositioned for other platforms.

Clearly PalmSource and PalmOne are going in different directions. PalmSource wants to put their software atop many devices that will come out of Korea, Taiwan, Japan and China -- countries where pen input is very workable (though Korea has a phonetic alphabet -- of which they are quite proud -- and is less pen dependent than China). On the other hand PalmOne has talked of using Microsoft's OS. I'm not impressed with PalmOne's hardware direction -- actually I'm incredibly unimpressed -- so this may not be bad. Maybe PalmOne just needs to go away.

I'm very curious as to what Apple might do with this new picture. Does it change their calculations at all?

Given that the Palm platform has most of its extremities in the grave, this isn't a bad move -- though it feels late in the game.

No comments: