Credible hints that Palm Inc is planning a new generation of Linux devices

[cervezas]

Palm has been silent about last months announcement of the ACCESS Linux Platform (ALP). But maybe that's just because the earliest time they'd be likely to release a Treo that runs on ALP would be a couple years from now. Palm isn't one to pre-announce products or roadmaps that far in the future. With PalmSource saying that ALP needs another nine months of work before it's ready to release an SDK, Palm might be biding it's time before going public with its judgment about the heir apparent to the Palm OS throne.

What we can say with some confidence is that Palm is quietly working with Linux behind the scenes. The hints that you'll have some kind of Linux running in the Palm of your hand are stronger and more credible than the hints we got a year and a half ago that a Windows Treo would be coming down the pike. For one thing, the anonymous sources that first informed CNET that Palm was exploring Windows Mobile (correctly as we now know) also reported that similar research was being undertaken into Linux. That leaked out in November of 2004, a month before we learned that PalmSource was acquiring China MobileSoft and charting a Linux future for the Palm OS. Was Palm getting ready for PalmSource's "Palm OS for Linux," or were they looking for other "partnerships" (CNET's word) to build their own Linux alternative? Difficult to say, and after all, these "sources" only were talking about research, not actual products.

More telling are the job openings at Palm over the last 7 months (as far back as I've been watching). In September of last year Palm had openings for 17 Linux engineers on their web site. Today there are 21 positions (16 in engineering) that specifically mention Linux experience as a qualification. Most don't have Linux as prominently in the job title as they did in September, but one tantalizing job description reads in part:

Linux Engineer- Mobile Handset, Embedded Systems

Overview:

As a senior software engineer, you will play a key role in the architecture, design and implementation of enabling technologies for a new generation of Palm devices. Working as part of the overall system team, you will work on the underlying support for various software applications.

Duties/Responsibilities:

You will be responsible for the design and development of components of a new software platform.

"New generation"... "overall system team"... "new software platform"... it does sound unmistakably like Palm is working on their own Linux platform. The job title uses the word "handset", implying that this platform might be intended for a future Treo line, but this conclusion might be undercut by a qualification later in the posting that specifies only "handheld" device experience being a plus (implying that telephony might not be part of the plan for this platform after all). I overanalyze, but you start to get the picture.

There are at least four possible interpretations I could make of this:

1. Palm is preparing a full-blown Linux-based version of Palm OS that they can continue to give the Palm name to and have complete control over.
2. This is part of the still-secret "Third Business" that Jeff Hawkins says Palm is cooking up. We don't know much about what this new category of device might be, but it might be a different enough animal that Palm would not consider using Palm OS or ALP. No news about this new product line, but for what it's worth, it got mention in the news again yesterday.
3. Palm is working in tandem with ACCESS and PalmSource to deliver their own user experience and applications on top of the evolving ALP platform, much as they did for previous versions of Palm OS. I'm doubtful that ACCESS would let a licensee--even Palm--be part of the actual design and building of the platform itself, but perhaps Palm might be given permission to replace the Garnet emulator in ALP with an emulator that runs an updated Palm OS.
4. Recognizing that the aging Palm OS Garnet is not going to satisfy their customers for two years until an ALP Treo can be released, they are developing a stop-gap system to buy themselves time. A simple solution: develop something like ALP but without all the extra APIs: just the Palm OS emulator and a few critical native Linux apps like the browser, email, phone, and messaging clients that could run in their own processes to improve the multitasking capability of the platform.

I discuss the last option in a little more detail here . All four possibilities are probable, but I consider the last one to be the most likely interpretation of what Palm is doing playing around with Linux in their basement. It's conservative and relatively easy but could add a lot of value to the Palm OS platform and (most critically) enable Palm OS to comply with the UMTS standard that GSM operators are using for their 3G networks. At the moment Garnet cannot do this, which is why the 700p is only slated for Sprint's EV-DO network at this time and we don't have hints about versions for Cingular or T-Mobile.

[TadW]

Thanks for your article, David! It's very good, although I didn't think of Palm seriously approaching Linux before. If I listen to yesterday's conference call, I had the weird feeling that Palm is more than happy with its decision to get in bed with Windows Mobile. They talk of corporate customers as their big chance to win new market shares. And obviously, this goes along Windows Mobile in conjunction with Microsoft Exchange Server to provide similar services like a BlackBerry does.

The question is, are there proven solutions based on Linux to satisfy the demands of corporate customers?

[cervezas]

The question is, are there proven solutions based on Linux to satisfy the demands of corporate customers?

I think Palm is focused on corporate customers because they see this as relatively low-lying fruit that could be plucked by integrating Windows Mobile (with some nice tweaks) into their already wildly popular business smartphone. But I doubt they really believe they'll be able to grow the way they need to by just expanding in the corporate market, which is small compared to the consumer market. I think like most smartphone vendors out there, Palm is trying to crack the mystery of how you get consumers to graduate from feature phones that already do quite a bit to smartphones that offer real mobile computing. Their motto is "the future of computing is mobile computing" but they know they've got to convince the average Joe, not just corporate IT, that this is the case.

So does Palm think they are going to be able to make breakthrough consumer smartphones without the kind of control over their platform that they've had historically--in other words, within the constraints of Microsoft's Windows everywhere vision? My guess is that they don't. They may not even feel comfortable being constrained by ACCESS's vision, although that's starting to look like a pretty exciting and consumer oriented vision. If I'm right, then the reason they are looking at Linux is not to further penetrate the enterprise but to pursue the bigger goal: getting consumers to catch the mobile computing idea.

It's not that Linux is consumer-oriented in and of itself--it's just that it's got a modern kernel and services that they can use to modernize the system that has done so well in the consumer handheld market: Palm OS. Remember: before ALP was announced Palm had expressed enthusiasm for PalmSource's Palm OS for Linux plans and stated that they would migrate to that from Garnet once it was ready.

[doctorow]

Originally Posted by cervezas

But I doubt they really believe they'll be able to grow the way they need to by just expanding in the corporate market, which is small compared to the consumer market.

I disagree. RIM serves only corporate clients, and they are more than seven times as big as Palm. So there is still plenty of room for Palm to steal BlackBerry market shares.

[cervezas]

Sure, and I think the Windows Mobile Treo line allows Palm to make a good run at RIM. But as large as RIM is, their shipments are practically an order of magnitude less than Nokia's S60 Symbian phones. Nokia sells 7-8 million of those semi-smart phones every quarter. Palm can't afford to ignore that much larger market and I'm doubtful that they think their partnership with Microsoft is going to be a lot of help for them in penetrating it.