This installment in LinuxDevices.com's “Executive Interview” series explores PalmSource's plans with respect to morphing Palm OS into a middleware and application stack that runs on top of Linux, with an aim of targeting mobile phones and… other wireless-enabled devices. While visiting the company's annual developers conference in San Jose this week, we sat down with PalmSource VP of Engineering Mike Kelley and Director of Product Marketing John Cook to discuss PalmSource's Linux strategy and plans. Here's what we learned . . .

---

Q1. When will we see PalmSource middleware that enables all the legacy Palm applications to run on Linux?

A1. Kelley — First half of 2006. Our intent is to support both [categories of] applications in slightly different ways. The Garnet applications which are on 68K run in Garnet on an emulation layer we call PACE. That same API will be available in Palm OS for Linux.

The native APIs for Cobalt are called Protein, and are being carried forward largely unchanged to Linux. Although, there are cases where they will change. We have endeavored to make a best-practices Linux platform out of Palm OS Cobalt. So, there have been a few cases where we've looked at APIs and decided to make some changes so that they capture the style and essence of Linux programming. But, our intent is it will be very straightforward for a developer who has an application running on Cobalt to recompile it and get it running on Palm OS for Linux.

Q2. Does PalmSource still intend to ship the Cobalt OS?

A2. Kelley — Yes, we definitely have licensees working with it. GSPDA publicly said they are going to ship a Cobalt phone. We can't discuss the confidential plans of our licensees, however.

Q3. If Cobalt exists, and its usable, and production-ready, why haven't we seen any PDAs based on it?

A3. Kelley — That's more of a business question. When we went to market with Cobalt, we were having discussions with customers who were interested in Linux. Interest in Cobalt, whether on our proprietary micro-kernel or on Linux, has risen since our Linux announcements. The model we have to move the OS forward carries a lot of the higher-level features of Cobalt forward onto the Linux platform very cleanly. So licensees have a very clear migration path.

Q4. You acquired several products along with China MobileSoft, which you now call PalmSource China. Are those products a relevant part of your roadmap, or was the acquisition merely a means to gain customers in China?

A4. Kelley — The acquisition was a means to gain customers in China. However, those products are a very active part of our product line. We expect to introduce technology for smartphones, as well as featurephones, that leverages very heavily off of the CMS technology.

Cook — They have a very robust business today, around their RTOS-based stuff. Around 70 customers. They've almost doubled [their customer base] since being acquired by PalmSource. The secondary thing is that there's a fair number of companies that are interested in taking that business to other parts of the world, including other parts of Asia, India, Europe and America as well.

Q5. Do you know anything about the Linux-based smartphone that's about to ship with PalmSource China's smartphone stack?

A5. Kelley — Yes, but I'm not the right person to brief you on that.

Q6: PalmSource has said it will focus on the Asian market, and mobile phones. When can we expect to see Linux-powered phones (or PDAs) in Asia that can run Palm applications? How about in the US?

A6. Cook — We have to be kind of careful about what we say about that. We got badly burned last year in discussions behind the scenes with licensees about their enthusiasm. In building phones, you also have the operator and carrier element, that has a big influence. We've been saying we'll have something out next year for the OS, but it doesn't mean there might not be some licensees that have early access, and might try to do something.

Kelley — We also have a very competitive set of licensees, and we can't afford to comment on their plans.

Q7: Which RTOSes does the CMS mobile middleware support?

A7. Kelley — Nucleus is probably the biggest one, that most of the development work has been done on. Some are Asia-specific. Just to be clear, the CMS mobile applications are not Palm-based. But they have a lot of customers that they've been doing business with. And our objective is to grow that, and leverage that. But that's a small product, and our aim is to have a large uptake of our big product. That was our first acquisition.

Q8: Will you primarily focus on smartphones, or featurephones?

A8. Cook — There are a lot of people who push the limits of what a mobile phone can do: like with the Treo smartphone. But featurephones are getting smarter too, and more features are going into them. We feel there's an opportunity to push a robust OS down further into the mainstream. At the same time, there's RTOS-based phones that could run out of steam. As you get into 3G networks, they want to be able to do more. Instead of a picture phone, you want a video phone. There's a lot more interest in the middle of the pie than ever.

Kelley — Operating systems exist to help you manage complex software. The 3G featureset being driven by operators has absolutely pushed the software on the phones into the “complex” category. So, I think we've seen huge interest in using a “real” operating system such as Palm OS on Linux to address that space. That's by no means limited to the smartphone segment.

Q9: Can you comment on Microsoft's Windows Mobile focus being narrowly on the “smartphone grail,” while leaving the Windows CE featurephone to partners like Intrinsyc [story]?

A9. Kelley — Microsoft has always fallen back to “Windows Everywhere” as their basic approach. We've always believed devices need to be optimized in form-factor, software design, and differentiated to meet different needs. When you talk about phones, there's probably 20 verticals.

Q9: Is the move to Linux motivated by a desire to broaden PalmSource's customer base beyond PalmOne? If so, will you target other device categories, beyond PDAs and mobile phones?

A9. Cook — The short answer is, “yes.” It's no secret that we're out talking with a lot of handset vendors and consumer electronics companies all the time. But more and more doors have been re-opened by our move to Linux. Linux has momentum in the mobile device space. Linux is interesting to companies both as a preferred OS and an alternative OS. From that standpoint, this move has been well-perceived.

Q10: What is PalmSource's grand strategic vision? Is it handhelds and software platforms? Is it embedded, in the general sense? Is it predominantly the mobile phone space? What sorts of products are you targeting — medical instruments? telematics systems? Your software's very UI oriented.

A10. Kelley — We have a very strong focus on phones, because it's a huge market, we have great traction there, and it's a very innovative market. It's not, however, an exclusive focus. There's no question there are other opportunities for us. Our general vision is mobile devices where we can add value will tend to be wireless-enabled, in the future. I hate the “convergence” word, but we can talk about pervasive wireless. Even though there's a wide range of diverse products, there's a basic need to be in communications while mobile, whether to talk with people, or communicate, or get new games, or get a new song. That's a very powerful driver.

Q11. What benefits does Linux offer, as a PDA/mobilephone OS, compared to Palm OS? How about challenges?

A11. Kelley — Linux has a lot of momentum. I was at SGI when we kicked up our Linux migration there. The pace at which Linux penetrated the server market was really amazing, and caught people off-guard. My personal opinion is that the pace of Linux penetration in embedded is going to blow that away. It's a very good solution for embedded. There's tremendous opportunity to standardize the basic parts of the operating system that don't need to be a value-add. The basic Linux services are reliable, secure, pervasive — everything that's necessary to power the basic OS.

To build a phone on that is still a complicated project. Which is good, because that's what people pay us to do. That's a huge benefit.

There's also the strategic aspect. Many are attracted to Linux as an alternative to other major players in this space that they view to be maybe not good partners in the long term. Not naming any names.

We're very committed to both benefiting from that and contributing to it. We look forward to riding the wave.

Q12: Contribution-wise, you've already joined CELF [story]. What are you currently doing there?

A12: Kelley — We're a new member, but we're looking to see where we can help.

Q13: Dr. Nagel, in his keynote, said Palm has expertise in the mobile field that no one else can match. What did he mean by that, and what will PalmSource contribute?

A13. Kelley — Two examples. It's a challenge to build a really compact and complete solution. Linux provides only a piece of the solution. There are other vendors targeting the phone space on Linux. But no one else has our credentials for having taken actual complete hardware products all the way through, completely qualifying them, and taking them to market.

For example, power management is really challenging. There are some great open source frameworks for power management, but to actually build it out and productize it for any given architecture is very involved. So that's an area where we expect to see a lot of interest.

These are not areas we view as sustainable differentiators, though. Our value-add is building great frameworks, great user experiences, great application suites, and technology that is very flexible so that licensees are free to go and further differentiate their products.

There's a lot of value in what we bring right down to the metal in Linux today, because it is really difficult to finish a fully productized embedded Linux product. It isn't at the point where you just go download it, drop it on your chipset, and say “Wow! it worked!”

In phones, one thing the Linux community is still learning to deal with is proprietary intellectual property. Phones have a lot of technology that is highly proprietary, such as wireless stacks, media, and codecs, etc. One of the areas we help our licensees is finding a way to protect their intellectual property rights. A large contingent in Linux would love it all to be free, but most recognize that there are areas of intellectual property that are not going to be moving into the public domain any time soon.

Q14: Will PalmSource partner with other commercial embedded Linux and software stack providers such as MontaVista, Wind River, and Trolltech?

A14. Kelley — There are opportunities for collaboration with all of them. Past that, until we announce specifics, I shouldn't speculate.

Q15: When Trolltech had established their software stack for phones, they open-sourced their software stack for PDAs. Do you think PalmSource might do something similar?

A15. Kelley — I wouldn't want to promise that. But I wouldn't rule it out.

Q16: What benefits will a combined Linux/PalmSource solution offer over competing stacks from Symbian, Microsoft, and others? How about challenges?

A16. Kelley — Versus Symbian and Microsoft, one big advantage is, “It runs Linux.” I believe that is actually a tangible advantage for licensees.

Our product has always been very lean, mean, and efficient, and that is an area of ongoing differentiation. Also, our user interface is often cited as the best there is.

Q17. How is the integration of CMS going? What are CMS engineers currently tasked with? How about PalmSource engineers?

A17. Kelley — It's going very well. We're very pleased with the pace of integration. We have three geographic divisions now. There's Sunnyvale. There's Mt. Pelier in France, which you may not be aware of. And now we have Nanjing (China). All three of those are very involved in every project, and we have almost perfect geographic distribution.

Q18. In general, how has the reaction been to PalmSource's announcement of plans to embrace embedded Linux as an underlying software platform?

A18. Cook — We didn't have a big announcement this year [at our devcon]. We didn't have a big OS release, like last year. But we ended up getting twice as many attendees as we planned for.

Kelley: It's exceeded my wildest expectations.

LinuxDevices.com: Thanks very much, and good luck!

---

About the interviewees

Mike Kelley is vice president of engineering for R&D responsible for software development of the Palm OS platform, including telephony, wireless, multimedia, and applications. Kelley began his career in 1983 developing hardware and software products for high-end 3D graphics production, managing teams at some of the early leaders in the industry, including Vertigo, Inc. In 1989, he joined Apple's Advanced Technology Group, and was subsequently promoted to director of the Interactive Media Group, shipping numerous award-winning technologies, including QuickTime VR, QuickDraw 3D, and the first 3D hardware accelerator for Mac. In 1996, Kelley joined SGI as senior director for Cosmo Software, shipping breakthrough Web authoring tools such as CosmoWorlds and advancing the VRML Internet standard. He then became senior director for the Advanced Graphics Division, developing SGI's high-end graphics supercomputers and strategic technologies such as OpenGL. From 1999 to 2003, Kelley was vice president of engineering at ePeople, experiencing the dot-com wave while developing award-winning Web based enterprise software for collaboration, CRM, and knowledge management. His individual achievements include more than a dozen patents and numerous technical publications. Kelley studied Mathematical Physics and Philosophy at Simon Fraser University, in British Columbia, Canada.

John has more than 20 years of high-technology experience in consulting and senior level corporate positions. Most recently, he was vice president of product management and marketing for Palm, Inc. Before that, he spent almost 10 years with Apple, where he served in product marketing and business development roles for the company's hardware, software and multimedia divisions. John holds several patents for technology used in mobile and wireless devices. He holds a bachelor's degree from Auburn University.

---

 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.