How we'll move to Linux - PalmSource CEO
And is the PDA dead?
By John Blau, IDG News Service | Published: 08:00, 15 February 2005
As competition intensifies in the market for smart phone OSes (operating systems), PalmSource is looking at ways to carve out a piece of market share. Earlier this month, the company acquired China MobileSoft (CMS) in a move aimed not only at gaining additional smart phone expertise but also Linux know-how.
At the 3GSM World Congress in Cannes, PalmSource unveiled four new applications that draw on its recent CMS acquisition.
IDG News Service interviewed David Nagel, president and chief executive officer of PalmSource, shortly before the start of the wireless conference and exhibition, which runs through Thursday.
You appear to be making some significant shifts in your strategy, moving beyond PDAs to smart phones and even Linux. Why?
Two years ago, we began to see the real growth opportunities in the handheld-computing market were shifting from stand-alone devices, like the original Palm Pilot, to devices that combine features of the handheld computer with communicator devices, whether it's a smart phone [Treo 650 review] or a BlackBerry-type device. We began refocusing our development efforts to serve that market, which we saw as the future.
Last summer, we brought out a product called Cobalt, which was our smart phone operating system. We have a number of products in development based on Cobalt.
Before I talk about Linux let me talk about China first. At about the same time we began our smart phone shift, we began to look at China, which we view as a market with enormous growth potential, particularly in the smart phone or communicator segment. There are 320 million subscribers in China today. It's the largest mobile market in the world, adding roughly 5 million new subscribers every month.
We also realised that to succeed in China, we will need to have more Chinese development and customer support presence. We considered either building our own development group or acquiring a company. Initially, we thought we would build a development group. But we soon realised that this would be difficult and would take time. So we looked around for a company that already had a capability there. That search led us to China MobileSoft. We closed that acquisition two weeks ago.
Why CMS? Did it have something PalmSource lacked, like Linux expertise?
CMS has three kinds of products. The first are wireless applications, like multimedia messaging, WAP browsers, e-mail and games. These applications are typically shipped with smart phones, which combine the capability of a handheld computer with a phone. They are also in feature phones. These phones are positioned in the middle between high-end smart phones and low-end basic phones, which offer telephony and messaging. Feature phones are like smart phones with all kinds of applications but they are closed devices. You buy them as an appliance. They come with preinstalled applications, which you can't load, and they have a proprietary operating system.
The second type of product in the CMS portfolio is a feature phone operating system. The feature phone market today is actually much larger than the smart phone market. Of the some 600 million mobile phones sold worldwide per year, several hundred million are feature phones. China is the largest market for feature phones.
The third product type is a smart phone platform, similar to Palm OS, but based on the Linux kernel and services. This was interesting to us. The more we looked at the work CMS had done with Linux and the direction we were taking with Cobalt, we saw a lot of benefits in using Linux as a foundation for our future offerings.
What kind of benefits?
First of all, when you build an open platform for mobile phones, you find yourself confronted with an extensive set of forward looking developments that you need to support, such as a whole variety of different microprocessors and radio modules - each requiring a significant amount of software to integrate with your basic platform. In the case of Linux, the work required to develop all these drivers is already done. Many chip makers do a reference port of their chips and create the software for integrating those chips to standard embedded Linux.
So with Linux, you're able to avoid a certain amount of development work?
Typically, the first release of a chip is just made available with Linux drivers. You may have to modify these drivers or improve them. But the basic work is done.
And manufacturers don't do this right away for the other operating systems, such as Windows Mobile, Symbian and even PalmSource?
Yes, that's correct. Most, if not all chip makers, now ship an initial implementation that runs on Linux. Implementations for the other operating systems follow. So again, by moving to a Linux platform, we avoid a lot of work. This translates into cost savings in development.
Does Linux also open doors in China?
Indeed. Linux has really become the government-supported platform of choice for the IT sector in China. Windows, of course, is also in China and has widespread use on desktops. But it's quite clear that the Chinese government has made a series of decisions to support the development of Linux to ensure at least one alternative to the Microsoft monopoly. If you want to do business in China, having an offering based on Linux is a much better start than having one based on some other operating system.
Are you primarily interested in using Linux in products targeted at the Chinese market or worldwide?
We first thought Linux would be very valuable in the Chinese market. But now, we believe it makes a lot of sense to use Linux in a broader context moving forward. We will be porting our Cobalt application level and all user interface programs to run on top of Linux rather than running on a micro-kernel that we originally made as part of Cobalt.
How much of your OS software will be Linux, how much PalmSource?
The micro-kernel part of a platform is maybe 15 percent of the total code base. But it's a very important 15 percent and the part that obviously determines the basic character of the product. We believe Linux is the operating system of the future for a wide range of devices, not only for handhelds but also embedded devices, appliances, consumer electronics.
You aren't the first to look at Linux in the mobile phone space? How will you differentiate?
There is only a small number of smart phones in the world today using versions of Linux. They haven't been particularly successful for a couple of reasons. The time required to boot up a Linux-based phone, for instance, is too long - as it is, by the way, with Symbian and Windows Mobile - between 20 and 45 seconds. This is unacceptable. People expect phones to come on in a couple seconds. So work is necessary to decrease the boot time. Another problem is power management. CMS has already done some development work to improve boot time and power management in battery-operated devices such as smart phones.
Will the shift to Linux put a squeeze on Cobalt and some of your other operating system development work?
We believe that by putting our software on top of Linux, we are providing a powerful combination. We offer tens of thousands of applications already created for Palm OS. That's a real benefit. People typically buy handheld devices because of their applications. With our huge number of applications and recognized user interface, we believe we bring enormous value to the Linux community and, at the same time, are able to take advantage of Linux virtues to create a new platform for wireless devices.
When can we expect to see your first Linux-based product?
We haven't announced any availability. We just closed the deal with CMS. My engineers would shoot me if I gave them a date they hadn't fully absorbed.
What about security? Will the move to Linux help?
With Cobalt, about a third of our effort in developing that program was focused on delivering a really secure platform. Security is absolutely essential in wireless, whether wireless networking for computers or wireless networking for phones and devices. We want to avoid the horrible security problems that exist in the Microsoft PC world. Most of the software releases that Microsoft sends out today are security patches.
We believe smart phones should develop along a much more secure path from the beginning and view Linux as an inherently better platform than Windows from a security point of view. It's a newer platform. People have been able to take an advantage of what people have learned painfully in the Windows context.
What exactly are you doing to ensure security?
We support a capability for signed code. Imagine a piece of code that has a signature on it, basically saying that this is a trusted application. Unless the code has been signed by some legitimate signing agency, which often is the carrier or wireless operator, you can't load the application on the smart phone. We also provide a lot of the encryption capabilities as well as secure VPN technology.
Today's phones offer just about everything you can imagine - e-mail, music, photography and now even TV. What application is still missing?
Over time nothing. These little devices will be as powerful as today's desktop computers but, of course, with smaller screens. Ease of use is crucial - that's our OS hallmark. It doesn't do you any good to buy a smart phone that has 30 functions if you can't figure out how to use them. We enable these devices to be used by ordinary people without having an IT staff standing behind them to explain how to do something. Mobile devices can't behave like computers. They have to be much more refined.
Is the PDA dead?
It's not dead. We still see 3 to 4 million of these devices being sold with our operating system. What we're finding, though, is that the market is evolving. There are people, like my wife, who still use a PDA. Although she has a smart phone, she just recently bought a new PDA because she does digital photography and uses it as a photo album. Some people will want a combined device, others will want two. The market will sort this out.