Palm Making Behind-the-Scene Changes to the Foleo

[Ed Hardy]

Palm, Inc. has announced that it will be working with Wind River to improve the version of the Linux operating system that is at the heart of the Foleo Mobile Companion.

During the Foleo's development, it was based around the Linux 2.4 kernel. However, Palm developers made extensive changes to this make it more appropriate for this upcoming device.

Going forward, Palm and Wind River are going to work together to make these same modifications to the more recent Linux 2.6 kernel, so the Foleo can be based on that.

(view large image)

"Wind River's open Linux platform gives us the reliability, performance and stability we need to continue driving innovation," said Mark Bercow, senior vice president, Business Development, Palm, Inc.

Are You Supposed to Salute a Kernel?

Although a significant change, it is not the sort of thing most Foleo users will notice, as it only involves changes to the Linux kernel.

The kernel is the basic foundation of the operating system, and performs many essential tasks. But lying on top of it is the user interface (UI), which is what people interact with. Changes to the kernel don't have to have any effect on the user interface.

In addition, operating systems include Application Programming Interfaces (APIs) that developers use to make their applications capable of exercising both low-level hardware features like graphics and wireless connectivity and higher-level features like UI toolkits and databases.  

Like the UI, which users see and interact with, these developer APIs won't necessarily change when the Foleo kernel is updated. This means that applications written for the initial version of the Foleo shouldn't be affected by any changes going on behind the scenes.

Development Tools, Too

One group who will notice this change is developers. One of the major effects of the modifications to the Foleo's kernel will be support Wind River Workbench, an established, Eclipse-based integrated development environment (IDE) for embedded Linux development. Going forward, developers will be able to use this suite to create applications for the Foleo.

"The Wind River Workbench IDE is regarded as being one of the best tools for mobile Linux development in part because of excellent debugging capabilities on either a host PC (emulated hardware) or on the target hardware itself," said David Beers from Pikesoft Mobile Computing.

Wind River will also be providing customer support to help Palm and developers extend the Foleo's capabilities.

Related Articles

• Brighthand's First Thoughts Review of the Palm Foleo

Thanks to all the people who helped with this article.

[Hook]

Thanks for the info, Ed.

[Antoine Wright]

I don't know. Why before a item is publically available make such a release about something dang near elemental to the device working as something that has changed. If anything, it would have made more sense to make this announcement if there was already this device in the immediate pipline and several other devices that would make such a proclamiation of the developer tools a thing that should be mentioned. This is a item that I don't have a good feeling about.

[cervezas]

Originally Posted by AWright

Why before a item is publically available make such a release about something dang near elemental to the device working as something that has changed.

Keep in mind that the announcement was made at a Linux developer conference and that's who Palm was really speaking to. Foleo customers have little reason to care about this, but to developers the announcement means that the Foleo will have top-notch development tools that improve their productivity and enable things like debugging directly on the target hardware.

When Palm started working on Foleo, this Wind River kernel, services and toolchain didn't exist. Now that the mobile Linux ecosystem is starting to blossom it makes a lot of sense for Palm to work with a partner that manages all the kernel updates (Linux isn't like Windows--it changes all the time), develops and maintains professional developer tools, and provides support services to Palm and Foleo application developers. It's a very natural move, not a sign of some kind of internal problem as it was spun on PalmInfocenter. It's also a pretty nice feather in Wind River's cap.

-David Beers

[Ed Hardy]

Basically, what Palm is doing here is bringing in a company expert in Linux development to help it with the features of the Foleo that users won't see.

This should result in a product that more stable and easier for developers to create applications for, while not having any significant effect on the hardware or UI that Palm spent so much time working on.

It also shouldn't affect the software developers have already begun working on.

Update: cervezas beat me to it, so suffice to say I agree with him.

[Antoine Wright]

Thanks for the clarity. I still feel a bit uneasy, and at the same time, there is reason to. Seems like the mobile Linux path is a lot newer than much of the news and companies talking about it would have one to believe. Making the announcement does fit to a large extent, and if I am Wind River I am happy about it.

The question I have is the same one that I have with EVERY mobile Linux variant, are there steps being made to ensure compatablity across those variants in terms of UI expectations, APIs, and future/legacy support. While I think (as most would) that Linux is a great idea, the fragmentation gets me, and this announcement did nothing to really bolster that confidence. Lest WindRiver is a part of a great Linux consortium that has the ear of most of the Linux community and this announcement then does have a better frame to fit it.

[minimalposter]

Originally Posted by Ed Hardy

Update: cervezas beat me to it, so suffice to say I agree with him.

Its good to see you here cervezas/David.

[Josser]

Originally Posted by AWright

Thanks for the clarity. I still feel a bit uneasy, and at the same time, there is reason to. Seems like the mobile Linux path is a lot newer than much of the news and companies talking about it would have one to believe. Making the announcement does fit to a large extent, and if I am Wind River I am happy about it.

The question I have is the same one that I have with EVERY mobile Linux variant, are there steps being made to ensure compatablity across those variants in terms of UI expectations, APIs, and future/legacy support. While I think (as most would) that Linux is a great idea, the fragmentation gets me, and this announcement did nothing to really bolster that confidence. Lest WindRiver is a part of a great Linux consortium that has the ear of most of the Linux community and this announcement then does have a better frame to fit it.

Are Linux variants 100% compatible? i.e. like apps for WinMob5 are compatible with WinMob6

[Gazpacho]

Colligan and Hawkins have always felt very strongly about Palm developing its own OS in-house, so it can control the destiny. While bringing Colligan and Hawkins back into Palm when merging Handspring into Palmone, I'm not quite sure whether they were happy that Palmone lost control over the OS, as that was more Nagel/Benhamou/Bradley's idea. And we all know Benhamou and 3Com's involvement was the primary reason anyway for Colligan and Hawkins to leave Palm, the company they'd created in the first place, at the end of the '90s.

Having transformed the new Palm into a Handspring II today (Palm's only viable product line is Handspring's Treo), and having removed the last 3Com bastion (the same Benhamou) from the board of directors, Colligan and Hawkins also regained control over PalmOS before moving on to an in-house Linux OS for the next-gen Treo's.

Apparently, the OS-requirements for the Foleo and the Treo are so different, that they outsourced a lot of work for the FoleoOS to Wind River Systems, while still keeping everything in-house for the new Palm/LinuxOS for future Treo's. I really wonder how different these two pieces of technology are.

[cervezas]

Originally Posted by AWright

The question I have is the same one that I have with EVERY mobile Linux variant, are there steps being made to ensure compatablity across those variants in terms of UI expectations, APIs, and future/legacy support. While I think (as most would) that Linux is a great idea, the fragmentation gets me, and this announcement did nothing to really bolster that confidence. Lest WindRiver is a part of a great Linux consortium that has the ear of most of the Linux community and this announcement then does have a better frame to fit it.

I dislike fragmentation of the Linux desktop, and worry about it on handsets, but for the Foleo Palm is trying to do something that is so different from any Linux desktop or mobile operating system I know of--and those differences are so interesting--I really don't care if you can't install Gimp with standard DEB or RPM packaging. I'm much more excited about seeing the realization of a new vision for the PC (the first, really, since the Mac) than I am worried about the extra effort it might take to port an app from another Linux platform.

Ironically, there is a downside to releasing a platform that has easy compatibility with a ton of free software. A lot of that software is only going to be so-so when ported to a special platform like this (it will probably be optimized for 2GHz x86 processors, not 312MHz ARM processors, for example) which means that user's first attempts at installing 3rd party software are likely to be so disappointing as to cause them to conclude that "Linux sucks." I know also that there is a common sentiment among many accomplished Palm OS developers that if Foleo ships with a common Gnome Toolkit UI (GTK+) they will be reluctant to invest the time and money required to create applications for it. Why? Because of a fear that they'll be competing with the large community of Linux hackers who give their software away since they don't need to make a living at it. There's a perverse incentive problem there that makes free software kind of a double-edged sword. So, in terms of software quality, I actually think Foleo might be a better product if the APIs and UI toolkit don't adhere *too* closely to standard Linux, if there even is such a thing.

I'm sure that's probably a minority view (it's hard to argue against "free as in beer") but I'd like to see Foleo apps that have been designed from the start to be used on Foleo hardware with Foleo (not desktop) users in mind. I believe many, maybe most of those apps will need to be quite a bit different from common Linux desktop apps. I'd like to see some "Zen." Even if that means fragmentation and fewer apps at the beginning, I think this particular fragmentation will show off the unique qualities of the device best.