Comments on: PalmSource and MontaVista Team on Next Gen Linux Phones

PalmSource and MontaVista Software, a leading Linux provider for the telecommunications market, today announced they are teaming up to help further accelerate the development of next generation Linux-based mobile phones. PalmSource has joined MontaVista's Mobilinux Open Framework partner program; and MontaVista has joined PalmSource's Palm Powered Mobile World (PPMW) program. As partners, MontaVista and PalmSource can leverage their collective expertise to create integrated solutions for handset vendors and mobile operators looking to build Linux handsets.
Return to Story - Permalink

Article Comments

 (12 comments)

The following comments are owned by whoever posted them. PalmInfocenter is not responsible for them in any way.
Please Login or register here to add your comments.

Start a new Comment Down

AOL Phones?

mikecane @ 8/9/2005 2:14:34 PM # Q
MEX: Why is AOL buying Wildseed?
http://www.pmn.co.uk/20050808aol.shtml
>>>However, in developing this solution, Wildseed also created a full Linux OS optimised for mobile handsets and sophisticated handset customisation platform. It is these key assets which AOL is interested in. Armed with this technology, its own extensive media libraries, the AOL mobile instant messaging franchise and Tegic’s existing customer base in the mobile industry, AOL has all the elements of a rich mobile user experience.

Is a $20 cell phone on the horizon?
http://news.com.com/Is+a+20+cell+phone+on+the+horizon/2100-1039_3-5823239.html

Reply to this comment

that won't save'em

auto194419 @ 8/9/2005 4:07:03 PM # Q
as if BeOS wasn't enough. as if those poor chinese bastards weren't enough. now they partner with someone else! i can't beleive it!
RE: that won't save'em
LiveFaith @ 8/9/2005 4:25:26 PM # Q
Yeah, I guess that proves it. If they made no partnerships with other industry players, then I reckon they would be doing well? Strange logic there. :-/

Pat Horne; www.churchoflivingfaith.com
RE: that won't save'em
cervezas @ 8/9/2005 5:55:37 PM # Q
Linux is all about partnering and collaborating. You can't be involved in the open source community without becoming part of the ecosystem of partnerships. It's the movement from the "cathedral" mode of development to the "bazaar" if you're familiar with Eric Raymond. And the whole point of putting Palm OS on top of a Linux kernel instead of a proprietary one is so that PalmSource can leave a lot of the kernel and system support development to partners and focus on the part of the OS that is really Palm OS.

The Palm OS middleware is still proprietary and commercial, but PalmSource has a lot to gain from companies like MontaVista that deliver a Linux kernel with all the system support software for particular chipsets and hardware. It's partly because PalmSource was trying to do it all themselves that they got into trouble with getting Cobalt out the door. Better to have partners like MontaVista helping with stuff like this than telling licensees that they have to roll their own version of the kernel, services, drivers, etc.

MontaVista for their part seem pretty enthusiastic as well. They gotta see the prospect of having Palm OS with its 30,000 apps and loyal following running with MontaVista Linux under the hood to be a pretty nice picture.

It's all good.

David Beers
Pikesoft Mobile Computing
www.pikesoft.com

RE: that won't save'em
LiveFaith @ 8/9/2005 6:36:39 PM # Q
Ya know, I've often wondered where PSRC would be today had it chosen to take this course instead of the original OS6 route?

Does anyone remember the comment form one of the execs about looking at all the other platforms (incl. Linux) and determining that OS6 was the best way? Hindsight = 20:20, and everyone knows they blew it with Cobalt. But imagine what PSRC could be today with this OS/middleware already shipping on PDAs, phones and who knows what else in the funnel.

It really can be the best of > 1 worlds here. I really hope PSRC makes it to the wedding feast before their oil runs out. It's just a tougher ballgame two year$ later.

Pat Horne; www.churchoflivingfaith.com

RE: that won't save'em
sr4 @ 8/9/2005 6:45:04 PM # Q
I'll be my usual cynical self, and wonder if Linux base will really make a difference. In cobalt for example most of the restrictions of garnet are carried over (strange multi-tasking, and one thread being able to crash all). Is the current kernel causing the problems, or the software architecture surrounding it?

Surur

RE: that won't save'em
Gekko @ 8/9/2005 7:01:02 PM # Q
>Ya know, I've often wondered where PSRC would be today had it chosen to take this course instead of the original OS6 route?

Same place. Except rather than an unwanted, stillborn BuggyCobalt, we'd have an unwanted, stillborn BuggyPalmLinux.

Chef Nagel and team would cook up a mess no matter what the recipe or ingredients.

So let's not plan any more of this "coulda woulda shoulda" and "ifs and buts".

Thank you.



RE: that won't save'em
cervezas @ 8/9/2005 7:34:05 PM # Q
Surer wrote:
I'll be my usual cynical self, and wonder if Linux base will really make a difference. In cobalt for example most of the restrictions of garnet are carried over (strange multi-tasking, and one thread being able to crash all). Is the current kernel causing the problems, or the software architecture surrounding it?

First of all, that's not exactly true about the multi-tasking in Cobalt. It's true that stuff running in background threads has to share a process with the other background threads. Everything in PACE has to run in a single process, too, of course. It's like a virtual machine. But any running application that's updated to the Protein API gets to run in its own process and protected memory just like in any other modern OS. All that is a big improvement over Garnet. Even the sharing of a process for the background threads (IMO, not a big deal) has a payoff in terms of better performance relative to the way WM does it. The reason for this is that being able to put an upward limit on the number of processes spawned allows the ARM processor to do some very fast context switching. That means multi-tasking has much less of a hit on performance.

At any rate, all of this is about the kernel, not the Palm OS framework built on top. From what I understand, ARM Linux is more like WM in the way it deals with multi-tasking. That's good and bad. Good because everything gets its own process (more robust). Bad because when you do multi-tasking it will have more of a performance hit than it would with the Cobalt kernel.


David Beers
Pikesoft Mobile Computing
www.pikesoft.com

Give the Be-derived development team the HOOK, already!
The_Voice_of_Reason @ 8/9/2005 9:40:37 PM # Q
I'll be my usual cynical self, and wonder if Linux base will really make a difference. In cobalt for example most of the restrictions of garnet are carried over (strange multi-tasking, and one thread being able to crash all). Is the current kernel causing the problems, or the software architecture surrounding it?

The architecture Palm chose (i.e. not true multitasking) with their unique kernel was the problem - not the bulk of the framework that sits on top. IF PalmLinux makes it to market it would have REAL multitasking similar to Windows and most other modern OSes. It's about time. The problem is PalmSource has VERY few people capable of helping fuse a Linux kernel to whatever framework PalmSource has been able to salvage from Cobalt (PalmOS 6). As I've said before, PalmSource needs to OUTSOURCE as much as possible to people who actually know what they're doing. They don't have time for more arrogant Be-flavored "we're gonna roll our own OS all by ourselves" crap. In fact, even with outsourcing, I think PalmLinux is DOOMED to repeat Cobalt's DOA history.




------------------------
Sony CLIE UX100: 128 MB real RAM, OLED screen. All the PDA anyone really ever wanted.
------------------------

The Palm Economy = Communism™
The Great Palm Swindle: http://www.palminfocenter.com/comment_view.asp?ID=7864#108038

RE: that won't save'em
PenguinPowered @ 8/9/2005 11:51:52 PM # Q
Hmm.. very interesting

No, not the press release. That's fluff. What's interesting is what people are reading into it. I don't see anything in that article that says that PalmSource is going to use MonteVista's Linux kernel, or that MontaVista has signed a deal with any of PalmSource's licensees. We'll just have to see what this deal really means. I don't think that "partner" means what you might think it does.

Meanwhile, "arm Linux" is a Linux kernel. You get exactly what you get on a desktop. Including Linux process management and kernel threads, meaning that applications people can partitition their aps into processes that are multithreaded.

Oh, and you can get it from the open source community for no fee, if you'd like to check it out. The omap patches at muru.com added to a 2.6.11 or 2.6.12 kernel result in a very stable system on a TI Perseus 2 development system, for example. I understand that the folks over at handhelds.org are doing interesting things around the Intel arm processors.

The interesting question, ala Unix phones will be trolltech versus maemo versus PalmOS. Too soon to tell how that's going to pan out.


RE: that won't save'em
cervezas @ 8/10/2005 8:29:08 AM # Q
I don't see anything in that article that says that PalmSource is going to use MonteVista's Linux kernel, or that MontaVista has signed a deal with any of PalmSource's licensees.

Well, you're right, it doesn't say that. But it's been clear since the DevCon that PalmSource wants Palm OS for Linux to be kernel-agnostic so their licensees can choose to work with a vendor like MontaVista if they don't want to roll their own. A lot of phone vendors are deciding not to roll their own these days since ARM Linux isn't something you can just plop down on any arbitrary collection of hardware elements and have it "just work."

There are other aspects of this partnership like working together on tools and coming up with reference handsets to make it easier for vendors to get phones to market fast.

I get the impression from PalmSource that they haven't decided whether they will do their own Linux distro or not. Obviously they need something that they'll use for development purposes in-house, but whether they'll package it all up with the middleware and system support software and call it a PalmSource Linux distro isn't clear to me and may not be clear to them yet. If I had to guess from what I've heard most recently, they're leaning toward "no" which means they'll be more like a TrollTech than a MontaVista in the market.

David Beers
Pikesoft Mobile Computing
www.pikesoft.com

RE: that won't save'em
PenguinPowered @ 8/10/2005 4:19:20 PM # Q
Now that's a reading that makes more sense: cooperate with MontaVista so that licensees who don't want to use PalmSource's reference kernel to roll their own can use MontaVista's.

My aging memory tells me that at DevCon PalmSource did say that they would have a reference distribution. It's something licensees would want, so I'd be surprised if they don't.


Reply to this comment
Start a New Comment Thread Top

Account

Register Register | Login Log in
user:
pass: