Access Linux Platform at LinuxWorld SF

David Beers was in attendance at LinuxWorld San Francisco where PalmSource was exhibiting the Access Linux Platform. PalmSource also held a full day of hands-on tutorials in which they built and debugged native applications for mobile Linux. In this article David brings us up to date on ALP, its progress and other details on and what was being presented about ALP at LinuxWorld. Read on for the full report.

Access Linux Platform at LinuxWorld SF

by: David Beers
August 20, 2006

PalmSource Booth at LinuxWorld SFPalmSource and parent company ACCESS were a major presence at LinuxWorld in San Francisco this past week. While other mobile Linux vendors like TrollTech and a la Mobile had announcements about upcoming reference handsets for developers or new Linux phone platforms, PalmSource was the only one to demonstrate their new operating system on real handsets. They were also the only one to give developers a full day of hands-on tutorials in which they built and debugged native applications for mobile Linux. The three PalmSource Developer Day sessions were well attended with standing room only at times, and the demos phones, reference board, and developer tools all showing the ACCESS Linux Platform (ALP) drew crowds on the expo hall floor of the conference.

ALP is still a work in progress. PalmSource reports that it is on schedule to meet its end of year deadline for delivery to licensees. But the difference between what was demonstrated this week and the ALP of six months ago is pretty dramatic. The crude demo launcher from the February 3GSM conference has been replaced with a tastefully themed launcher that presents not only Palm OS Garnet and native GTK applications, but also applications written against PalmSource's new MAX framework. MAX stands for "Mobile Application eXperience" and PalmSource has contributed new widgets to the Gnome Toolkit (GTK) to enable the kind experience that Palm OS users have come to expect--as well as some new ones focused on the smartphone user experience. Native MAX PIM applications that use these controls and other advances are now on display. HotSync has been rewritten to be compliant with OMA standards and Internet-compatible for synchronization and backups via WiFi or wide-area network. And, importantly, there is a complete suite of developer tools now. Based on the familiar Eclipse development environment and other open source tools that PalmSource has optimized, these performed without a hitch for around 50 developers during a three-hour cycle of GUI design, coding, and debugging. (More on the developer sessions in a later article.)

Unfortunately, while attendees were free to try out the Haier N60 flip-phones running ALP in the PalmSource pavilion, PalmSource was very strict in forbidding pictures of the handsets or reference board. The shots I was able to take are of applications running in the ALP simulator on the Linux laptop that PalmSource provided for use during the Developer Day sessions. These images show screens that are similar, but not exactly like what LinuxWorld attendees saw on the Haier handsets.

I was told that the Haier demo phones have specs that are typical of the kind of handset that PalmSource has in mind as a primary target device: QVGA (240 x 320) screen resolution, a standard phone keypad, and preferably a touchscreen. While ALP is designed to work well with non-touchscreen phones, Palm OS applications that lack full support for the five-way navigator will require touchscreens, as presumably will smartphones intended for Asian languages that are better supported by handwriting recognition than a keypad. The reference board that was on display had what appeared to be a VGA (480 x 640) screen and a full alphanumeric keyboard to show the range of hardware configurations that are possible.

Naturally, the first thing you see of ALP is the launcher. My reaction was that it looked a lot like Palm OS: a grid of icons that could be tapped with a stylus or finger, or highlighted and selected using the five-way navigator, much like on a Palm OS Treo. On the small Haier handsets there was room for nine icons at a time--the large-screen reference board and desktop simulator appeared to have room for as many as twenty. The most obvious difference from the Palm OS was the color: a bold blue graphical background image with white fonts and icons that were highlighted in a gold border when they received focus during navigation. The colors, backgrounds, borders, and controls in all the MAX applications running on the device were rendered in the same theme, creating a consistent look and feel throughout. In the developer sessions I learned that these themes are developed using scalable vector graphics so that they can scale naturally to different screen resolutions. PalmSource expects that these themes will be an area where the wireless carriers will create a branded experience, but it appears to be an opportunity for user customization as well. I wasn't able to ascertain whether "skinning" your phone with third party themes will be an option, but it seems likely.

ALP Ghost - Palm OS emulatorNo matter which of the four supported APIs an application is written against in ALP (Palm OS Garnet, Java, GTK+ or MAX) their icons are displayed in the same screen, so initially it is transparent to the user how a particular application was developed. But when launching a Garnet application you see the Palm logo for a second or two before the application launches: it's the same splashscreen you see after a soft reset on a Palm device. Most Palm applications that are not designed to handle QVGA resolution, and these display in a square format with a black area at the bottom of the screen. This raises two questions: will there be Palm OS APIs to better support full-screen QVGA applications, and could the 80 x 240 area at the bottom of the screen be used for a handwriting input area? As far as I was able to see, Graffiti, a key component of the Palm OS, is absent from ALP at this time.

PalmSource had a Compatibility Station where Palm developers could try out their applications running in ALP's built-in Palm OS emulator. The applications I brought for testing as well as those of the two other Palm developers I spoke with ran perfectly in the emulator. About two thirds of the developers in the Developer Day sessions were Palm developers.

One of the most important new capabilities that ALP adds over the Palm OS is multitasking. Thanks to its modern Linux architecture, ALP is fully capable of running multiple applications concurrently. But PalmSource has wisely left it to the judgment of developers to decide whether their applications provide value running in a background thread. Unless specifically written to run in the background after another application is launched, ALP applications will exit and free up memory just as Palm OS applications currently do. This contrasts with Windows Mobile and Symbian systems where the default behavior is to leave every application in memory after the user closes it and launches another.

Many applications will benefit greatly from being able to run in the background, playing music, downloading email, or listening to a network for incoming data. Such applications have access to the screen even when they run in the background. For example, an email client can pop up a small window reporting the arrival of a new email. This window could display over a browser window, for example, while the browser continues to render a web page. ALP also provides an option for background applications to contribute icons to a status bar, which runs across the top of the screen. I was able to play an MP3, launch another application, and then bring up a mini control panel for the music player by tapping on the player's icon in the status bar. I could adjust the music and close the control panel without ever exiting the other application. On a small screen the status bar is scarce, prime real estate, of course, and PalmSource expects that the wireless carriers will jealously reserve access to it from their own built-in applications and services. But even if third party developers are restricted from contributing to the status bar, I was told that ALP will have APIs enabling developers to create task-switcher applications that deliver similar functionality.

ALP Contacts ScreenA quick look at ALP's PIM applications revealed screens that bore more than a passing resemblance to their Palm OS forebears. Performance was snappy on the Haier's 400MHz processor, but these still felt like "work in progress" to me, lacking some polish and features. Lots of applications were missing from the demos: the Phone application, email and messaging clients, notes and to-do, and the camera application, for example. Still, it's not too surprising that the platform had to be substantially complete before the applications that ride on top of it could be fully developed. Keep in mind in looking at these images that the background image, the fonts, the buttons, are all determined by a theme that can be configured at the system level. Also that the simulator exhibits a larger screen than most phones probably will. When you get a real ALP phone in your hand you can expect it to look quite a bit different from what you see here.

ACCESS continues to keep its NetFront browser on the forefront of advanced mobile browsers and even on the demo handsets I was impressed with what I saw. I'm not familiar with NetFront's current capabilities on other platforms, but on ALP it now has a thumbnail view of your bookmarks that makes navigating to favorite sites easy to do at a glance. One of the PalmSource engineers confided that this feature would be extended to the browser's history stack so that you can navigate forward and backward through the history of a surfing session without having to load each page as you go. This is one of the features that users have cited as making Nokia's new S60 browser one of the best for mobile devices, so NetFront is staying competitive.

Overall, I was impressed with the progress on ALP and happy to see it in the flesh. A lot of questions remain to be answered about how it will evolve between now and the end of the year. And then there is the question that is on every Palm user's mind: will we see ALP on future devices from Palm Inc? PalmSource officials were, as usual, "unable to comment on the plans of their licensees." But I came away thinking that at this stage in its development at least, ALP does indeed look very much like a successor to the Palm OS--quite frankly, more so than I expected it to. Whatever negotiations and discussions are taking place between Palm and ACCESS as I write this, it seems that Palm could do a lot worse than dub ALP--perhaps in a customized form--as the official next-generation of the Palm OS.

David Beers owns Pikesoft Mobile Computing which specializes in custom software development for Palm OS, Windows Mobile, and Java devices. He posts his thoughts on mobile technology and software development at his blog, Software Everywhere.

Article 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 View Full Comment Thread

ALP-OS doesn't have a snowball's chance of surviving.

The_Voice_of_Reason @ 8/20/2006 7:10:52 PM # Q
Sorry, Beersy but I'm not buying it. Access has a kludge of an OS and has effectively chopped the testicles off the PalmOS platform. Meanwhile they pretend to actually give a damn about PalmOS users. How long will it be until they completely dump the PalmOS emulator? 1 year? 2 years? Access isn't fooling anyone here.

I hope Palm is able to reach an agreement with the Access people to acquire rights to continue development of PalmOS, because otherwise we may as well all switch to mature platforms like Windows Mobile and Symbian when our PalmOS devices finally die. ALP is about as attractive to me as Copeland and BeOS were. Thanks, but no thanks.


RE: ALP-OS doesn't have a snowball's chance of surviving.
Foo Fighter @ 8/20/2006 8:36:05 PM # Q
Good God! These buttons are nearly identical in appearance to the mockups I created! I demand royalties!!!

Elitist Snob,

RE: ALP-OS doesn't have a snowball's chance of surviving.
ddd @ 8/20/2006 10:05:47 PM # Q
That's Ubuntu icon beside the application menu.

RE: ALP-OS doesn't have a snowball's chance of surviving.
cervezas @ 8/20/2006 10:12:02 PM # Q
ddd wrote:
That's Ubuntu icon beside the application menu.

Yep. We were all running the tools in Ubuntu 6.06 (Dapper Drake) laptops that PalmSource provided for the event. The simulator is running ALP in User Mode Linux, a way to run a Linux system virtualized within another:

David Beers
Pikesoft Mobile Computing

RE: ALP-OS doesn't have a snowball's chance of surviving.
KultiVator @ 8/21/2006 3:56:39 AM # Q
The_Voice_of_Reason wrote...

Sorry, Beersy but I'm not buying it.

Now there's a surprise Mr Doom & Gloom.

How can you possibly weave a negative spin out of such a promising report? Mr Beers is not trying to sell you anything, he's just providing his feedback on a key event that a great many people here are interested in reading about.

Whilst PalmSource/Access still have a lot of individual elements to work on to make this OS a success, I'd have thought you'd be glad to see signs that we might have an alternative future OS option that doesn't involve feeding revenue through to Billy-Bob Gates.


p.s. David Beers: Good article - thanks for taking the time to provide commentary.

RE: ALP-OS doesn't have a snowball's chance of surviving.
freakout @ 8/21/2006 5:28:09 AM # Q
^^ I'll second that. Promising indeed.

It irritates me still that Palm haven't commented on ALP. Dammit, how much could it really hurt them to just tell us which way they're planning to jump?

I apologise for any and all emoticons that appear in my posts. You may shoot them on sight.
Treo 270 ---> Treo 650

RE: ALP-OS doesn't have a snowball's chance of surviving.
KultiVator @ 8/21/2006 6:44:09 AM # Q
Freakout wrote:

It irritates me still that Palm haven't commented on ALP.

Yeah - I think they're too busy playing the corporate game of maintaining share-price, playing tough, sitting on the fence when perhaps it'd be a better strategy to show some leadership and give us tech-fans something to get excited about.

Realistically though, are Palm dynamic enough to take ALP and really run with it? Perhaps they are now too reliant on WinMob for their Treo offering to make a move to something better - or at least wish to appear that way in order to secure the lowest possible price for their ALP license!?! Who knows?

One thing's for sure, it'll be a sad day if the only ALP handsets available in the West are from unknown Chinese manufacturers (but it could be argued that's what we've been buying from Palm since at least the days of the original Tungsten|T).

So, perhaps we might see release hardware by 3rd quarter 2007?


RE: ALP-OS doesn't have a snowball's chance of surviving.
scstraus2 @ 8/21/2006 7:59:53 AM # Q
It is Palm that is killing the PalmOS, not PalmSource. It's not PalmSource's fault that Palm is too wankerish to move to a version newer than 5 years old. I will toss my treo in the bin the moment I can get any phone with a new PalmSource OS. Screw Palm.

RE: ALP-OS doesn't have a snowball's chance of surviving.
ChiA @ 8/21/2006 9:18:23 AM # Q
It is Palm that is killing the PalmOS, not PalmSource.

The PalmOS wasn't good enough for Palm, the blame lays firmly with Palmsource for not offering a compelling enough product. It doesn't even matter why Palm and other licencees were less than thrilled by Cobalt, what matters is that PalmSource's Cobalt failed to meet the needs of ANY and ALL PalmOS licencees.

Blaming Palm for the death of Palm OS is like the local baker blaming his business failure on his customers no longer buying his bread, yet "neglects" to mention it was because he tried to make a quick buck by switching to a new recipe of mouldy, overpriced bread.

To date,
That Cobalt loaf has been sitting on the shelf for a good two years and nobody, not Palm, Sony, Acer, Garmin, HP, Motorola etc have picked it up.

If Cobalt was the most fundamental change in computing since the introduction of Visicalc then companies would have been queuing round the block to get a slice of Cobalt.

Now it's not even worth feeding to the ducks - you'd be arrested on charges of animal cruelty.

Statistics are like bikinis. What they reveal is suggestive, but what they conceal is vital. Aaron Levenstein

RE: ALP-OS doesn't have a snowball's chance of surviving.
ChiA @ 8/21/2006 9:43:23 AM # Q
It's not PalmSource's fault that Palm is too w**kerish to move to a version newer than 5 years old.

The decision was up to Palm, not PalmSource. After all, nobody's knocking down your door to force you into upgrading your 286 EGA PC to a Dual Core Xeon machine.

Whether it was a wise move for Palm is another matter - a matter for the Palm directors and their shareholders to resolve. Again, customers aren't forced into buying what Palm has to offer. The custom will come for the right products at the right price. This is something the Palm directors have to focus on.

Statistics are like bikinis. What they reveal is suggestive, but what they conceal is vital. Aaron Levenstein

RE: ALP-OS doesn't have a snowball's chance of surviving.
just_little_me @ 8/22/2006 2:46:34 AM # Q
Stop the world. I want to get off. I actually agree with TVoR's headline... dammit.


RE: ALP-OS doesn't have a snowball's chance of surviving.
The_Voice_of_Reason @ 8/22/2006 5:28:27 AM # Q
Stop the world. I want to get off...


Be my guest. Hope you don't mind if I launch you into orbit by just breaking my foot off in your a$$ instead of sending you a one way ticket into outer space on the Shuttle. Next stop: Clue Planet™. Get on board, JLM.

By the way, it's time for you to start helping (the flailing) Beersy out by doing some Palm Apologizing for your masters here:


Reply to this comment

ALP looking good

dkirker @ 8/20/2006 11:18:10 PM # Q
Ok, so I too also had a chance to go to LinuxWorld this year. Sadly I could not go to the Developer Day, but I did spend about 2 to 2.5 hours at the ACCESS PalmSource booth. I had a chance to talk to PalmSource's Developer Relations Engineer, Bill Lee as well as another individual. Both were very helpful.

I describe some of my experiences at my blog, (I apologize for the crummy pictures, I wasn't too sure if I was allowed to do them and I didn't have time to make them perfect), but I will add to them here.

ALP is the only mobile Linux platform (on cellular handsets) that offers more than one executable option to the user (generally the only option is Java). ALP offers as David said: PalmOS Garnet, GTK, MAX, and Java applications. GTK and MAX kind of go hand-in-hand. There is an option that allows the developer to skin his or her application, which makes it look 1000000 times better. Native linux apps run on top of a mini X Windows system. In the past people had complained that this would be slow, but I only noticed a .5 to 1 second delay when launching (the X Windows cross-hatch patter was displayed, but that will probably change in the release version). The delay will probably drop by the time of the release.

At the PalmSource booth, one of the demoers had Bejeweled loaded on one of the phones. In the black space that David was describing there was a small DIA with home, menu, full screen, etc. buttons reminiscent of the latest PalmOS Garnet Simulator. Bejeweled seemed to run with no slowups. Bill Lee did say though that when running ALP over UML developers will notice lower than normal performance with ARM applications because UML actually emulates the ARM code, which tends to be heavy.

Remember, any current comments about ALP right now reflect a pre-pre-production state. ALP will get much better in time.

I also held a discussion with Bill Lee and this other PalmSource employee about multi-tasking. Their comments were as David said, it will be up to the developer. I had made some suggestions though about possibly allowing an option for the user that would allow him/her to choose a multi-task approach or a Palm launch style approach (close current app when launching a new). I also suggested using a taskbar like what QT uses (displays icons in the taskbar) or like what Palm does, a drop down that displays active applications (as opposed to recent apps, or maybe both). I was told that they would take this back to their engineers.

All-in-all ALP looked very very promising. I think that ALP will have a pretty successful future. There will definitely be more improvements in it between now and release (no time frame, but Bill thought January to February or March of next year looked about right).

Also, as a little comparison for those who have tried out the Linux ports that the Hack&Dev community has worked on. ALP was much faster than these ports.

I expect great things to come from ACCESS PalmSource. ALP is far away from being dead or vaporware like Cobalt was (RIP Cobalt, Long Live ALP).

-Donald C. Kirker

RE: ALP looking good
scstraus2 @ 8/21/2006 7:57:09 AM # Q
Finally some good news! I'm very impressed with what they have so far so much earlier than I expected anything. Most of the stuff a developer needs to get started is there. I would have liked to see the PIM suite looking a little bit more PalmOS like and ready for prime time, but overall an impressive offering.

If Palm doesn't adopt it, screw them, I'll go elsewhere. I know they won't make anything better, as they've never done anything but ruin the software they've taken over.

Reply to this comment

Font Support?

KultiVator @ 8/21/2006 4:03:43 AM # Q
Hi Guys,

Are you able to comment on font support in ALP - specifically on whether scalable (vector) fonts are likely to be directly supported?



RE: Font Support?
ChiA @ 8/21/2006 4:21:47 AM # Q
I hope that ALP will support Unicode fonts and multiple languages too. I'll be surprised if they don't when you consider ACCESS is aiming for the Asia-Pacific market.
RE: Font Support?
scstraus2 @ 8/21/2006 8:04:25 AM # Q
I'd assume that if they're using a scalable vector UI, they'd also be using scalabale fonts (maybe not vector, but some type at least). Fonts are definately one area that needs big improvement over current PalmOS>

RE: Font Support?
cervezas @ 8/21/2006 8:42:50 AM # Q
I can't say for sure about the fonts, but keeping ALP's GTK heritage in mind I can say that we probably have a much improved situation over the limited scaling of fonts in Palm OS. It was a priority in Cobalt, you'll recall.

One thing I should mention is that scalable vector graphics are very processer-intensive things to render. I would expect (in fact it was even hinted to me) that a lot of stuff would be somehow pre-rendered to avoid a performance hit. If that's true we shouldn't be expecting stock ALP devices to have Mac OSX-style dynamic scaling of fonts and graphics (i.e. smooth growing and shrinking based on stylus movements).

David Beers
Pikesoft Mobile Computing

RE: Font Support?
#*377 @ 8/21/2006 2:48:48 PM # Q
Well, GTK and FreeType play well together, maybe we'll get that built in. And the nice part about moving to linux is that it should be easy to port. The real challenge, as with all eye candy, is going to be the performance of the blitter. Hopefully we'll have some kind of framebuffer access (crosses fingers.)

Also, while I'm all in favor of having nice app frameworks (MAX, J2ME), I hope that PalmSource keeps true to it's DIY roots and doesn't discourage this sort of thing.

Reply to this comment

ALP = looming Palm OS disaster

VampireLestat @ 8/21/2006 4:57:37 AM # Q
Let's all get ready to migrate to WM because this ALP is going to fail. Why you ask?

This is not Palm OS on top of a new kernel. It is a new OS that has a Palm OS emulator; and only 68K emulation to boot (and only for WELL BEHAVED PROGRAMS!). If I wanted a Palm OS EMULATOR, I would buy a PPC with the StyleTap Palm OS emulator!

Bring back Cobalt. I am not spending a second more even thinking about ALP, this is trash, it is not Palm OS, it will spell the death of Palm OS. No go. Either it goes, or I go to WM.

RE: ALP = looming Palm OS disaster
VampireLestat @ 8/21/2006 5:06:58 AM # Q
And another thing...

This is not Linux, so don't let anyone be fooled by the "free" aura of the Linux name. There is nothing Linux about this crap. It does not run Linux programs, and for the end users, they are paying for the OS when they purchase a new device.

So what the hell? If you are going to pay for a multitasking OS, why not go for WM? At least IT has a based of tens of thousands of established programs. ALP does not tap into the Palm OS software base, it cripples it.

ALP will destroy the Palm OS economy. If you are a developer who makes $$$ from selling Palm OS programs, you better have WM versions ready soon because ALP will a) present programming challenges and new costs and b) no one will buy those programs.

You can't build a whole brand new generation of ALP programs overnight. Who the hell is buying this bs?

Consumers will go with WM and those who desperately need a given Palm OS program, they will run it in StyleTap.

RE: ALP = looming Palm OS disaster
VampireLestat @ 8/21/2006 5:15:08 AM # Q
I really feel sad for the guys at PalmSource. They don't deserve to go down in flames and failure like this. Man...

Guys, go apply for a job at Palm Inc ASAP, while you still have time. I can't believe Palm Inc. is going to buy this. They got to be making their own OS. They might still be needing programmers in the near future.

RE: ALP = looming Palm OS disaster
VampireLestat @ 8/21/2006 5:26:57 AM # Q
Oh, if any of you own Palm Inc. stock; the second Palm announces it is trying ALP, call your broker and sell everything. If I had Palm stock, I would.

Anyways, I am confident Colligan is not a total retard; the man is after all one of the famous founders of Palm OS, he has got to have SOME brains. Colligan surely can see this will never fly with the Palm OS community. He better have plans to make a TX with WM and/or plans to make a PALM OS ON A WHATEVER KERNAL, not a new OS with a Palm OS emulator.

He has to find a way to get OS 5 and / or OS 6 full property rights away from Access; he just has to.

RE: ALP = looming Palm OS disaster
KultiVator @ 8/21/2006 6:52:39 AM # Q
Vampy dude...

Read the article again, Mr Beers indicates that current arm appls are handled by the emulation (so not just the legacy 68k apps you mention).

The fact that BeJewelled is running without noticeable slow-down is a pretty good testament to the quality of the emulation - IN THIS PRE-ALPHA OS!

Nuff said.


RE: ALP = looming Palm OS disaster
scstraus2 @ 8/21/2006 8:10:34 AM # Q
This is an idiotic thread. Access demoes almost flawless emulation on an Alpha release that no one expected to see anything of for 6 more months, and you're complaining? What would you rather have? Frankengarnet version 5.9.999.999.999a?

RE: ALP = looming Palm OS disaster
cervezas @ 8/21/2006 9:00:49 AM # Q
VampireLestat wrote:
This is not Palm OS on top of a new kernel. It is a new OS that has a Palm OS emulator; and only 68K emulation to boot (and only for WELL BEHAVED PROGRAMS!). If I wanted a Palm OS EMULATOR, I would buy a PPC with the StyleTap Palm OS emulator!

Bring back Cobalt.

Actually, Palm OS has been emulated ever since the first ARM device was released many years ago and it's emulated on Cobalt, too. There are some technical differences, but none that have a major impact on users, unless the developer has been using undocumented, unsupported APIs that PalmSource has said for years could change at any time.

I actually heard things at the conference that led me to believe that Palm OS may even be somewhat enhanced by running in ALP. One engineer I spoke with hinted that the Palm OS isn't completely sandboxed within the emulator--that it could access certain resources outside it. Don't know what those resources might be but lets just take a simple example: the file system. If the Palm OS virtual file system can access the same file system that the Linux apps are using, that could open up some interesting ways to extend Palm OS applications by leveraging Linux services running in another process. Imagine a Linux daemon that periodically fetches RSS updates, email, or GPS position data in the background and saves it to a file that a Garnet application can use. This could drastically improve the perceived performance of Palm OS applications that use such data since most of the heavy lifting could be done before the Palm OS application even launches. Not saying I know for sure this is possible, but I got the impression it might be.

David Beers
Pikesoft Mobile Computing

RE: ALP = looming Palm OS disaster
#*377 @ 8/21/2006 11:38:04 AM # Q
Well, if you could write to the old palm filesystem from the new OS, then you would have a way to update those good old PDBs automatically (write the PDB as a file to the old VFS from the new OS and then use the old garnet DM calls to turn that back into a PDB.)

It's not giving the old app multi-threading, but it certainly would take care of one of the major reasons for wanting threads in the old OS, IMHO.

Using Linux to beef up Garnet's capabilities
cervezas @ 8/21/2006 11:57:19 AM # Q
Well, if you could write to the old palm filesystem from the new OS, then you would have a way to update those good old PDBs automatically (write the PDB as a file to the old VFS from the new OS and then use the old garnet DM calls to turn that back into a PDB.)

Another way of looking at it, yes. Underneath the emulation layer I'm sure those DM read, query, and write calls are proxying down into the Linux file system (probably to SQLite). At that point it becomes a matter of security configuration whether non-Palm OS applications can read or modify this data.

It's not giving the old app multi-threading, but it certainly would take care of one of the major reasons for wanting threads in the old OS, IMHO.

Exactly. Not all of the major reasons, but one significant one. Of course the code that operates on this data in a background thread is going to be Linux code. But since ALP can run Linux daemons it's possible that there will be background services available (even in ROM) that would be of interest to Palm developers. This is all very speculative, of course, and there's a real question about how much ACCESS really wants to make this possible rather than nudging Palm developers more firmly over to MAX for these benefits. It's just interesting to think about.

David Beers
Pikesoft Mobile Computing

RE: ALP = looming Palm OS disaster
dkirker @ 8/21/2006 1:12:27 PM # Q
VampireLestat, where do you think ALP is not Linux?

If you take the time to read one of ACCESS PalmSource's whitepapers, specifically "Making Linux Mobile, The ACCESS Linux Platform" under the section titled"The Access Linux Platform" you will realize that Linux is the underlaying technology.

Also, as David said, PalmOS apps have been emulated in PalmOS 5 and 6. The only thing that is not emulated is the ARM code. Why, because, the processor is ARM based. Mark my word, any PalmOS app that worked in PalmOS 5.4 will work on ALP. This is because GHost is PalmOS 5.4 (Garnet).

I am sorry if you don't see any future in ALP, but from what I took away from this conference, ALP's future is very bright.

RE: ALP = looming Palm OS disaster
AdamaDBrown @ 8/21/2006 4:10:42 PM # Q
I believe that what VL is saying is that there's no Linux software base for this OS. Mobile Linux won't run desktop applications any more than Windows Mobile will, and mobile Linux program compatibility is fragmented between versions. So they're basically starting from a blank slate in terms of software.

RE: ALP = looming Palm OS disaster
dkirker @ 8/21/2006 4:44:45 PM # Q
To an extent.

Bill showed me a demo on how simple it is to make a Linux app work with ALP and how to alternate between the themed version and GTK version of the app (when compiling).

The app he used was gworldclock, Google it. He first showed me how it ran on Ubuntu (little Ubuntu page on gworldclock: and then did a quick recompile with a little xml file (defines theme, app name, essentially probably just creates a small database file that ALP uses for the launcher) and had gworldclock running in ALP in seconds. Gworldclock by the way is a desktop app. So, it really depends. You can get some Linux desktop apps to run in ALP. Mainly the issue is with the interface. Most desktop apps (well take Konquerer) have to be recompiled for the mobile device because they need to be redesigned for smaller devices.

Anyway, there is an app base. 3-4 bases to be exact.


desktop Linux app migration not a problem
cervezas @ 8/21/2006 5:04:28 PM # Q
dkirker is right. Thanks to the decision to use X Windows it looks pretty trivial to recompile an existing GTK desktop application for ALP and we spent some time doing just that in the developer sessions. For GUI apps you'd want to reconsider how to lay out the screens, of course, and you wouldn't be able to run something like the Gimp because of resource constraints. But the beauty of open source is you could unplug unecessary features from many desktop Linux apps and make other tweaks to get versions that would run on a handset. You could also do some creative and interesting stuff with non-GUI Linux applications that run as services on the device and talk with your GUI apps. It's going to be a lot of fun.

One more example: a lot of my enterprise and government clients have me develop SuperWaba applications for them (like J2ME only much better) and all those apps should be supported very easily on ALP since the SuperWaba VM has already been ported to Linux. SuperWaba enables you to create and use native libraries, so that's going to open up a nice new world for that platform.

David Beers
Pikesoft Mobile Computing

RE: ALP = looming Palm OS disaster
Sam H @ 8/26/2006 12:02:43 PM # Q
Beersie, only you could describe anything involving the X Window System as 'a lot of fun.'
RE: ALP = looming Palm OS disaster
freakout @ 8/26/2006 12:56:14 PM # Q
Hey, the X Window System isn't so bad. You just have to get it drunk. It's a demon at parties. Dances like a mofo.
RE: ALP = looming Palm OS disaster
Sam H @ 8/26/2006 2:51:50 PM # Q
Doing anything with X is an exercise in pain. Open sourcing a project can achieve a lot of things, but (on its own) it can't fix fundamental design flaws.
Reply to this comment


twrock @ 8/21/2006 6:52:33 AM # Q
"...the kind of handset that PalmSource has in mind as a primary target device: QVGA (240 x 320) screen resolution...."

Remember back in the day when the Handera 330 had a QVGA screen and the arguements went back and forth about the "best" screen resolution? This just cracks me up. Those Handera people got a lot of things right.

Nice report David.

It might not be the "mythical color HandEra", but I'm liking my TX anyway.

Reply to this comment

Too Little, Too Late

Gekko @ 8/21/2006 7:16:58 AM # Q

This is a marathon race and WINMOB is about to cross the finish line and ALPO is still sitting on the couch eating potato chips.

RE: Too Little, Too Late
scstraus2 @ 8/21/2006 8:06:39 AM # Q
Correction, Palm is on the couch eating potato chips. We never even saw Cobalt. It could have been the greatest thing ever and we wouldn't know. Frankly I don't even care. I just want an upgrade over Garnet. It is Palm that is killing PalmOS, and it will be them I ditch first.

Right now Access is trying to save PalmOS, while Palm is trying to kill it. I'm very happy with this performance of an Alfa release. I'll buy a chinese handset if I have to.

Screw Palm. They can go destroy WINMOB.

RE: Too Little, Too Late
Foo Fighter @ 8/21/2006 12:31:29 PM # Q
>> "This is a marathon race and WINMOB is about to cross the finish line and ALPO is still sitting on the couch eating potato chips."

This isn't a winner-takes-all race, and no single platform is destined to achieve total supremacy. So nobody is going to be crossing any finish line here, because there isn't one to reach.

Access's conundrum lies in the fact that Symbian and Windows Mobile are the established players in this game (as is Linux, but on a completely different scale) and are already well on their way enjoying greater adoption and market penetration while ALP isn't even in the race yet. Given the platform's roadmap, it will be at least another year beyond release (providing Access ships this OS in Q4 as promised) before ALP even appears on hardware. We may not even see an ALP-based smartphone until 2008. Just that reality is bad enough, but it gets worse...

The next trick will be to convert all those "thousands" of PalmOS developers PalmSource claims to be sitting on into ALP devs. It's not a foregone conclusion that developers will even embrace ALP. Many could simply stick with what they already know (Garnet), leaving ALP to inherit a library of elderly legacy applications with few native MAX apps of its own.

Seducing developers into learning and adopting this new platform may prove to be an uphill battle as the majority of what were once strictly Palm developers are already now writing code for Windows Mobile and Symbian as well. If Palm thumbs its nose at ALP, choosing instead to limp along with Garnet, you can forget about ALP becoming the successor to the PalmOS platform.

Some may see this upheaval as the appropriate time reevaluate their own product roadmaps and cast their lot with WM and Symbian, cutting the cord with PalmSource.

Will that happen? Who knows. It's impossible to know how the chips will fall until ALP actually hits the market, and we can gauge developer and vendor reaction.

Elitist Snob,

RE: Too Little, Too Late
joad @ 8/21/2006 5:42:40 PM # Q
>Screw Palm. They can go destroy WINMOB.

It seems pretty apparent that Microsoft doesn't need their help - they seem to be doing just fine making it an awful OS without anyone else's assistance.

Reply to this comment
Start a New Comment Thread Top View Full Comment Thread
Achtung! Only the first 50 comments are displayed within the article.
    Click here for the full story discussion page...


Register Register | Login Log in