Comments on: PalmSource Acquires China MobileSoft, Will Develop Palm OS for Linux

It did not Amiga much them much...
The point is that PalmOS-under-linux will be inferior to customized linux... Well, maybe not technically inferior, but the price/overhead/cost/features ratios won't be good for PalmOS...

That's more a death announcement to me...
Farewell, palmOS...

http://www.chinamobilesoft.com/product/mlinux.asp

"All opinions posted are my own, and not those of my employers, who are appalled."

As for running on top of Linux? It might not of worked for the Amiga, but the same idea seems to be doing wonders for the Macintosh platform. Macs have really gained a lot of respect in the *nix community since OSX was released.

If PalmSource can pull it off, this could really open up a lot of new markets for them in non PDA areas. Palm OS on your TiVo anybody?

>What happened to the Cobalt kernel?

Still there in Palm OS Cobalt. Palm OS for Linux is an addition to the line, not a replacement.

>Why switch to Linux so late in development?

We didn't switch, we're adding. We'd been thinking of supporting Linux for some time. CMS had already done a lot of Linux work, letting us accelerate our development.

>What happened to the Be technology in Cobalt?

Still there, and we plan that it'll be in Palm OS for Linux as well. We'll be moving the Palm OS Cobalt software frameworks (graphics, multimedia, wireless, etc) onto Linux. Much of the Be technology went into them.

>Swithing the OS core is not as simple as it seems,

Agreed.

>I doubt many apps would be compatible with the new operating system.

Our intent is to include the Palm OS emulator in Palm OS for Linux, so properly-written 68k apps should run. We'll also support apps written for the new "Protein" APIs in Palm OS Cobalt. Those apps may require a simple recompile in order to run in Palm OS for Linux.

>Sounds like they're skinning the palm interface over linux

I don't know what's the best term to use, but the way we're describing it is that we'll put the Palm OS interface, apps, and frameworks on the Linux kernel.

Hope this helps.

Mike
CCO, PalmSource

I know the buzz says there will be separate OSs, but watch the gemstones go flying out the window if this takes off.

I'm honestly left simply hoping for the best.
I really hope this result in us the users getting better operating on more and better systems.

I do wonder about PACE emulation here. Let's see 68K apps running thru a Cobalt emulator which is built on top of Linux kernal. Whew, somebody better get another truckload of "Zen". We may be runnin' out. :-o

Pat Horne; www.churchoflivingfaith.com

PalmOS can still stand on it's own as far as I'm concerned. It's funny all the comments we run into about "limitations", while in the real world, others wonder how we can do the things we do. Yea, I'm a hardcore user who has gone through at least 12 devices in 4 years including those with Windows Mobile, and we no longer have PPC evangelists looking for confrontation at our workplace of 13,000. However, that is not because I'm any great expert, it's because I have a very capable tool and know how to use it.

This change is not about fixing weaknesses, it's about embracing a growing segment of technonologically astute users, and providing for future expansion.

I'd say more, but I'm in a rush right now...

The last paragraph is the big announcement - they'll put Palm OS on top of China MobileSoft's mLinux. That is good news.

"voice of chaos"??? W T F???

So you joined PIC yesterday and your first post was to hint at the Linux Palm before it was officially announced? Interesting. Conspiracy Theorists are working overtime.

Ryan, this Palm "insider" is hurting my feelings with his user name. Is he just another developer, or is he really an Astroturfer working for Palm, "created" to ask a bunch of nicey-nice questions for Michael Mace to answer + somehow spin this flip flop into a positive? ;-) Palm wouldn't do that, would they?

TVoR

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

I smell Astroturf.

Looks like Palm is sinking to the kind of manipulation games Microsoft is legendary for. Wow.

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

http://www.palminfocenter.com/comment_view.asp?ID=7359#101041

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

http://www.palminfocenter.com/comment_view.asp?ID=7359#101041

A dumba$$ from TapWave. Way to be, G. Wrap your loose lips around Mike Cane's flaccid member.

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

Sharp went Symbian!!

http://www.symbian.com/press-office/2004/pr041206.html

PDAs are dead.

They all think it's now the Smartphone.

Everything there is great, but let's look at this from both the short term and the long term for the software developer.

From the short term, at this point there are no Cobalt devices shipping. Even before a single device has shipped you announce a new OS platform. Since it will support PACE, that is the API set that I can guarantee maximum portibility, ergo for at least the short term PACE is my only real alternative.

In the long term the OS set that you provide will be decided by the marketplace. Maybe Cobalt will be more successful, maybe the Linux-based Palm OS will win. Maybe they will win in different segments. In either case, if I want to really take advantage of what the platform has to offer, I will focus on what is selling.

The move to Linux was smart. I like it a lot. You've just got to make me understand why I would want to move to Cobalt - right now I'm glad I haven't spent a lot of time learning the Cobalt API's.

Don't tell that to "Jumpin'" Jean Louis Gassee - the guy who sold them BeOS for millions and who is now the same guy who is now the CHAIRMAN of PSRC!!! What's wrong with that picture?

Benhamou sure knows how to cut deals!

-------------------------------------------------------------------
Disappointed Palm User:
"For PalmOS 6, I had expected that Palm would have been brave enough to think UNIX flavor + pretty GUI + PACE clone (the "emulator" I naively envisioned being folded into the Clean Sheet™ OS) running on modern 400MHz processors. Give 30 experienced engineers 1 year to bind all these elements together. I think it could have (and should have) been done."

PalmSource Dominatrix:
Easy claims, for someone who is essentially posting anonymously. Tell me, what is your engineering background to be able to make such a statement? Have you thought about these things:

* What the heck do you mean by "UNIX flavor"? Do you know what that is? (Bonus question: do you know what it is about the ARM architecture that makes it very difficult to efficiently implement the traditional Unix process model?) Are you talking about licensing an existing kernel like Linux or FreeBSD? Sure, licensing this stuff makes sense... which is why we licensed STREAMS for our IO subsystem. The kernel itself is a very small part -- we actually have only a few engineers working on ours, not much more than what would be needed to do customization and integration if we had licensed it. Oh and think about this: Be already had a fully functional Unix-like protected memory kernel, one that had been in production use for many years. Cobalt did not ship with the BeOS kernel. Are you saying the Be engineers are so stupid that they just kind-of forgot they had this thing?

* What is this pretty UI? Are you talking about running X Windows? Maybe embedded KDE? What are the trade-offs of these? If you aren't talking about licensing something, how is this any different than what we did for Cobalt?

* PACE clone. Cobalt -has- PACE. That is what runs all of the existing applications. The main difference I can imagine between what is there and what you are thinking is that in Cobalt PACE (for the most part) doesn't take care of translating the old system calls into what is needed in the new system architecture. Instead, we implemented that part of the compatibility layer outside of PACE so that (a) ARM native applications could also use them (to ease porting of existing applications to ARM); and (b) for the initial release we could re-use and grow the existing APIs instead of introducing a completely new application framework. This decision was a significant measure in cutting down on the time to market: by making the traditional APIs available in ARM it was much easier for us to port all of our own existing apps, and we could avoid a -ton- of work in designing, documenting, and testing a full set of new APIs.

Enough of the spin. Enough of the games. Get your house in order and let us know when you know when you finally figure out what you're doing (assuming that ever happens).

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

Congratulations, Mike. Where would you like your prize (a used Tungsten E) mailed to?

It amazes me that Michael Mace is brazen enough to show up on Palminfocenter with his dog and pony show. Many will find his attempts at spin doctoring to be an insult to the intelligence of the Palm enthusiasts who look to PIC as the place for HONEST information regarding the PalmOS world.

I'd be shocked to see that the market buys PalmSource's spin. PalmSource just reached a 52 week low price today and is still falling. Realistically, it will likely drop below $10 within a few days.

Desperate moves from a desperate company. Palm has basically just written off the Be acquisition, Cobalt and the past three years of development, all in one fell swoop. And does PalmSource think the market will keep waiting for them forever? Brilliant.

Just wait until you see what's in the PalmSource's SEC filing on December 22!

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

Plus, he's too probably busy "sucking at [Palm's] malnourished teats" to care!

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

"Now we have some Linux nerd who says, Oh, I'll just type 'make' and boom! PalmOS on Linux."

I think part of the problem here is that there is a lot of misunderstanding over what a kernel is. This confusion isn't helped by the fact that, though Linux is a kernel, mostly people use the name "Linux" to describe an entire platform.

Here is basically what a kernel (including Linux) does: manage processes, manage threads, load and start executables, provide shared libraries, allow you to do IO (with persistent storage and networks), and allow you to send basic commands to hardware such as "send a byte on a serial port".

That's basically it. Note that Linux doesn't even give you the famous shell. That is why you have so many to choose from (bash, csh, zsh, etc., etc): they are all built on top of these primitive services that the kernel provides.

So when people use the word "Linux" what they are usually referring to is the Linux kernel plus a huge amount of other stuff. This is what Linux distributions do, take the basic Linux kernel, and then put together with all of the other pieces (shells, servers, X-Windows, KDE, etc., etc) needed to make the kind of full platform you expect.

Now when we say we are going to make Palm OS run on top of Linux, we are talking about the -real- Linux, the kernel. On top of the kernel, instead of a lot of the pieces you get in a normal desktop distribution (X-Windows, KDE, file browser), we will be placing the current Palm OS platform components basically as they exist in Cobalt.

Now it just so happens that a large amount of the effort we invested in developing Cobalt was to redesign the entire Palm OS architecture so that it would work with and make use of a modern OS kernel with multi-threading, protected memory, etc. One not completely unlike Linux (though not nearly as similar to Linux as the BeOS kernel was, but that's another story).

What this means is that we expect the new Protein SDK we designed for Cobalt to basically work as-is on Linux. There is quite a bit of effort we will need to do inside the system to do the port, but with Cobalt we have already addressed the vast majority of the issues that anyone outside of PalmSource would see: how we make these APIs work in a protected memory system, how we expose multi-threading and multiple processes in them, etc. That is why applications developed for the current Protein APIs will compile as-is on this future Palm OS on Linux.

"I've investigated the Linuxed Sharp Zaurus. You practically need a goddammed engineering degree to load apps onto that thing. And even the Linux cultists couldn't tell me HTF to get FONTS onto it to use with the Hancom WP app!"

You absolutely positively don't need to worry about this, because what we are talking about here is the Linux kernel. We will be providing all of the higher-level services, the same ones you know and love today. To an end user the final platform will work, look, and feel basically exactly the same as it does on our existing kernel.

"We've already had hackbod here (I like her; so shut up, youse guys!) defending the telecom-dictated security measures that made Cobalt much more "closed" that PalmOS has traditionmally been. And now you expect these same telcoms to go for LINUX?! Something so OPEN? And not just Linux, but CHINESE Linux?"

If we were to take the linux kernel, create a process, and let you run whatever code you wanted in there, we could make it so your code could do absolutely nothing. (Unless there are bugs in the kernel that allow you to get around its security, which the Linux community would be aghast at and quickly fix since they like to tout their security compared to Windows.)

The fundamental feature that allows the system to enforce this kind of security is protected memory, which both our kernel and Linux has.

When talking about these things, it is important to be careful about over-loaded terms like "open". In Linux, what "open" means is that anyone has access to the source code and can modify what it does. This does NOT mean you will be able to have any arbitrary access to a particular system running Linux, if that system doesn't let you. Sure, you will be able to get the source code for the kernel in that system, but the only advantage at all that gives you in avoiding security is that you might be able to find a bug in the kernel code to take advantage of.

"Not one device with Cobalt yet. Eleven unnamed ones set to announce next year -- allegedly (and, yes, not ALL of them will be doing Cobalt -- and that could mean as few as ONE allegedly will!)."

Do you mean some of these may be running the Linux version of Cobalt? If so, given the time it takes for a device to get to market and the fact that this acquisition hasn't yet completed, I think it is fairly unlikely you'd see Linux-based devices next year.

"But, damn, these were *Be* people! HTF could you have so much talent and yet have so LITTLE to show for it?"

Wow, strange to see a nice comment about Be engineers. Thanks. :)

For what it's worth, I consider the current Cobalt product to be a tremendous engineering achievement, which everyone at PalmSource has reason to be proud of. I can't wait to buy a device that is running it!!! And the vast majority of that work will carry over to PalmOS for Linux... just like much of the work we were doing at Be carried over from BeOS to the Cobalt kernel (with a brief stop in Windows Land).

--
Dianne
PalmSource Application Frameworks Manager

The OS described above is, in fact, little more than what we have today on Palm OS Garnet, except running on top of Linux instead of the in-house kernel it currently uses. But you basically couldn't use anything in Linux, because none of the rest of the Garnet platform knows how to do anything with multiple threads, protected memory, multiple processes, or anything else like that. About the only advantage you would get with it is you might be able to write Palm OS applications directly in native ARM code.

Nice try, DK. Your post doesn't make any sense. It's pretty obvious that the "Unix flavor" listed as part of what I said should have comprised PalmOS 6 includes a Unix-type kernel. Otherwise, what would be the advantage over using PalmOS 5? Think about it.

"Palm has basically just written off the Be acquisition, Cobalt and the past three years of development, all in one fell swoop."

This is absolutely not true. While we don't know many of the details right now about how all of this will work out, it seems likely that nearly (if not all) of the Be technology that is currently in Cobalt will carry over to the Linux flavor. This gets back to the fact that the overall platform is much more than the kernel, and the kernel we ended up using in Cobalt was not from BeOS. (As an example: did you know that during its lifetime so far Palm OS has gone through three different kernels, and this is merely a fourth?)

Nobody is trying to pull something over on you. Read right from the announcement: "bringing the benefits of Palm OS to the Linux community, including ... software frameworks based on the best of Palm OS and BeOS". This is all that stuff on top of the kernel that, to the developer and the user, largely makes Palm OS what it is. It is the Data Manager, event queues, windows, forms, notifications, attention manager, sync, and a ton of other things. What this announcement means is that we are basically bringing all of these Palm OS things (along with the significant new infrastructure we have in Cobalt) over to Linux, very much the same way that MacOS X uses Mach/FreeBSD for its kernel.

This announcement also doesn't change the fact that Cobalt -does- exist, licensees -are- working with it, and they can ship devices using it -today-. The work on Linux is a longer-term plan, which can be done partly because we have the Cobalt platform today that we can ship. And as has been stated, this switch will be largely transparent to users and developers: they will see the same user experience, the same Protein APIs, etc. In fact it should be much more transparent than the switch to Cobalt, since here we are only changing the underlying kernel rather than all of the system architecture on top of it.

--
Dianne
PalmSource Application Frameworks Manager

Again, nice try. PalmOS 6 (Cobalt) has been given the hook. Licencees gave it a resounding "No thanks" (is it true TapWave actually laughed in your face?); it was a year late; and it's STILL awaiting several features to be added, including some mandatory for use as a smartphone OS. PalmSource is attempting to salvage whatever it can from the past three years, ergo the Linux-based Plan B (9?) from Outer Mongolia. Good luck trying to see how much of your Cobalt-related code can be recycled. "Reduce/Reuse/Recycle code" - that will be PalmSource's offical mantra for PalmLinux. Should make for a really Green OS.

Bottom line is that PalmSource screwed up and is desperately attempting to salvage whatever it can from its investments. You actually STILL might have had a chance to succeed it only you had a bit more time. Unfortunately, the rest of the market isn't standing still and PalmSource is going to get squashed by the big boys. As someone that worked for Be, this scenario should be eerily familiar to you. Maybe it's time to start looking for another eVilla to licence PalmOS 6 to.

TVoR

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

Pass the donation plate, Reverend.

Linux=the bandwagon du jour in '05 (according to PS at least)

What do multiple bandwagons end up becoming? A convoy of wagons.

What do wagon convoys do when the going gets rough? Circle them tighter and tighter and tighter until the inevitable doom occurs...

Again, friends, I'm going out on a limb here and predicting IBM will have SOMETHING to do, hardware, software, or otherwise, with whatever house of cards Linux/Palm OS amalgamation ends up being cranked out in the next year or so. Anyone remember the PR hot air IBM was generating two or three years ago ("we invested $1 billion+ in Linux because it's the wave of the future" etc)?

Now that Acer's safely out of the picture, P1 can start to worry about Lenovo cranking up handheld production.

Nevertheless, IBM can go to the pinstripers and say "here's your PDA running Palmlux (my vote for the new Tower of Babel POS/Linux hybrid), here's your smartphones running the same thing, here's your thin client, here's your Windows desktop that still has our name on it if you still care for such things, here's the drastically chapened Thinkpads still with our name on it, here are your servers, mainframes etc etc etc. Now here are our software & service offerings"
It all makes such perfect sense.

This linux version will use the EXACT same APIs as Cobolt. Its mainly the kernal and some services that are linux. What does that mean? It smeans the code will run fine (just change the compile settings in PODS) on this. The dont need to learn any new APIs. The code for Cobolt is really the Protein APIs - and this is exactly the same whether is is OS 6 Cobolt or OS 6 Linux. ......at least thats the theory;)

"All opinions posted are my own, and not those of my employers, who are appalled."

That said, let's remember that JVC never put out their PPC and that ViewSonic dumped theirs.

According to PalmSource (http://tinyurl.com/5jc73), "Palm OS for Linux will be able to run many third party Linux applications and services (GUI applications will need to use the Palm OS APIs)."

Later in the same letter is this question and answer:

"Q. Will you enable developers to port or write "native" Linux applications that run under Palm OS for Linux?

"A. Yes. Given the open source nature of Linux, developers will be able to write applications directly to the Linux core. There are many development tools for the Linux community that will address the needs of this audience very well."

I enjoy poking fun at P1/PSRC's foibles quite a bit.

I didn't want to go on doing it without also saying that I appreciate you coming on here to try to inform us.

Whether you are speaking for yourself, OR for who you work for, it takes a lot of guts to come on here and suffer our slings and arrows.
For this, I salute you sir.

I want PalmOS and the hardware running it to be something we can cheer about, and I wish you best of luck in your efforts.

For now, I'm staying with my TH55, but I'll continue to look on with interest and sincere hope.

Here's the name: PalmNix.

As in: Just nix this friggin thing!

I'm coining the name Punix.

-------------------------------
Editor, http://Pocketfactory.com
Contributing Editor, http://digitalmediathoughts.com

But I like the Palm OS and its hardware (so far, so good!), its evolution is what I'm really worried about!

I wanted to buy the latest hardware technology with my preferred OS but compared to SYMBIAN (Smartphones & other mobile phones) and Windows (multiple mobile OS flavors), why should Palm always follow and not lead, or at least be neck-and-neck with the competition?

With Palm, it really is a good thing to wait if you want to get your moneys worth! But what would it do to benefit the platforms evolution, to PS and its licensees business in the long run if this long-wait continues?

Hope it could be better.

m505 > Visor Prism > Visor Prism + Visor Phone + EyeModule + FlashPlus > Treo 600 ( this holiday season! )

Because they didn't have the application developers, GUI, user community, or brand name. Palm, on the other hand, has all that.
What Palm needs is device support, and Linux can deliver that because phone makers and device makers have it already ported and have lots of drivers for it.

Why should a "vendor" be able to produce a stable anything? Vendors sell stuff, they don't produce stuff.

Anyway, PalmSource is an application and library developer and has never been much of anything else. They used to buy their kernel somewhere else, and now they get it for free from Linux. No big deal.

"Improve further on the anemic built-in apps?"

Anemic as they may be, they are the reason people buy Palms, because they still beat the available alternatives.

The kernel is pretty unimportant these days, as long as it's fast, mature, stable and the company using it, has a measure of control over it. You're getting the APIs and the GUI, so tell me, what's the difference?

The real question of course, is whether or not it will work. Personally, I think the company is now on its third shock from the paddles, and is still not showing any real signs of brain activity.

So yes, no matter what the press release says, Cobalt is finished (but can you imagine what would happen if they actually admitted that?! It would be like Apple and Taligent/Copland all over again!), and this is the new plan.

As I said, I'm not convinced, but then it was pretty obvious that Cobalt had hit the skids a long time ago, and they have to try something. Good luck to them.

Stay in school, Bubba.

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

At any rate, since the writing's clearly on the wall for the broken-down old Mule that is the Palmconomy, there's never been a better time to: A. Thank Ryan for his timeless, often thankless work put into maintaining the site. B. Make the offer that if there needs to be a "Save PIC" shortterm fund established anytime soon, I am sure that all of us "old school" regulars can chip in a few PayPal $. And yes, I do click those banners...sometimes.

I completely agree that you should consider putting a PayPal donation link for PIC on the site's main page.

If things get ugly in the Palm Economy in 2005, you shouldn't have to scrounge for money to keep the site going. And please don't become another PalmStation.

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

http://www.palminfocenter.com/view_story.asp?ID=6866&curpage=4#94839

And if Apple does indeed put credence to that iPhone thing, whether or not it would converge with the iPod evolution and if it proves to be a worthy alternative to the Treo; and if Cobalt and / or this Palm OS over Linux thing wouldn't get its act over and done quickly right away to produce new hardware as promised; I might be persuaded to go for Apple once again for my mobile fix!

Bad thing for Apple (Newton division!) and eventually a good thing for Palm though, because they've benefited from the Newtons demise in the past. But the iPod & Mac OS X is now tightly integrated with each other (and by itself), than what the Newton & Mac OS was back then. And if PS couldn't get their act together in time, the reverse could happen as well for the demise of Palm.

I like the the Palm OS, my old Visor Prism and the Treo 600, (but not the Treo 650), eventhough I crave for it's Bluetooth connectivity and other minor features added. But what I really want now is an all-in-one convergent device that does everything effectively, that's all.

m505 > Visor Prism + Visor Phone + EyeModule + ThinModem Plus FlashPlus + SmartPad > Treo 600 this holiday season!

Edumacate yourself about what has happened to MAJOR American companies like Microsoft, Chrysler, etc when trying to do business in China. These companies drool over what they see as a massive, untapped market. Only problem is, Western Rules Don't Apply. It's like the games of chance at a carnival. "Step right up folks, win yourself a piece of a BILLION person market!" Bend over and hope they're kind enough to at least use a little lube - things are gonna get rough.

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

I fail to understand your argument of how that MODEL has FAILED in China. GM, Microsoft, and all the other did not have that model as they conducted business in China.

Really? I thought PalmSource has ALREADY had PalmOS "stolen" from it. Something about a former licencee who's name begins with "A". Who was that again?

The only thing open source about a future PalmLinux might be the Linux-based kernel. Everything else that sits on top of that kernel (GUI, PACE, etc) is modular and proprietary. Pirates will take it all, thank you very much. Pirating PalmOS (any version) is as easy as pie. If you think otherwise, you evidently don't know what you're talking about.

TVoR

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

Surely this is nonsense. Open source means ANY competent coder can provide support. Close source means only people who can see and inspect the code can really understand the behaviour of their software.

Again, the original statement is nonsense.

Surur

A few particular corrections:

"recycling old Be designs"

There is pretty much no code in Cobalt that actually appeared in BeOS. The Be-related technology in Cobalt is mostly a set of UI and media frameworks we had been designing for a while before the Palm acquisition. These were intended to be a "next-generation BeOS", to address many of the design flaws in the BeOS frameworks that had become a particular problem when we did the stripped-down version for BeIA. As it turned out, a lot of the problems we were addressing for this new system design were directly applicable to a next-generation PalmOS, and so were used for Cobalt.

That said, the technology in Cobalt that was started at Be makes up a small (though in some ways fundamental) part of the whole system. Some parts of Cobalt were being designed by Palm when the acquisition happend (the kernel, IOS, and new Data Manager), some parts were implemented by Be and Palm engineers using the new Be frameworks (the Window Manager, Sound Manager, Notification Manager, Attention Manager), some parts were code that was ported from Garnet and then extended (basically all of the applications, the Form Manager and everything on top of it).

"insist on having Perfect Code"

One of TVoR's delusions. I have posted comments before of clear examples showing this is not the case.

"weren't able to pull off the dramatic PalmOS 6 transition on schedule"

Cobalt went GM 6 months later than originally scheduled. For the most part engineering didn't believe in the original schedule, but agreed on it with the condition that the resulting product at that point would never actually show up on a device. :) Ultimately, it was decided to push out the schedule so we would have a real GM.

Even so, being 6 months late creating a new operating system is hard to view as a failure.

"PalmOS 6 still isn't complete"

Which begs the question of what isn't complete about it.

"told PalmSource to go fcuk itself"

Think TVoR will tell us who his inside source is for this information?

"PalmSource finally pulled the plug on PalmOS 6"

Given that we are still too early in the process of working on "PalmOS for Linux" to completely know what it means, it is remarkable that TVoR seems to know better than we do. That said, in broad outlines it should be all of the higher-level system services (starting, for example, with the core frameworks from Be and things built on top of them) running on top of a Linux kernel. For the pieces between the kernel and user-space frameworks (Data Manager, IOS), it's still not clear exactly what they will look like.

This is hardly "pulling the plug" in Cobalt. Cobalt is our main product today, and as the announcement says it is not being cancelled. We fully expect a significant number of devices to ship with it. We continue, today, to be doing significant work on "Cobalt with the original kernel" for specific licensees and future major releases.

"picking pieces of Be code out of the PalmOS 6"

The funny thing about this is that the code that originally came from Be is probably the easiest to deal with -- it has already been ported from the BeOS kernel, which just so happens to be a Unix-like kernel not that different (on the outside) than Linux.

Another day, another port. :)

--
Dianne
PalmSource Application Frameworks Manager

I actually feel sorry for the Be engineers remaining at PalmSource. 10 years ago, they were part of a vibrant company that represented all that was good about the 90s tech craze. BeOS was in many ways far ahead of its time and could easily have claimed its place in history as the new Mac OS. Instead, greed got the best of Pépé Gassee and the company quickly fell on hard times, desperately prostituting its ideals to fetid c r a p like the eVilla just to stay alive.

It must have been painful to go from dreams of coding the ultimate desktop OS to cobbling together a lowly PDA OS. And now to be told that they will be replaced by a group of engineers from a Chinese Linux startup? Wow. That's gotta hurt. Paxil's going to be in short supply in Sunnyvale. Hope no one does anything drastic.

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