Reactions from the Mobile Tech Industry

Ah, but they are. Sure, they'll reduce it down to the core functionality that is needed, but Linux was designed as a desktop/server OS. There are much better designed operating system and kernels for mobile/embedded devices, including BeOS.

They are betting on the advanatages of using a more popular operating system with more developers who are familiar with its internals and the availability of more drivers for chipsets (often handled by the chip developers themselves). Oh, and it's free.

I think PalmSource is probably right with this decision, but don't let yourself believe that there are not better operating systems out there for this task.

Too busy getting actual work done on my Zod.

"And this coming from a guy who said Bluetooth has killed WiFi! Bwahahahahahaahah!!!"

Never said that. I said Bluetooth *and 3G cell data* would kill WiFi. A friend of mine is ditching his Tungsten C this weekend for a Bluetooth-enabled T5. He's tired of me being able to check my email and RSS where he can't.

As far as article comments go, I'd like to hear from some ex-Palm Inc/PalmOne/Handspring staffers. Are they predicting doom & gloom like the rest of the PIC faithful?

Rant:
Going back to the previous comments about the "Zen" of the older units. The other day I met with a guy who still kept a HS Edge in his briefcase. I briefly played with it. That could have been a phenomenal unit had it just had a bit more oomph under the hood and an SD slot instead of that stupid mini-Springboard. In fact, P1 should have immediately tried to shoehorn a 320*320 color screen, SD & BT into that fellow as soon as the Handspring acquisition became official. Now THAT'D be a proper T|E2!

It is BeOS that was designed, from the ground up, as a desktop operating system. Its tightly integrated design also meant that targeting it at a different device type was going to be a lot of work.

Linux, on the other hand, is widely used in embedded devices, cell phones, and handheld devices already. Linux implements the most widely used standards for embedded operating systems. Linux and its predecessors run on machines with a few hundred kilobytes of RAM. And Linux cleanly separates the GUI from the kernel, allowing Palm to do just what Palm wants to do.

"They are betting on the advanatages of using a more popular operating system with more developers who are familiar with its internals and the availability of more drivers for chipsets (often handled by the chip developers themselves)."

Yes, that's the point: Linux is already widely used in handheld and embedded applications, so it is already supported.

"I think PalmSource is probably right with this decision, but don't let yourself believe that there are not better operating systems out there for this task."

Maybe there are, but BeOS isn't one of them. BeOS was a bad design even for desktop use, and it made absolutely no sense to port it to handhelds.

Just so you're clear on this, BeOS was never designed to be a traditional desktop OS. When it was designed, desktop systems had the power of today's PDAs. It has a microkernel design and was specifically geared toward handling multimedia. The BeOS folks did years of work on using it for non-desktop systems.

Targeting BeOS to other device types was not a lot of work. Getting Linux to even be a reasonable possibility for embedded systems was. Go back and look at the history of it. The big difference was that BeOS was closed source and done by a handful of developers in a short period of time. Linux efforts are open and have far more effort and time behind them.

I don't specifically want to argue for BeOS here. I'm glad that we have Linux as an alternative to Windows, because it is clearly the winner there. The simple fact of the matter is that Linux won out with PalmSource because of cost and support, not because it was the best software solution (and that is not saying it is a bad solution) for the task. It was the best decision for a company that wants to make money.

Didn't people complain long and hard about win ce being a cut down windows system, and therefore a resource hog with little optimisation for handhelds? Who's playing the same game now?

Surur

First, you claimed that Linux was "designed as a desktop/server OS". That's simply wrong. Linux was designed to be a work-alike of UNIX, nothing more and nothing less. UNIX derivatives have been used for decades in embedded applications and have requirements that are a fraction of those for BeOS. Also, Linux is one of the most popular embedded operating systems and already powers a huge number of mobile and embedded devices; claiming that it's a "desktop/server OS" is just unreal.

Second, you claimed that BeOS is a "much better designed operating system for mobile/embedded devices". One will never be able to settle that argument definitively, but just because there is a zealous and vocal group of BeOS advocates out there doesn't make that true. To me, BeOS is a textbook example of poor engineering: it does what it does fairly well, but it does completely the wrong things.

I hope we will eventually get something better than Linux. I also hope its successor will look nothing like BeOS because BeOS got just about everything wrong.

The relationship between Linux and desktop software is no different from the relationship between Windows XP and the Windows GUI and between Macintosh and the Macintosh GUI: it's a kernel with a separate user-mode GUI environment, composed of a window server and lots of dynamically loaded toolkit libraries.

UNIX/Linux has had that architecture from the start, while Windows (3.1->NT) and Macintosh (OS9->OSX) basically had to start over to get there because there previous systems were architected so poorly.

and it translates very poorly into a desktop environment for the average user.

There is no basis for that claim; Linux desktop usability is comparable to that of Macintosh or Windows.

Cheaper, yes, more powerful - arguable. PCs starting pulling ahead of Macs in the power stakes towards the late 90s. Prior to that it was neck and neck.
Whether it's back to neck and neck now is debateable but I don't want to turn this (yet another) PC vs Mac argument - especially in a Palm OS forum!

>" Eventually, despite the Apple loyalists' protests that Macs were infinitely better in every way, more customers chose PCs, and it's been that way ever since. Apple got hit by a death spiral because they didn't diversify and compete, so they lost the market leadership position. Sound familiar? "<

Ya, ya ya. I think it's more a case of PC market grew much faster than Mac market and Apple NEARLY went into death spiral as a result of their mismanagement in the mid 90s. Yes, Apple has small slice of computer market BUT, unlike PalmOne and PalmSource, it's a PROFITABLE slice and APPLE has US$4 billion in the bank - that's what counts whatever business you're in.
Compare with what's happened to IBM, the inventor of the PC. It's PC division has been offloaded unto a Chinese company; it can be said that the PC business hasn't worked out for them. Unlike Apple, many PC compatible companies have come and gone.

The only company which has CONSISTENTLY gained the greatest benefit from the PC compatible market since its creation (both of market and company) is MICROSOFT.

Relating all of this back to PalmOS, it appears that PalmSource has struggled with OS 6, in the same way Apple struggled with Copland in the 90s and Microsoft seems to be struggling with Longhorn now. It also infers that Palm's BeOS aquisition was a bad move, why didn't they make the move to Linux then instead, now they're having to throw stuff out and start anew - just as Apple did when it threw out Copland, went for NextStep/Rhapsody and transformed it into OS X.

It also suggests this is a last chance move by PalmSource in it's battle to stay alive - I ask rhetorically just how many major licensees of the PalmOS remain now?

Linus wrote Linux because he wanted an OS that he could tinker around. That it has evolved into workstations and servers is just one market path. I've had a Sharp Zaurus for several years and the startup time and battery life are real pain points (and QTopia isn't so hot an environment, either, IMHO). If mLinux has solve those problems as they claim then they have a good jump on the Linux/Palm OS experience (well, at least the old Palm OS experience, maybe ;) ).

(Personally, while I think Palm OS on Linux is a good idea, the better aspect of the deal is the move into the China market. And I really don't see many vendors jumping on the Be-based Cobalt now - if I didn't have the device ready to ship at this point I'd at least want to know the timeframes for the Linux-based OS first, and I would probably wait for it. CMS also seems to support the same ARM vendors as PalmSource, so it's a very good fit from that standpoint.)

LMAO! You're kidding right? If you said over the past decade, you'd be pushing it. Decades??? Hilarious!

Yeah, you might of seen some "embedded" systems running Unix derivatives in the 1990's, but they were running on relatively high-end hardware for the time. Maybe even in the 80's but once again, it would've been on very high end hardware for the time. Hardly what I'd call an embedded system, and most Unix derivatives at the time didn't have much support for real-time stuff either.

Requirements-smirements ... who cares about requirements. Any decent OS today from a company other than Microsoft should be able to run circles on today's 100-400MHz PDA processors. I'm talking about pluggable micro-kernel achitectures, consideration in the design for low power requirements, real-time support, etc. Things you want for mobile/embedded devices. I can't think of one Unix-related OS that goes (or at least had its origins) in those directions. Sure Linux, with A LOT OF EFFORT, has gone in those directions, but just imagine that kind of effort behind an OS that was designed that way from the start.

> just because there is a zealous and vocal group of BeOS advocates
> out there doesn't make that true.

Well, I'm actually NOT one, but you probably think I am. I'd just like to see a mobile device with an OS that was designed for the task, and Linux isn't it. I hadn't had a chance to take a look at Cobalt yet, but I had been hoping PalmSource would use that opportunity to put together an OS designed right for handheld devices.

> To me, BeOS is a textbook example of poor engineering: it does
> what it does fairly well, but it does completely the wrong things.

Well, get your textbook out and educate us about what's wrong with BeOS (or Cobalt for that matter). When you keep telling me that Linux is better because everyone else is using it, your argument doesn't hold much weight.

Cobalt should have an option to inefficiently multitask two applications with notification of activity in a background process, so long as the user desires. This issue can be addressed by allowing the people to turn on and off multitasking in Preferences for individual applications. That would kill two birds with one stone - it would allow true (inefficient) multitasking to those of us who demand it, and turn the feature off for those of us who lack the system resources on their device. This, in my opinion, should be the types of things PalmSource should be investing R&D into - how to give customers what they demand.

-Bosco
NX80v + Wifi + BT + T637

LMAO! You're kidding right? If you said over the past decade, you'd be pushing it. Decades??? Hilarious!

UNIX has been used as an embedded OS for communication switches, network switches, high-end printers and copiers, industrial devices, and high-end machinery for more than 20 years. POSIX arrive in 1988 and was quickly adopted by embedded OS vendors.

Yes, and if you looked any further into it, QNX isn't a "UNIX derivative" despite the name.

And if you looked even a little further, you'd see that both QNX and Linux implement the UNIX APIs.

http://www.qnx.com/products/rtos/