__/ [ Daniel James ] on Monday 08 May 2006 11:26 \__
> In article news:<slrne5mp5r.i0k.sorceror@xxxxxxxxxxxxxxxxxxxxx>, Ray Ingles
>> The PalmOS was nice for its time, but it's just not up to multitasking
>> and such. You could shoehorn cooperative multitasking in without too
>> much trouble but it's just not the same.
> IIRC PalmOS is built on top of a multitasking real-time OS that Palm
> licensed from another company (KADAK?), but PalmOS itself doesn't make use
> of the multitasking features of the underlying kernel. There may have been
> a licensing issue here?
Several years ago, concerns were raised not only about the inability (or
unwillingness) to support multi-tasking. Another principal concern has
been (and still is) unicode and internationalisation. I wonder if there is
a backward compatibility trap and I very much suspect so. You did allude
to this yourself. Not only might cross-device beaming break, but also
third-party code. Moreover, think of issues like packet size, clipboard
stack/size, memo pad limits and so forth.
> What puzzles me is that - now that PDA technology has moved on and
> multitasking is recognized as an important feature for a PDA OS - PalmOS
> still does not do preemptive multitasking. Surely Palm would have altered
> the PalmOS front-end code to use those multitasking features by now if
> there weren't something deeply embedded in the PalmOS design that made that
> hard to do? The underlying RTOS apparently already supports it, so the
> problem must be in the GUI layer ... or possibly in the difficulty of
> maintaining compatibility with the huge base of working 3rd-party
We bagan a certain elaborate discussion on this as soon as Palm announced
their move the the Linux kernel in an official way (they exclaimed the
intent early last year, if I recall correctly). One of the certain
speculation was that Palm OS applications will run on an emulation-type
layer. I can't remember the technical jargon and acronyms involved, but I
can track that long thread if you are interested.
>> What I'd do is pretty much what they are already doing - build a Palm
>> emulator on top of Linux and add some hooks so you can call back and
>> forth. The Palm stuff as such will gradually fade away.
> Putting the PalmOS front-end on top of linux is probably a great idea if
> you want to be able to license PalmOS as a front-end to PDA/smartphone
> makers who already have linux running on their hardware. It is probably not
> going to make it any easier to add multitasking to the GUI layer.
Interesting points. I have always thought of the move to the Linux kernel
as a gate to better interoperability with more devices and applications.
In fact, studies have foreseen a future where only Windows- and
Linux-based devices prevail (so long Symbian and other small vendors). I
think your point affirms such an analysis.
Roy S. Schestowitz
http://Schestowitz.com | GNU/Linux ¦ PGP-Key: 0x74572E8E
1:10pm up 10 days 20:07, 7 users, load average: 0.37, 0.36, 0.60
http://iuron.com - next generation of search paradigms