Palm OS Cobalt Video Demo

I would have posted it two years ago, had this delivery technology been in place.

Mike, you can't stream/post a torrent file within an article. I realize there are plenty of other ways to disseminate a movie, but google video makes it pretty easy and I don't have to worry about bandwidth or people not having any additional software. It's a one click interface.

-Ryan

Everybody's a critic...

"I wish I better understood why it never made it to primetime. I mean, I know the PalmSource people were pretty much incompetent when it came to developement"

Cobalt never made it to primetime because PalmOne never adopted it. Palm had a bass ackward licensing scheme from day one that charges for upgrades. Hardware makers have an incentive to stay at their current OS because they have to pay huge $ to change.

The signs of problems where there when Handspring stayed at OS 3.1 long after OS 4.0 was released.

Palm Source *should* have had a more even subscription based model rather than having a huge lump sum when the OS changes. The idiots at PalmSource management were basing their projections on a huge windfall that never came and the idiots at PalmOne secretly decided to avoid paying.

PalmSource development wasn't incompetent on the technical side and Cobalt was/is a very solid OS. Palm should have never split into two halves and they shouldn't have split with an infeasible licensing structure.

Cobalt has all of the good things that Garnet had, plus it fixes many of the short comings and errors. A lot of money was spent on a great OS that was never used.

I mean, I know the PalmSource people were pretty much incompetent when it came to developement after they were spun off from Palm.


Umm, no. Quite the contrary, as you can see in the video, they knew their stuff. The issue is that in every software products life there comes a time where you have to go back to go forward. This is very painful from a business standpoint as it takes time and resources. OS X was Apple's pain, Vista is Microsoft's, FireFox is the result of people stopping and realizing that at some point, crufty code has to go. PalmSource broke off from Palm as they realized this and wanted to go down the path of working towards the future. PalmOne just wants to sell devices. They know they can keep selling Garnet devices and people will buy them so why pay extra for Cobalt? Now granted, they finally hit a bump in the road and now we get WinCE devices. Why? Because Palm just wants to sell devices.


It was a catch 22 from the start. Without breaking off, Cobalt would never have happened... But because of the break off, there was no reason for Palm to put forth any effort to upgrade. That and there is the We-Don't-Want-To-Be-The-First mentality. Why, when the ARM architecture could go big or little endian did Palm switch to little? They didn't want to be the first. I was at that PalmSource (oddly enough on the same end of the room as the video) and the *big* event at that show was Sony. The room was packed (standing room only) as they pulled out like 6 new devices. The PalmOne presentation was a ghost town. Not even a product announcement. Nothing. We all thought that Sony was going to be the first to do Cobalt (the pinlets where made for allowing of entering of Asian characters) Then of course they dropped out.

PalmSource's mistake was going that far with no commitment from *anyone* to release a device. PalmOne's mistake has been not looking at Cobalt as a path forward (how many Treo issues are directly related to the fact that they are *still* patching Garnet?) Right now I see Access in a better spot than Palm. Access has the Asian market and embedded devices ahead of them. Palm is about to have it's ass handed to it in the WinCE market with no clear direction in the only market that they had the lead in, PalmOS devices. Time will tell. GSPDA seems to be a good match for Access (finally). Prior to the buyout, it looked like some kind of death spiral with PalmSource having no one to buy and PalmOne having nowhere to go. Things are just starting to get interesting now.

"Without breaking off, Cobalt would never have happened... But because of the break off, there was no reason for Palm to put forth any effort to upgrade."

Cobalt was conceived long before Palm split. Actually OS5 was a temporary "half-way" solution from the beginning. Apps didn't run fully native and obviously OS5 was never intended as a long-term solution. That would be like Apple's Rosetta or Blue Box becoming permanent. OS5 was supposed to be a temporary patch!

If Palm hadn't broken apart, a native cleaner OS would still have happened.

WareWolf is completely right that most of the Treo instabilities are due to an OS that is stressed and strained to the breaking point held together with fraying strands of duct tape. PalmOne solves the problem by sticking with Garnet and hiding/downplaying problems in the media. --Pathetic

Playing around with WinMobile phones is a great way of diverting media attention away from Palm Treo instabilities.

Personally I think the PalmOne engineers have done a excellent job, but management has forced them to use an OS far beyond its original design. Some of the Garnet basic architecture is little changed from the US Robotics Pilot developed in the 1990's.

Surur, I had the same thoughts when watching this video. Very reminiscent of the BeOS multimedia-intensive demos. Sad, too, because BeOS was able to eek out every bit speed from the CPU and graphic chips and this Cobalt demo looked to do the same. As you've noted, BeOS's main obstacle was their inability to get driver support, and that's what we've heard as a big reason for why PalmSource made the switch to Linux. But, to be quite honest, I don't see the two as being analogous.

First, BeOS was always a small player trying to make it big. At the time PalmSource had Cobalt ready to go (supposedly), they were in a fairly commanding position.

Second, driver support for a PC is critical. Driver support for a PDA or smartphone? Don't they pretty much need to write those drivers from scratch, anyway? And even for things like 3rd party Wi-Fi SD card support, it's not like there's fifty different manufacturers of SDIO Wi-Fi cards. At most, you're talking about four or five. So working to get working drivers for those with a non-universal OS shouldn't be that big of an undertaking.

I would contend that PalmSource made it's own grave. They were too fat and too lazy. They let/made their licensees do too much heavy lifting while they squandered away the big money they were pulling in from licensing fees on executive salaries, huge buildings, and who knows what else.

I've never been a big fan of this shift to Linux. The original Palm OS was designed from the ground up to run with a small kernel and be optimized for a PDA. Adapting a server OS to a PDA just because Linux is the current "hot buzzword"? Not a plan for success, IMO. I really would like to know what (if anything) was so bad about Cobalt such that they never got any licensees to develop products using it.

http://Tapland.com
- Tapwave Zodiac News, Reviews, & Discussion -

TVoR knows what was/is wrong with it but he/she/it ain't telling...

Actually I know too, but I'm not telling either.
(See how easy it was to make that claim?)
Now if someone really does know something more specific than we already know about the "why" of the Cobalt no-show, it would be fun to hear. Otherwise it's just a big tease, and I'm not interested in playing that game.

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

twrock;

You must not have noticed the sarcasm dripping from my earlier post. Anyone who is 'in the know' is either currently employed by Palm/Access/PalmSource and/or tied down under a NDA or simply doesn't care enough about the rabble rousers on PIC to spill the beans.

It's possible we may not know The Ugly Truth for another couple years. How long did it take for the inside scoop on Copland to come out of the Apple camp back in the late 90s?

Pilot 1000-->Pilot 5000-->PalmPilot Pro-->IIIe-->Vx-->m505-->T|T-->T|T2-->T|C-->T|T3-->T|T5-->TX

...the people who decided it was a good idea to adopt BeOS for a handheld OS.

Who'd that be? Not us, certainly.

Since I don't expect Dianne to offer (yet another) correction to this persistent misrepresentation here, let me act as Rumor Control and offer the facts:

- BeOS was neither "adopted" nor adapted by Palm/PalmSource. We did purchase Be's intellectual property and hire a number of their (ex-)engineers.
- No substantial amount of BeOS code was taken wholesale into Cobalt.
- For starters, Cobalt ran on a kernel developed by the Palm kernel team, and not on the BeOS kernel. If, as you seem to be suggesting, BeOS was somehow simply wrestled into a handheld version and retitled "Cobalt", that'd appear a fairly odd oversight...

Actually the media player in the Cobolt sim DOES work. You have to create another folder in the directory that poses as expansion media, right where the 'Audio' directory is. Create a folder named 'Video' (i just guessed and it worked). It's been over a year or so since trying that so i don't remember the supported codecs - just mpeg2 or 4 i think?? The trully cool bit that we all should be crying about was the API for the media engine. The player is just a simple little player written using that engine. Every programmer would have been able to easily add video and audio to their apps. check the cobolt API docs.

>>What this means is that the coolest programs for Cobalt (The media player, graphics intensive stuff) cannot run because the ARM processor isnt emulated in the sim

Remember also that 'the cool' stuff in the video that Ryan posted above, is all running on the Cobolt Simulator.

Just another point - as understand it, none of the stuff you see in Cobolt simulator that is core OS stuff (UI, the underlying functionality of the built-in apps / the apps them sleves) are 68k. It is all Cobolt APIs: ARM on a device, x86 dlls on windows.