I didn’t really know where to put the final two sections of this article, so I decided to create a small miscellaneous chapter specifically for them. In this chapter, I’ll first detail the importance of Palm OS licensing, after which I’ll dive into the future of Palm OS, as seen from 2004 (which means Cobalt).
Licensing Palm OS
While often overlooked, you can’t really write a detailed article about Palm and the Palm OS without diving into the many companies that licensed the Palm OS for use on their own devices. Sure, Palm itself commanded the vast majority of the Palm OS market, but the licensees played a crucial role in the evolving ecosystem: they became test beds for new features, technologies, and ideas.
From very early on in its existence, Palm intended to license Palm OS to third parties. Instead of scavenging around trying to scrounge up a licensee or two, like so many modern platforms have to do, Palm scored big right at the very beginning: five months after the original Palm Pilot started shipping, they scored IBM. The industry giant was going to sell rebranded Pilots, which eventually arrived on the market about a year later in September 1997 as the IBM Workpad 10u. IBM continued to rebadge Palm OS devices until 2001.
Vertical markets were also interested in Palm OS. And so, in March 1998, Symbol launched the SPT1500, a Palm III-like device with a larger forehead to accommodate a SE 900 Pico Scan Engine (apparently a revolutionary small barcode scanner). You could also get a version of the SPT1500 with wifi built-in (no joke – it had spectrum24). As far as I can tell, this makes it the first Palm device with wifi built-in.
From 1999 and onwards, there’s an explosion in Palm OS licensees, many of which would play crucial roles in the on-going development of the platform and its supporting ecosystem. HandEra (formerly TRG) launched the TRGpro in October 1999, first to introduce a Compact Flash slot. In 2001, HandEra would launch the HandEra 330 – not only was this one of the first (not the first, though, but we’ll get to that one in a second) devices with a ‘high’ resolution display (320Ã—240 instead of the usual 160Ã—160), it also introduced a feature that would become standard on all higher resolution Palm devices: a virtual Graffiti input area.
October 1999 also saw the introduction of the first ever Palm OS device with mobile phone technology built-in, making it one of the very first smartphones on the market: the Qualcomm pdQ. Hardware-wise it was identical to the Palm IIIe, but came with CDMA technology so you could call and even browse the web and send and receive email. Weirdly enough, the phone and Palm OS features were separated; the phone functionality ran its own firmware, and communication between that firmware and the Palm OS was limited. Later versions, after Qualcomm was acquired by Kyocera, improved this integration substantially.
Other than loads of Palm OS devices from smaller, lesser-known companies (my favourites: the AlphaSmart Dana and the Tapwave Zodiac), several big names would become Palm OS licensees: Nokia (never released a device as far as I know, but there’s this), Samsung, Acer, Fossil (for smartwatches – currently all the rage again), Garmin, and Lenovo. As interesting as these were, they were small-time compared to the two big ones: Sony and Handspring.
Sony became a Palm OS licensee in 1999, and unlike some of the other licensees, it really tried to push the envelope for the Palm OS platform with its line of CLIÃ‰ devices. While the first CLIÃ‰ – the PEG-S300 – didn’t exactly set the world on fire in 2000, it did have a unique design aesthetic that broke the mold a bit, and it had a more expensive version with a colour screen, too. This wasn’t the first Palm OS device with a colour screen, but it was among the first.
After getting their feet wet with those first few Palm OS devices, Sony really started to drag the Palm OS platform forward, kicking and screaming. Released in March 2001, the CLIÃ‰ N700C was the first Palm OS device to sport a “high” resolution display, and it did so in a way that didn’t break application compatibility: it simply doubled all the pixels from the standard Palm OS resolution of 160Ã—160 to 320Ã—320 (all that is old is new again). It wasn’t until November 2002 that Palm itself released a device with a 320Ã—320 resolution, the Palm Tungsten T.
Even before Palm adopted the 320Ã—320 resolution for its higher-end devices, Sony had already stepped up the game with what I think is the most iconic of all CLIÃ‰ devices: the CLIÃ‰ NR70. The NR70 introduced the characteristic flip-and-twist design with a full QWERTY keyboard; you could flip the screen open like on a clamshell phone, and you could then twist the display around and close it back up to arrive at a more conventional form factor. It sported a resolution of 320Ã—480, although applications had to be designed to support it (Palm wouldn’t adopt this resolution until two years later (!) with the slider Tungsten T3).
The NR70 had a more expensive brother: the NR70V. The V-model added yet another first for the Palm OS platform: an integrated digital camera with 0.1 megapixel. Sony also released the somewhat odd-looking PEGA-MSC1, a digital camera in the form of a Memory Stick which could shoot pictures at a maximum resolution of 320Ã—240. You could use this with select CLIÃ‰s with Memory Stick slots. Just to illustrate how far ahead of the curve Sony was: only two years later did Palm release its first device with a camera (the Zire 71).
I’ve always wanted a CLIÃ‰ back when they were still current, but they were quite hard to come by where I lived. A few years ago I bought two CLIÃ‰s from one of our French readers, one of which is the PEG-NX80V, one of the later flip-and-twist models. Its camera no longer works and the display has odd discolouring issues (I intend to open it up to see if I can fix it), but the whole form factor is quite pleasant. The keyboard is rubbish (impossible to properly type on), but twisting and turning the display is very satisfying. It’s sad the mobile world has pretty much converged on boring slabs – I’d love Sony to release a modern, dual-screen interpretation of the flip-and-twist design.
Other than these specific firsts that Sony brought to the Palm OS ecosystem, a big focus of the company was multimedia – imaging, video, ATRAC and MP3 playback with remote control-equipped headphones (my NX80V has a stylus tip built into the remote!), and so on. This wasn’t really a focus for Palm, but Sony showed them how it was done. In fact, Sony was so far ahead by this point that not Palm, but Sony was the first to launch a Palm OS 5 device (and, thus the first ARM device). Heck, Sony even went so far as to develop its own unique ARM processor for CLIÃ‰ devices.
Before announcing the end of the CLIÃ‰ line for European and North-American markets in June 2004, Sony had one last crazy trick up its sleeve for the world: the CLIÃ‰ PEG-UX50. The UX50 (and its cheaper sibling, the UX40) was a really tiny Palm OS laptop with a twisting display. Pretty it was not, but it more than made up for it with its sheer awesomeness. I’d love to have one of these.
Japan got to enjoy the CLIÃ‰ line for a year longer still, but in early 2005 the curtain closed in Japan, too. However, Sony just couldn’t resist going nuts one more time, and in late 2004, it released a… Palm OS tablet: the PEG-VZ90. This crazy-looking contraption sported an OLED display, and has so much ‘want’ written all over it that it was only available in Japan. I’m fairly certain when I say none of us have ever seen one of these exotic beasts in real life (yet another reason to visit Japan).
And with that, the CLÃ‰ line met its end. Sony had made significant contributions to the Palm OS ecosystem, and its focus on improving the multimedia aspects of the platform benefitted all Palm OS users. Their sales never even came close to that of Palm itself, but they left a definitive mark on the Palm industry.
Handspring is an entirely different animal than Sony or any of the other Palm OS licensees. The company has a very unique history in that it was founded by none other than Jeff Hawkins himself. In 1998, Hawkins and Donna Dubinsky had become unhappy with where 3Com – then owner of Palm – was taking the platform, since they realised that the PDA business wasn’t going to stay around forever. “The handheld computing business is going to go away,” Hawkins said at the time, “It’s going to become part of the cell phone business.” Hawkins was always one step ahead of everyone else.
So, Hawkins and Dubinsky left Palm, and founded a new company called Handspring. They licensed the Palm OS right away, and one year later, September 1999, it released its first two PDAs: the Visor and Visor Deluxe. Handspring’s first distinctive feature was the Springboard expansion slot; a proprietary slot that could house just about anything from memory cards to cameras, GPS modules, cellular modules, MP3 players, wifi, BlueTooth, barcode scanners, games, and much, much more. The importance of the Springboard slot’s ability to house mobile phone technology is evidenced by the fact that all Visor PDAs were equipped with a microphone specifically for phone expansions.
The craziest accessory every to be made for Springboard? I doubt you want to know, so skip the next quote box if you don’t.
Raynet Technologies unveiled today the world’s first body massager Springboard module for the Handspring Visor. Designed by Singapore Corporation, Raynet Technologies Pte Ltd, the Raycom Personal Massager converts the Handspring Visor handheld into a massaging device, while maintaining other computing operations.
It’s even been reviewed. I didn’t bother to read it, but here you go.
Handspring continued to build and sell the Visor line until it was discontinued in 2002, because the company had something far more special up its sleeve: the Handspring Treo. The Treo 180 was the first of the line, and came in a version with a QWERTY keyboard and a version with a Graffiti input area (the Treo 180g). Handspring released several more models before it was acquired… By Palm.
Following the Merger, Palm would continue to release several new and updated Treo models. They were quite popular – in fact, in the US, Palm enjoyed a smartphone market share in late 2005 of 31%, second only to RIM’s 55%. As far as I can tell, the Palm Centro, released in late 2007, is the last Palm OS device Palm ever made – before switching to webOS. It ran Palm OS 5.4.9.
Handspring innovated both in and outside the Palm OS ecosystem with the versatile expansion slot and, more importantly, with the Treo. Thanks to the Treo and its success, Palm was given a crucial stay on the way to its demise, which gave it just enough time to develop the next chapter in the company’s history – but that’s a story that has already been told.
Where to go from here (if here is 2004): Palm OS 6 ‘Cobalt’
Before I let you go, there’s one last chapter to this story before we wrap things up. This won’t be a happy chapter, and it won’t be about remarkable aspects of the Palm OS you didn’t know about, nor will it be about cool technology from the ’50s. With this last chapter, I won’t be able to strike the string of nostalgia either – because we’re going to talk about something that has never been released.
For those of you who’ve seen Fucking Ã…mÃ¥l (and for those of you who haven’t: shame on you), remember that gut-wrenching scene that dragged on – intentionally – for what seemed like bloody hours where nobody showed up for Agnes’ party? Where director Lukas Moodysson forced you to watch the naÃ¯ve hopefulness of Agnes’ mother as a stark contrast to Agnes’ depressed anger?
Well, Cobalt is like Agnes. Nobody showed up for Cobalt’s party either.
So far, I’ve been largely successful in ignoring the ownership situation of Palm OS, but for this last chapter I can no longer get around it. I’m not going into too many details, so here’s the Cliff’s Notes: Palm thought it was a great idea to create a wholly owned subsidiary dedicated to the development of Palm OS, called PalmSource, in 2002. They thought they had an even greater idea when they spun PalmSource off as an independent company in 2003. PalmSource was then acquired by ACCESS in September 2005.
Before being acquired by ACCESS, PalmSource worked on the future of Palm OS: Palm OS 6, better known as Cobalt. PalmSource took this matter quite seriously, and went for an almost ground-up rewrite; they stated around 80% of the code was rewritten. Cobalt was meant to bring the Palm OS into the 21st century, and included many features that had long eluded the platform. Sadly, since not a single licensee adopted Cobalt, information on it is patchy, at best.
First and foremost, just as Jim Schram had already predicted way back in 2002 when talking about the then-new MCK kernel in Palm OS 5, Cobalt finally exposed the system’s inherent ability to multitask to application developers. It was a fully ARM-native 32bit operating system with multitasking and multithreading, so multiple applications and services could run at the same time. It had a background processing model designed to “reduce most memory problems commonly associated with multitasking in mobile devices”.
Related to this, Cobalt also introduced protected memory, so that a crashing application would no longer be able to drag down the entire system. In addition, the maximum RAM and ROM limits were both raised to 256 MB, with the ability to raise this even higher in the future. Loads of security features (like industry-standard encryption stuff for enterprises) were added as well as core parts of the operating system, and it had a brand-new communications framework based on STREAMS, which supported multiple concurrent communication sessions.
A particularly interesting aspect of Cobalt was its new, extensible multimedia framework. As most of you will know all too well, Palm bought Be, Inc. and the BeOS in 2001 (yes, OSNews dates back that far!), and since BeOS is software, it ended up at PalmSource after it was spun off. Cobalt’s multimedia framework and its “rich graphics and multimedia features” came from BeOS, and may even have been designed and written by former Be engineers (I can’t find confirmation that actual Be engineers worked on it, though).
Coded by our Be superheroes or not, the BeOS-derived framework brought Palm OS up to speed with the rest of the industry, and provided it with, among other things, rich graphics support with “paths, rotation, gradient fills, anti-aliasing, outline fonts and transparency”, and supported screen sizes of up to 32000Ã—32000 pixels (!). It also supported various audio and video formats, and developers and licensees (ha!) could easily add support for more.
Cobalt included an emulator so it could run applications written for Palm OS 5 and earlier. However, only “well-behaved” applications would run without modifications, while more complex ones would need modifications. Supposedly, Astraware’s David Oakley ‘ported’ Bejeweled to Cobalt in less than half a day. Of course, developers could also write applications specifically for Cobalt, which were called Palm OS Protein applications. Interestingly enough, these could apparently be compiled for both ARM and… x86. I have no idea if ACCESS had plans for an x86 Palm OS.
On 10 February 2004, Cobalt was released unto the world – however, nobody, not even Palm, licensed it, preferring to ship Palm OS 5.4.9 instead. So, PalmSource went back to work and released Cobalt 6.1 in September 2004, which added an “enhanced” user interface (it looks exactly the same as before, just with slightly differently-designed buttons), and integrated a few things that were optional before (like BlueTooth and wifi). Sadly, nobody wanted Cobalt 6.1 either.
Since Cobalt never made it to market, there’s not a whole lot to add here – I can’t refer to reviews of devices, I can’t talk about personal experiences, nothing. All we have is a preview by Brighthand’s Ed Hardy, who notes, among other things, that yes, the BeOS-inspired multimedia framework could display a movie playing on all sides of a spinning cube. Which is very, very useful in real life computing scenarios both mobile and desktop, and don’t you dare tell me otherwise.
It’s worth noting that there is actually a freely available Palm OS Cobalt emulator that was provided by ACCESS. It’s still available, and runs on Windows 8 just fine. It runs the full Cobalt operating system in a window, and you can change several settings like resolution, memory, and so on. It’s pretty fast, but looks dated (despite the updated UI), and it misses several key applications (like the browser or an email client). Still, because you’re controlling it with a mouse and there’s little you can actually do with it, it can’t possibly form the basis of something resembling a review.
And this is where Cobalt’s story ended – or so I thought. I always thought that not a single Cobalt device was ever made – but during the long process of planning and writing this article, I stumbled upon an article by Bob Russell at mobileread, which confirms that Cobalt devices certainly did exist. The report is corroborated by David Beers, a Palm OS developer who actually handled one of the prototype devices.
A company called Oswin Technology, based in Singapore, designed and built two devices running Cobalt in 2005 that were offered for sale at the last PalmSource DevCon before ACCESS acquired PalmSource. Russell even provides three full-page scans of the advertisements for the devices, which list all the specifications and two beautiful photographs of the devices. One was a candybar phone, the other a slider (?), but otherwise had nearly identical specifications. They sported a FreeScale i.MX21 processor, and 64MB RAM/128 MB ROM.
They aren’t exactly beautiful or anything, but they are – as far as I know – the only complete Cobalt devices ever made. I wonder how many were sold at the time; I guess it depends on their price, and on just how many people were at that specific PalmSource DevCon. I wouldn’t peg the number at much higher than a few dozen, making this one of the most exotic Palm OS devices ever made – perhaps the most exotic. Interestingly enough, it seems that one of the two devices did make its way to market as the Zircon Axia A108 – running Windows CE.
The elephant in the room right now is of course why, exactly, Cobalt failed to attract licensees. I think the answer to this question doesn’t really have anything to do with Cobalt itself, but more with the storm that was brewing in Cupertino. By this point, all traditional Palm OS licensees had left the platform behind, except for Palm itself (and even they defiled their heritage by selling several Windows Mobile Treos), meaning that unless PalmSource could attract new licensees, it would have to rely on Palm.
However, I don’t think it’s far-fetched to believe that because several of Palm’s employees had ties to Apple, they were aware – to a certain degree – of the work Apple was doing on the iPhone, and that Cobalt would not be competitive enough. In other words, no matter how much of an improvement Cobalt was compared to Palm OS 5.x, Palm didn’t believe it was good enough to take on Apple’s upcoming iPhone.
And from what I’ve seen of Cobalt, they were right.
While Cobalt’s story ends here, ACCESS didn’t stop trying to squeeze more blood from the dry, withered stone that is Palm OS. Aside from Palm OS 6.x, PalmSource/ACCESS had also developed Palm OS 5.5, which was designed to run in a dedicated virtual machine so other possible smartphone platforms could run Palm OS applications (it made its way to Nokia’s internet tablets). Like Cobalt, nobody was really interested in this thing either.
However, this product, later renamed to the Garnet VM, was to play a crucial role in ACCESS’ next attempt at gaining traction in the smartphone market: the ACCESS Linux Platform. As its name implies, ALP was a Linux environment based around Gtk+ and GNOME technologies, which could also run mobile Java applications. On top of that, Garnet VM was built-in to provide compatibility with applications for Palm OS 5.x and earlier. Its user interface was essentially a slightly more modern-looking Gtk+ recreation of the Palm OS graphical user interface (although the last version, from 2009, doesn’t have much in common with Palm OS any more).
ACCESS failed to find any licensees for ALP. Its website stands as a 404-ridden tomb of broken dreams and unfulfilled aspirations.
As a sort-of conclusion to the Palm OS and Cobalt chapters, here’s a short overview of the major Palm OS releases.
- Palm OS 1.0: 1996
- Palm OS 2.0: 1997
- Palm OS 3.0: 1998
- Palm OS 3.5: 2000 (a point release, but a major one)
- Palm OS 4.0: 2001
- Palm OS 5.0: 2002
- Palm OS 6.0: 2004
- Palm OS 6.1: 2004 (September – the final Palm OS release)
And that’s all she wrote.