Here's the primary flaw in the arguments presented by that white paper:
Perhaps the consortium believed that next generation mobile handsets would have greater application requirements and need the support of third party developers like desktop programmers. In this case, a good API and a third-party developer program would naturally have been attractive. Unfortunately, the need for third party developers is limited in next generation devices, because the mobile Internet business model depends on content and connectivity via technologies such as WAP and K-Java, not proprietary third-party applications. The Microsoft and Netscape browser war made this point clear.
Wind's view of the operating systems on 3G phones is that they'll be nothing more than thin-client browsers with some sparse, pre-installed software by which one can access network apps. So they think that they don't need a hige third-party developer base, they don't need apps that can be bought off store shelves, and so on. This stance as just as flawed as Larry Ellison's "the network is the computer" vision.
Everyone should take a look at the 3G prototype phones displayed on the following page:
gsmworld.com
There's obviously plenty of room for a highly differentiated user interface. There's plenty of means by which non-network-based apps can be installed onto such devices the same way they're installed onto Palms, Psions, and Pocket PCs. I'm not just talking about standard organizer apps, but also multimedia plug-ins, video games, scaled-down business apps, and so on. Much more than just your run-off-the-mill web browser.
With this kept in mind, Symbian's going to have two huge advantages against Wind: the ability to support applications that are created by a third-party developer base numbering in the tens of thousands, and which can be purchased off of retail stores as well as "borrowed" from friends (another problem with a network-based approach...people love pirated software) as compared to a few sparse network-based apps on a VxWorks device, and the ability to create a virtuous eBay-style network effect given its widespread support (more phones supporting Epoc leads to more Epoc-supporting programs which leads to more user popularity which leads to more handset support which leads to...).
What OEMs need is not an organizer operating system, but a core real-time foundation and solid development tools so that they and their third-party partners can build next generation devices – a whole range of them – quickly and easily. A Java virtual machine on these devices will allow third party developers to create a wealth of mobile applications.
I think the last five years have shown how over-hyped Java really was. Sure, it's great for allowing internet applets to function regardless of the computing platform of the end-user, but for standard apps, programmers found it to be vastly inferior to OS-dependent languages such as Pearl and C++ in terms of flexibility and robustness. Once again, this goes back to the fallacies surrounding Wind River's "the network is everything" view of things. Besides, Symbian has Java support covered anyway:
techweb.com
So in conclusion, I find it difficult to believe that Symbian is just "a marketing campaign to hype market awareness." The huge investments made in Symbian by the world's leading handset manufacturers, the 3G prototypes that are meant to utilize the operating system, the demos that have been shown of Epoc-based handheld devices, the Epoc-based product release timelines given by Nokia, Ericsson, Sony, and others, the current widespread use of Epoc in Europe, and the overwhelming amount of support and interest in Epoc shown by third-party developers at CeBIT make it look like much more than hype; and as the failure of the network computer's shown, attempts to go against a client-centric platform with widespread application support by means of a network-based architecture that denies end-user flexibility tend not to meet with much success.
Eric |
