Now blogging at diego's weblog. See you over there!

the mobile platform wars


So it only took a couple of days since Motorola bailed out of Symbian for a rumor to surface that they would be releasing a "Microsoft-powered" phone, sold through Orange, later this month. Recently there was another rumor (or news? I can't find a link) that Psion was moving to WinCE for its mobile devices. In the meantime, Linux is making inroads into all sorts of devices. Oh, and, by the way, PalmOS is down but not out yet.

See a pattern emerging here?

Call it the mobile platform wars. In which Symbian has the tactical advantage, but is, strategically, in an entirely different position.

Symbian, by design, allows its licensees to tailor their offering heavily for different devices, which is great for licensees short-term (allows them to obtain early lock-in on features), but not so great for Symbian as a long-term platform. Long term, a platform cannot survive like that, and by extension neither can its licensees. The platform splinters irreversibly, because even though the licensees achieve short-term early lock-in on features, the platform itself has no lock-in.

Symbian won over Palm through faster innovation and larger deployments, but now Symbian is the incumbent, and the game is different. The new entrants are not competing on features, but on platform homogeneity.

It has happened before: Look at UNIX in the 80s.

In theory, you could port applications between UNIX OSes by sharing more than 80% of the code between them through a "standard" called POSIX.

In practice, almost no one did it.

And so the POSIX-UNIX "standard" allowed itself to be overtaken by both Linux and NT. Because once platforms are established, it all goes back to third-party developer support. Why? Because users care about applications and devices, not about OSes. They don't care if, say, the memory space is 32-bit flat. Developers do. Which is why third-party development should be active and growing for any long-term platform.

Which requires a vibrant community. Which requires tinkerers and small developers, as well as big developers. Which requires simplicity and portability within the platform, and a low-cost of entry (read: free, well-documented, well-supported, entry-level tools).

And is it a coincidence that, again, both a Windows variant and Linux are emerging as the greatest threat to an innovative platform? I don't think so.

Netscape, by the way, made similar mistakes with regards to developers. And they played a small but important role in their fall from grace.

But Symbian is improving, and listening. It isn't over. Yet.

Even if it comes to the worst, I can't see Symbian ceasing to exist. When you're talking about millions of devices sheer volume wins, so it's quite possible that in one or two years Nokia will just end up owning Symbian outright and it just will be Nokia vs. Microsoft.

But since Nokia is one platform and one vendor, Java would be a better choice.

Paradoxically, excellent Java support might allow Symbian to prosper by providing the best Java mobile platform around. Palm might yet come around as well and figure out that Java is their best weapon to fight the Microsoft/Linux juggernaut.

And in that case, once again, the only thing standing between us and yet another monoculture would be that sweet smell of digital coffee.

Categories: technology
Posted by diego on September 5 2003 at 12:11 PM

Copyright © Diego Doval 2002-2011.
Powered by
Movable Type 4.37