I'm not sure I totally agree with Adrian Kingsley-Hughes' blog post assumption that the nice homogeneous iPhone environment will be less secure (on Stevie-boy's 10 million unit estimate) than the security you get with Java ME technology having security through obscurity (of the WORA promise on cell phones).

See:

Is security an effect of the sucky WORA of Java ME???

Here's a quote:

 The fact is that it’s tough enough 
 to write legitimate mobile 
 applications using something like 
 J2ME (Java2 Micro Edition) that’ll 
 work smoothly across a small number
 of phones. The combination of an 
 abundant lack of standards and the 
 number of companies competing 
 aggressively means that no single 
 platform has managed to capture 
 enough users to create a critical 
 mass.

If virus writers can write a virus they will. But, put up enough security hurdles to make the virus writer's life difficult (while keeping legit MIDlet writers like Google, Opera, Digital Chocolate, JAMdat, etc. lives (hopefully) less difficult), results in what we see now, not sucky WORA. Virus writers do what every good engineer does: cost/benefit analysis. If it looks too costly (with too many security hurdles), they move on. You don't try breaking into a house with bars on the window, an ADP sign on the front lawn, and a German Sheppard named Bruno on the porch.

Hinkmond