Julian and William are discussing J2ME on Ecademy - apparently the games industry have gotten fed up and are going to sort it all out for us. I'd reply there if I could, but haven't been able to post into the forums for a few months now... so here we go:
"My understanding is that this will be a NATIVE game architecture, that is compiled code not a VM. The fragmentation not only increases costs, it reduces innovation. The performance penalities of running game code in a VM is another issue that is being addressed here."
Well, if they fix it tomorrow - which they won't - we'll have 3-5 years before whatever magical pixie dust they eventually release gets out there in large enough numbers to be worth bothering with. If this is a problem for you, either get out of mobile for a few years, or learn to deal with what we've got today (that's rhetorical - I realise that you're doing the latter William ;)).
Julian writes: "Less cynically, a common development platform, and a more open environment would reap huge dividends for everyone."
Actually, I think fragmentation creates opportunities. Look at the number of independent J2ME games developers, and compare to the number of small independent Nintendo DS developers out there. Given the disparity in screen sizes, keypad layouts, and processor performance, "write once run anywhere" was always optimistic. It should be, it was a marketing slogan after all.
Personally - I think it's easy to underrate the position we're in today. It is possible to develop, test, and launch a new game across 80 handsets in about a month. That strikes me as a massive approvement on the situation 5 years ago, when you couldn't even get dev kits or write software for the vast majority of handsets, full stop.