T9 in a J2ME environment: Grrrrr....

A small rant:

I happen to like T9 predictive text. Of all my mobile-owning friends (strange to think that there are people, software engineers even, out there who still don't own a mobile phone), I stand alone. I find that using 10 keys to select from 26 letters can work, and work fast and well. Oh sure I have proof-reading errors, but that's not the technology's fault.

So I have been using T9 in java text editors (I'm even trialling a free-trial-period commercial app, just to see if it works any better than the cruder, free one I'm using). And I am not a happy man.

I thought at first it was the program, which didn't really make sense to me. It didn't make sense because, guess what, it wasn't the program's fault.

It might be Sun's fault, it might be Samsung's fault (probably is, actually), but: It's broken. Sometimes it will work, with 95% of functionality (you can't go back to a previous word, select it, and then change it, which is stupid and ridiculous). Sometimes it just won't work at all, and you get non-predictive text. Sometimes it will give you random letters from the keys pressed, with no dictionary rhyme or reason behind them. Sometimes you get the first half of the word random, and the second half predictive, but what good is that?

I am feeling very miffed. Let down, even. I have a phone which, cleverly and ingeniously, cripples every cool thing I think I can do with it. I swear it's actually malicious.