I don't disagree with him but I still think Appcelerator is worth it. Because as annoying as Appcelerator can be it's still quicker than writing your app twice and Apple's devotion to objective-C makes it very hard to carry over code from other platforms.

I would give two warnings though (which are supported by the author's piece)...

1. You need to make sure your use case is covered in the Kitchen Sink example. If you can't take those pieces and put together what you want to build than you shouldn't bother with Appcelerator (the documentation is pretty useless)

2. You need to buy support. The community is non-existent but the staff support is pretty good. A lot of times their advice boils down to "just ignore that error" but that's helpful with a lot of the errors generated.

Also note Appcelerator just bought Aptana, a company that makes development software. So clearly they realize there's a problem are are taking steps to fix it.

Edit: One last thing in their defense. Though they've been slower than I'm happy with in making their platform easier they have added a lot of new features. Things like Bump, bar code reading, credit card reader support and in-app purchases aren't available elsewhere (though these were just announced so I don't know how well they work yet)