It looks like you have run into two very specific problems, the complexity of backporting MyPaint (in its current incarnation) from a newer platform to an older one, and the limitation of EQ (limited on Maemo 5 and, as far as I know, entirely missing from Diablo).

This does not indicate a general problem in the development process for Maemo though. I still maintain that it's vastly simpler to develop for Maemo than PalmOS, at least for someone with my background: I'm used to develop for Posix-compatible, Unix-like platforms. I can for the most part develop exactly as I do for my desktops, and, for non-GUI applications, just take whatever I need from the huge pool of existing Debian applications and re-build for Maemo. GUI applications need more work, but, except for some missing pieces the selection of libraries and tools is huge.

As for what will be developed specifically for Maemo (after all, what you want on a mobile device isn't necessarily interesting on a desktop and therefore maybe not already existing): Please note that this is limited by the number of active developers and the time the device has been out. These two numbers alone are vastly smaller than for e.g. iPhone, and probably Android as well. That taken into consideration it looks good to me, we're getting new useful applications every week at least. I just installed a couple of new ones today.