Open Office (or Libre Office rather) is not at unrealistic.
I looked into it half a year ago and it's clearly feasible to get a basic Cairo-canvas port going which can then be progressively be turned to use native interfaces (it might be that initial port would be way too slow btw)

The main blocker for me was that the work is really boring + I would never use it myself.
I prefer to focus on stuff I would find useful myself on my MorphOS system. That's why I wrote Scribble and ported Scintilla instead of doing similar work on a word processor.

So to conclude, OO is feasible for someone with the right skills for this kind of work on MorphOS + a large dose of motivation.