Emanuele Tamponi's work is really utterly impressively cool - here we have someone who goes through all academic papers, reads & digests them and produces unique software that's simply the first generally-available anywhere.
Emanuele's current work, on BLA, will make painting so much more natural and easy.
On the other hand, Sven Langkamp is working on selections, which is less glamourous, but when suddenly everything everything starts working the way users feel it should, it is a hugely important advancement.
When software changes from "it works, but it also kind of naggingly sucks" to "oops, I've done the job I fired up Krita too", that last oops isn't often conscious - but it does make a huge difference to the experience of the user.
On the less pleasant side of things, we've had a kerfuffle this week about our metadata implementations, but that's been resolved and we've found a way forward. Krita will rock, metadata-wise in the end.
Cyrille Berger, one of our most prolific and versatile coders, has decided that he didn't want to saddle me with the task of implementing it and will restore the code, but only when he thinks that it is ready.
Personally, I think he's very qualified to do that, and since he indicated that he had already mutually fruitful and conclusive discussions with the Strigi and NEPOMUK people I completely trust him with this task.
Recently, (I'll add comments from http://www.valdyas.org/~boud/krita_article.tgz)...
This article gave a highly-positive review of Krita, especially in light of his quibbles, namely with the text tool & ldr-to-hdr bracketing, both problems which are solved in trunk - thanks to Thomas Zander and Cyrille Berger, respectively.