I decided to participate in this year's gsoc and Krita is an obvious choice for me.
While over the last year many great features like Python scripting and 2D animation were implemented and still being worked on I feel the basic functionalities are still lacking performance- and stability-wise, at least on my hardware.
Currently I just try to get into the codebase again by fixing bugs here and there but once the applications open I want to know roughly what I want to contribute to and have the ability to tackle the task.
From my point of view performance is a big issue but looking over the idea page for last year's gsoc many of the suggested tasks are additional features like the ones I described before.

My question is, is enhancing performance or usability for a subsystem a valid task for gsoc in the eyes of the Krita project?
I ask as the attention judging from the proposed tasks rather seem to lie in extending its features and optimizations may be reserved for more knowledgable longtime contributors that have an easier time understanding and tinkering with the codebase.

We would like someone to improve Krita's performance, but that really needs someone with lots of experience. As in, extremely familiar with Krita's codebase and much, much smarter than any of the current developers. There is no low-hanging fruit left, at all. And... It's not the sort of thing that fits Google's ideas of what a summer project should be.

To clarify, it's specifically stated that a bunch of bugfixes or the like are considered a terrible summer of code project. The idea of summer of code isn't to solely improve the program: it's more to give the student a cohesive project they can set their teeth into and cohesively finish by the end of the summer.

However, there's actually a ton of opengl related ideas on the gsoc page for this year, such as partial opengl canvas updates, or using opencl/opengl for filters. You should join the Krita channel around 10 hours from now if you want to discuss what you need to know and where you need to focus. You can also take a look at our wishbugs or the krita:abyss project on phabricator...

But, whatever you do, and this is from someone who has finished the GSoC twice, do something that is within your area of expertise. Mine is colors. So my GSoC projects were about colors. If they were about openGL, I would not have passed them, because I know NOTHING about openGL. So please look at GSoC from your area of expertise, because we can probably think up a part of Krita that needs serious polishing and yet somehow lies in your area of expertise.

Thanks for your replies.
I expected that I am too optimistic but it was something obvious I wanted to ask about.
I referred to the Idea List of 2016 as I somehow completely missed that the new one for this year is already up. I skimmed through the proposed projects and they look great

I'll be lurking on IRC for now.

EDIT:
As nobody responded on IRC and I don't want to spam this wall of text over and over I post it here instead.

Yesterday, I asked on the forums what 'desirable gsoc tasks' are in the eye of the krita project. Thera recommended to ask here and "discuss what I need to know and where I need to focus".
After I discovered the idea list for 2017 two project caught my eye "Partial Canvas Updates + ROI" and "OpenGL-enabled Brush". I'm still familiarizing myself with the codebase and haven't looked at that corner of krita at all so far. Normally I would try to understand the current implementation and its problem space before asking/bothering anyone. Especially as my OpenGL knowledge is quite basic.
Rather than secluding myself I come forward to ask what my main focus should be to be prepared for the tasks I mentioned.