I'm highly invested in Ushahidi. It's a wonderful system for crowdmapping and if I'm not mistaken is in active development by at least one of the original developers of Kohana. Beyhond that, I find Kohana's convention over configuration paradigm to be very nice and easy. It took me a few hours to develop a module for Ushahdi based on it and my server installations are some of the most stable PHP web app installations I've ever had.
Saying all that, I think Kohana can and should be further developed for the sake of Ushahidi specifically and as a general purpose server framework.
I went through the various discussions in this forum and the previous one and later did a little research about PSR and it seems to me that what's stopping Kohana from being PSR compliant are issues that may be solved with some effort, especially around Class naming conventions and routing. But it doesn't seem to be such a big deal to make it PSR compliant and capable of moving forward like Symfony, Laravel and other frameworks.
Can you imagine what it would be like to load any PSR compliant module into Kohana and easily build web applications the way you are used to?
So what's really stopping us from doing those modifications to the framework? Is it that hard to accomplish? Doesn't seem so to me.
Finally, let's say it isn't a feasible move for Kohana. Would you be interested in migrating existing projects from Kohana to another framework?
Wouldn't it be the same effort to update the framework and keep running on a community owned, fully open-source infrastructure?