Re: [Freevo-devel] The future of Freevo(.org)

[Meine, H. <han...@gm...>]

Hi Dischi,

thank you, too, for all the work on Freevo and kaa, I had a lot of fun with it and had it running on a home entertainment box in the living room for some years.

However, after the birth of my first kid, the time available for "hacking" would diminish rapidly, and I learned to appreciate "ready solutions" (in my case, a PS3 with the PlayTV tuner) that would get 90% of things done right, beautiful, and mostly stable, and – most importantly – get rid of the responsibility in case something does not work.  (With my kids growing up, I see more hacking coming up slowly, again. ;-) )

I also learned quite some things from this project, and I still admire the coroutine things in kaa, a concept I had some difficulties to get my head around, but which I really love.  On the one hand, it is even easy to explain, but it is quite hard I think to convince people how much easier it would make their lives, when at the same time it looks like it would complicate things.

It is always interesting to take part in OSS projects and see how different those projects are, and how similar at the same time – for instance, IPython had very much similar problems to Freevo; the core developers started to work on a (more or less) complete rewrite and at the same time maintained the stable version, contributions for the latter kept flowing in, so that development was split between the two versions, a new maintainer for the stable branch was found, the new version had interesting features and concepts, but it was unclear how it would ever finish, the mistake was recognized.  IPython was lucky to get external support and funding for a small group of dedicated hackers to completely refactor the old codebase and then integrate the juicy bits from the unfinished new version, so everything turned out really great and eventually, all important features were retained (while getting a set of much more exciting new ones).

I also like how you split up Freevo into multiple parts and made it possible to use them independently (i.e. kaa.*); the whole architecture was (and still is) great and AFAICS maximizes the possibility that this code is not dead.

At the same time you're right about XBMC, which has a big community and is feature-rich, so I will probably have a bit on their side next.

So, thanks again, keep hacking and have fun!
  Hans