It's sad to hear you feel that way about it MIC, but I can understand that it happens.
Development is, intendedly, a bit loose until we've got some more basic features in and got used to SVN. Assigning a specific developer to add a next and prev menu item, and updating a web-page or something that he's working on this, which is about 1 minutes work, is just silly and impossible. Later, we can then tighten it up and look at more (and larger) specific tasks. We can then also take advantage of the branches in SVN so we don't affect the main codeline.
I think we've had a lot of functionality added during the latest week (although it's only in the repository right now, and not released), but we've got the player screen shined up, which was a requirement from some people, it shows ID3 tags (the code YOU sent me), it has a volume bar, realtime spectrum analyzer (about 10-20 lines of code only), next/prev song and in addition to that, I have working M3U playlist code.
I don't know about unstability, the player works 100% for me.
But, I was thinking of how releases should be handled. To make sure we have solid releases, only tested and working code should be committed to the repository. I will then, every few days (at the moment) make a public release.
Later, we probably need to set specific release dates, and list the features that will go into that release, then focus on those.
There is no immediate need for code (size) optimization. We have about 70 kB for further expansion and that is a LOT, considering that all base driver functionality is already there.
What I see as a feature we need to look at very soon is how the skin/themes can be handled in an effective and flexible way. Each individual user can then choose to design their UI as simple as they want.
Any views on this are welcome and I appreciate any feedback or ideas on how we can best take advantage of our resources.