FYI, We're not far from an incremental release of the Harvester with a few new features and tune-ups. I hope to have a little more information available about this shortly. In answer to your questions:
1. I've committed a fix to CVS to sort the archives list by default; see http://pkp.sfu.ca/bugzilla/show_bug.cgi?id=2430
for a patch.
2. Currently, only the administrator can delete a data provider via the Archives list. We've discussed changing this (e.g. to allow archive submitters to update records) but haven't reached a decision yet.
3. Data provider functionality isn't currently included, but several users have expressed interest in this feature, including one who is potentially implementing it; given the modular design of the Harvester (e.g. harvesters and metadata formats are plug-ins) it will probably be a complex task.
4. We haven't looked into the "Friends" elements, but this is certainly a good candidate for future inclusion.
Regarding styles and customization, we'll be adding a few features in the near future such as stylesheet and logo uploads via the GUI. In the meantime you'll need to do this by hand e.g. by replacing the logo file in the distribution directory.
Fuzzy querying logic would probably not be too complicated to add, given the Harvester's existing inverted index, but we haven't had much call for it; I'll bring it up with the team.
Public Knowledge Project Team