I don't think we need a tight naming scheme. I think we need an on-topic naming scheme together with some categories and tags.
A guidance might be the Linux development, if you look at the Ubuntu packages set, you see there is a limited list of categories, and I believe a package can fit in two categories.
When doing a search, one just types in what comes to mind 'source control', now the package manager searches through the available packages and returns a sorted list of packages. These can be filtered upon categories. I think there are several ways of sorting, but the first is:
Popularity With VIPM we can use the internet connection to have a database where the popularity of packages is traced. A system would be '+1' for each install '+2' for each package that is installed after 4 weeks. '-1' for each uninstall. This data can be anonymized, or only kept local in the current repository.
Relevance Relevance can be calculated upon, does the search term comes up in the packag name, tag-list or description
Code standard
Paid
Open source
License
I think the default sort should be 'Popularity', which can be influenced by 'Relevance' (for instance a popularity score with 80, but relevance 5, could be below pop=30, relevance=95).
Ton