The first order of business is to clean up the OPL taxonomy (the classification system by Subject/Chapter/Section). In the past, this was created dynamically where any new problems included in the OPL could create new parts of the taxonomy.
So, we will bring groups of webwork instructors together at a series of workshop. In small groups, they pick a new streamlined taxonomy for pieces of the OPL (where we hit everything before we are done). Next, they reclassify relevant problems into the new taxonomy. We provide tools to make this quick and easy. While doing that, they also rate problems by educational objective using a system with 6 classifications roughly based on Bloom's taxonomy. The development version of webwork lets users include these ratings in OPL searches with the library browser.
The final stage of cleaning up a portion of the OPL is to look for duplicate problems, and one's which are very similar, but different enough that given a choice, some people would want one and some would want the other. Duplicates are removed (in a way that keeps old set definition files working, but the duplicates are no longer seen by the library browser). Problems marked as similar become "more-like-this" groups. In the current library browser, only one problem from a more like this group is initially shown, but there is an icon to click on. Clicking reveals the similar problems (and clicking again hides them).
This work takes places at workshops where the grant pays most expenses and a modest stipend. If you are interested in attending a workshop, contact myself (firstname.lastname@example.org) or Jeff Holt (email@example.com).