Mark: I missed it at first too, but once I found it, I was a much happier panda Shelves at that point become organized by whatever you want them to be - for example, by category, rather than by location - and the location search lets you find them. Quite handy, imho.

* Treehugger,
* Nemo RDF, by me and isn’t close to beta,
* RDF Twig.

Also note another Norman Walsh creation, SXPipe. It would make multiple transformations much easier.

What I’ve been doing is scanning a set of books, selecting them all (order by scan time and click-hold shift-click from the top to the first book from the shelf) and then clicking “My Info”, setting the “Location in Building” field, creating a shelf, and dragging all the books in. Total time to do all that is ~30 seconds, and lets me get at the data both in the view (I have shelves on the left) or in the order by (where you can sort by Location In Building).

In the RDF/HTML output, there’s no information about which shelf something is in, since I don’t have a need for it; The information is mostly there so I can share a listing of what books I have with other people, and eventually so that I can aggregate other things like reviews in with the data. RDF gives me a nice unique identifier for my copy of the book that I can treat as equivilant to the Amazon copy, then pull in local reviews, stuff like that.

You’ll notice if you look at the actual RDF (/library/books.rdf if you don’t want to play with content negotiation) that the library is all at the end (in the rdf:first, rdf:last…. hm, that should be a bag, not a Collection, Collections are ordered, theses shouldn’t be, just realized that) and that it’s just one big collection. I may be a special case, but outside the efficient search of the actual program, I don’t need the storage data at all.