Salvete! As far as I can tell, this isn't being considered now for future inclusion. While there are improvements slated for reports, I don't think there's an overhaul planned that would bring things up to the level of fine detail that would be useful for the purpose of serving as, or going a long way towards a conspectus. RDA seems to have ducked the issue, but with the rules just now kind of sort of out http://www.rdaonline.org/constituencyreview/ it seems like they've temporarily ducked the subject stickiness. (I skimmed, it's early yet!) My mind goes back to the nifty feature in the Extensible Catalogue that essentially brute forced authority files. It would be neat to have a similar feature to operate on subjects where possible. The subroutine could check a given record, see if it were missing a Dewey, an LC, or a SuDoc number, then attempt to pull the missing field from the world at large. This could then go into a queue where a cataloguer could check what the computer was doing and confirm or deny changes. It's obviously more complex than the authority check that the XC lads are on to, but man would it be worthwhile. Having this coupled with a granular subject based circulation report would allow for ridiculously high quality collection development. Following my closing is a letter that went across the GOVDOC listserv from someone else about one of the current problems in only having things indexed under one subject heading, which was what got me rethinking this. Cheers, Brooke I'm going to reply to the list because I am also interested in this question. Our library has tried unsuccessfully to do this kind of analysis. Most of our current collection analysis depends on Dewey or LC call numbers. We map the Dewey and LC numbers to specific subject categories and then run a count of the number of items held by each library in our consortium based on that call number range. You can see the results at http://www.alc.org/acdc/INDEX.HTML The problem is we could never map Sudoc with fine enough accuracy to the Dewey or LC classification system, so Sudoc numbers are NOT included in our analysis, to my intense regret. There was some discussion a while back on govdoc-l about a crosswalk between Sudoc and Dewey/LC, but I don't recall that anyone had a ready made solution. Many of the oclc records for documents have LC numbers associated with them. It may be possible to include some govdocs by using the associated LC number contained in the bib. record. This would pick up some documents but not all. I asked Marcive once about how many GPO records came with an LC class number. They said it was a very small percentage. While attempting to use the predetermined LC class might be be better than nothing, but it still seems like a less than desirable solution. I would be interested in what others are doing.