id.loc.gov is AWESOME.
Not only is it the first (so far as I know) online free search and browse of LCSH (with in fact a BETTER interace than the proprietary for-pay online alternative I’m aware of).
But it also gives you access to the data itself via BOTH a bulk download AND some limited machine-readable APIs. (RSS feeds for a simple keyword query; easy lookup of metadata about a known-item LCSH term, when you know the authority number; I don’t think there’s a SPARQL endpoint? Yet?).
No, you can’t get the data in MARC. But I think the RDF and n-triple formats you can get it in have all the info you’d want from the MARC authorities? Anyone know what, if any semantic data might be in MARC authorities but not in the RDF/n-triples? If it’s all there, you could probably transform back to MARC if you wanted to — but I’d really prefer not to, thank you very much.
And one of my favorite features, LCSH to LCC mappings. Which are in LCSH authority records of course, but now you’ve got live access to the info in LCSH authority records to make this mapping. This could come in handy.
I wish there were a way to look up an LCC and get it’s LCSH mappings too, but there isn’t on the live database — but heck, you can download the data, presumably updated regularly. You could easily re-index yourself to provide this reverse lookup.
A more or less ‘demonstration’ mapping to a dutch vocabularly of some kind shows how this format could support mappings from LCSH to other (more useful to me) vocabularies.
I believe that NLM has LCSH MeSH mappings already, in XML. The data is there to have a service like id.loc.gov give you MeSH <-> LCSH mappings, which I’d really like to see too, and could do some useful things with.
Thank you Library of Congress. This is indeed the kind of thing our pseudo- national library should be doing to aid innovation in our use of our metadata — the cost to LC to provide this to us will be nothing compared to the long-term benefits we get from it.
Bibliographic Wilderness 
Well, you don’t get everything you’d get in the MARC authorities, for example, SKOS doesn’t really have capacity for subdivisions or pre/postcoordination.
So, http://id.loc.gov/authorities/sh2008122125 doesn’t have any relationships on “Slovakia” or “History and criticism”. Just “Hungarian literature”.
Of course the good thing here is that we now have a concept URI for “Hungarian literature–Slovakia–History and criticism” — somebody can come in and layer in another vocabulary that can deal with postcoordination and those that don’t care (those that really just care about the history of Hungarian literature in Slovakia) still have an easy identifier to use and a set of relationships.
Interesting.
I don’t see any reason the SKOS _couldn’t_ structure out the subdivisions more. Did LCSH actually choose to avoid specifying some things that were in the MARC as a matter of policy; or is it tricky to do this using the RDF structures being used there? If the latter, maybe they’ll figure out a way to do it in the future. Although I guess even if the former then, as you say, someone else can add it in on top. You’d need the original MARC to add it on top though, since the original MARC subfields would tell you which parts of a pre-coordinated string are labelled ‘topical’ vs ‘chronological’ vs ‘geographic’, something rather useful that is lost from the data there.
OK, this is really cool and all. I’ve bookmarked it and will be using it daily.
But wasn’t http://authorities.loc.gov/ also a place for “online free search and browse of LCSH”?