An idea which I’m working on started with the Filk Book Index, which lists bibliographic information and contents of small-run songbooks by science fiction fans. Everything is MODS under the hood, and it’s difficult for others to create the files. So I’ve started on a desktop Java application to edit MODS files. The idea is to support not the full, overcomplicated generality of MODS, but a particular way of using it that covers only the fields of interest and the way the XSLT expects it to be used.
Once I started on this, it occurred to me that it’s a small step from that application to other custom editors using largely the same code, but set up to support the local MODS profile. For instance, it could allow a limited range of roles from a pulldown menu, use relatedItem in prescribed ways, have a prescribed authority for languageTerm, etc. Each installation would have its customization.
From there it wouldn’t be a big step to doing the same for other XML metadata schemas such as PREMIS, but that’s for another day if there’s interest enough.
My first notion was that this would be a standalone Java application, with the ability to download XML files and perhaps to submit them to an FTP server. The initial discussions I’ve had with people suggest, though, that a web application would be of more interest. This runs counter to my intuition: Why go to the trouble of installing the application on the Web when you can just run it, and why use HTML forms rather than custom code? But maybe I’m missing something.
So here are my questions for anyone who’d like to respond: Does this sound interesting to anyone? In what form would it be most useful? Has anyone already done this so well that I shouldn’t bother? Answers starting with “I have money to spend on this” go to the top of the queue, of course. :)
Update: Well, lack of feedback is informative, too. On to the next great idea…
Mad File Format Science 
Format registries don’t SPARQL
Over the past few days, I’ve thrown together a simple Java GUI application to query SPARQL endpoints that purport to have file format information. The results have been less than sparkling. The only endpoint that I’ve been able to query successfully is the one for DBPedia (http://dbpedia.org/sparql), and the data there, extracted from infoboxes on Wikipedia pages, is too inconsistent for structured queries to be of much use.
The endpoints for Pronom (http://test.linkeddatapronom.nationalarchives.gov.uk/sparql/endpoint.php) and P2 (http://p2-registry.ecs.soton.ac.uk/SPARQL/) just aren’t responding, and I can’t find a published endpoint for UDFR. If anyone can help me out with these, maybe I can make some more progress.
What I have to show for my effort so far is a beginner’s knowledge of SPARQL and an application that can easily be tailored to query any SPARQL HTTP endpoint. If anyone’s interested in seeing it, let me know.
7 Comments
Posted in commentary
Tagged metadata, RDF, software, SPARQL