22 June 2009

Link intermediary?

I'm looking for advice. Here's the problem: what's a good persistent link mediator, along the lines of how a DOI resolver works for publications, but for something else that we're going to curate over time. Let me set the background.

We're assembling a genus-level revision of the Decapoda (roughly 3000 taxa). Along the way, we've put together a pretty definitive list of the references for the authorities. Because Decapoda are such a long-known group, lots of these are of the really nasty historical sort (multiple years of publication, individual papers with discontinuous page ranges, printed dates of publication that aren't the actual dates of publication, etc.). So the bibliography for this thing (some of us think, anyway) is nearly as valuable a piece of work as the taxonomy.

Add to that the fact that along the way we've been assembling PDFs of many of the primary references and making them available via the bibliographic entries (all at http://decapoda.nhm.org/references).

Like the taxonomy itself, the bibliographic information will change with time after the work is formally published. Taxonomy changes due to new science; these references will change as we get better information on true publication dates, authors, and so forth. Plus, we'll be adding new PDFs and links to DOIs for contemporary publications.

In the ideal world, we'd do a whizz-bang digitally-enabled journal publication where each taxon would be linked to the stable taxonomic repository for that information (e.g. WoRMS, where this taxonomy will probably end up). Reality: we won't get the information online there in time for the article publication, and we can't delay publication or some of the collaborators will forcibly rearrange our anatomy.

What we can do, though, is create links in the publication for each of the bibliographic references. Those would go to our curated web database of bibliographic entries, which also gives access to whatever PDFs or DOIs we've scared up for them (more with time, remember). That's a nice first pass.

But ultimately, as Keynes said, we're all dead. So we plan to hand off this bibliographic database to a stable bibliographic provider (the Biodiversity Heritage Library's new reprint initiative, for example). But that will happen "later" (when? who knows).

So... here's the question (if you're still alive): Is there a stable intermediating service on the Web we can use to direct these links? Initially, we'd populate the entries with links that would get you from a click on the journal article to a bibliographic entry on our website. "One day" we'd rewrite those entries so that the same click on the old archived journal article would get you to the new shiny bibliographic repository that will happily host our entries in the future.

I'm eager for suggestions.

Thanks!

2 comments:

  1. Dean, I think http://purl.org is probably the kind of thing you are looking for. It acts as a redirection service where you have a URL of the form http://purl.org/XXX which resolves to a URL that you specify. A number of vocabularies (including Dublin Core) use PURLs, and it's backed by OCLC, so I'm guessing it's reasonably stable.

    A more exciting (but risky) strategy would be to use URL shorteners such as http://bit.ly, which would give you access to statistics on usage, etc.

    ReplyDelete
  2. Thanks! http://purl.org looks good.

    Using something like http://bit.ly or other url-shorteners would work if they had a facility for rewriting the target URL for a given short URL. None of the shortening services I looked at provided an interface for that. Pity. It would be cool.

    So the best we can do for the moment is look at our own server logs. That, of course, will only work until we redirect the URLs away from ourselves. Oh well.

    ReplyDelete