My Old LibTech Blog (2013-2016)

It's raining fried eggs

Author: John Durno
Date: 2013-10-01

idlocgov

The shutdown of 'non-essential' US Federal Government services is having interesting and somewhat unanticipated fallout in the world of library tech.

Lately there has been a move toward using 'linked data' services instead of local copies of data for things like large bibliographic data sets. The motive behind this is impeccable: it's really hard to keep copies of large data sets up to date, and in a world where everything is networked, why should we have to? Why can't there be just one authoritative source for this data, and we all link to that?

So now we (or more accurately, giant bibliographic utilities) are developing linked data services like viaf.org, hosted by the non-profit OCLC, for name authorities (the canonically unique and unambiguous identifier for individuals) and id.loc.gov for Library of Congress Subject authorities, hosted by the US Library of Congress (similar, but for subject headings). They're pretty new, but these services have the potential to become core infrastructure for the future development of library catalogues, or whatever library catalogues become over the next few years.

So, great, no more trying to keep our local authority files in sync with the ever-changing, ever-expanding data sets maintained by these giant organizations. What's not to like?

Well, as we discovered today, there's this thing that technologists generally refer to as a 'single point of failure.' Before today, I think people might have just laughed if you'd suggested that something as venerable and well-funded as the Library of Congress might not always be around to deliver core services. But we're not laughing now, because currently id.loc.gov is offline. Fortunately, as it turns out, we're still running our out-of-date copy of the subject authorities so it's business as usual for us, and for most other places because as I say, it's early days in the linked data realm for libraries.

So assuming we're still keen to ramp up this whole linked data thing, the takeaway here should be that going forward we really need to ensure our data is available from more than one source. Maybe we don't all need to keep copies anymore, but we need to ensure the fate of our bibliographic universe doesn't hinge on a single organization, no matter how rocklike and stable it might appear to be.

Or as one of my colleagues put it: "Better not put all your eggs in one basket in the cloud! One day there may be a thunderstorm and it will all blow up with a lightening strike".

Today it truly is raining fried eggs.