Well, it finally happened to me.
Here I was, sitting in my hotel for SLA 2011, waiting for Amy Buckland to turn up so we can get dinner, watching Hairspray, and working on some metadata standards for data catalogs, as part of OKF. It’s something exciting and new, and when I was thinking of the necessary fields each record should have, it happened.
What would the unique identifier of each record be? Does it need the “identifier” field (which would probably be the record number or something)? Not really. I mean, what would that number be other than arbitrary? In linked data, it would be identified through all the other URIs. I was thinking like your old school relational database – each record needs a number as short hand for calling it, but that’s not necessary in linked data. It’s not the record, it’s the object. Or it’s to say the object is the record? Or let’s just toss out the records (except for this record).
So I finally get it. I think it took going outside my comfort zone (not everything can be like MARC), and a little Divine inspiration.