2012-03-06

On usage of empty fridges

It's been a while since we've not revisited here our favourite Tao Te Ching chapter, the one from which the title of this blog is derived. Yesterday I found it was a mandatory answer to Danny Ayers' post on Google+, quoting a quite old (2003) but still relevant post by Dan Brickley "Missing is not broken". Worth re-reading indeed, as well as Danny's comment : an empty fridge is still a fridge. And even more useful so according to 老子. The usage of a fridge is allowed by this very emptiness. Whoever came back once from the supermarket to find the fridge already full has understood that a full fridge is useless until you start to empty it. Granted, by Laozi times, fridges and supermarkets were barely concepts, but useful empty clay pots did the trick.
In the quoted post, Danbri presents very sketchy descriptions of foaf:Person instances. But if you look at  the foaf:Person class itself, in the vocabulary, it provides a minimal definition, quite close to emptiness. You will find that a foaf:Person is a foaf:Agent, neither a foaf:Organization nor a foaf:Project. Plus a handful of properties such as foaf:firstName sufficient to infer that something is a foaf:Person. But this very sketchiness of foaf:Person is what enables its tremendous, overwhelming use in linked data and linked vocabularies, as described in our previous post. Because it is almost empty in its definition, the foaf:Person pot can be filled by all kinds of semantic goodies.

I've stumbled a few days ago on an even more minimalist exercise if any, the lobid vocabulary. As said in the very small ontology file (no class, one property) comment :
Under this name space new properties and classes needed for the lobid.org service are defined . Already existing properties and classes which are (re)used in lobid.org aren't documented here.
This vocabulary was submitted through the LOV suggest form. When I received it, I hesitated between a mistake or a joke. But it was neither. The submitter, Adrian Pohl, was deadly serious about it. He's a librarian, and he's German. The perfect logic behind this minimal vocabulary is that the dataset at lobid.org could be described almost entirely using existing open vocabularies such as BIBO, FRBR, DC Terms, FOAF, DAIA etc. The only missing piece was a property to link a Person to a QR code encapsulating its contact details. Therefore this property is all there is for the moment in the lobid vocabulary. But following good LOV practices, this unique property defines its domain and range using FOAF.
There was no doubt this vocabulary was to be included in LOV, less for its minimal content than for the good practice of re-usage hidden behind its emptiness. Kudos to Adrian for the lesson.

PS : Following Danny's immediate feedback I struck the quasi-double negation in the first sentence. Hopefully the mind-rebooting effect is still there.