Ask Not

I’ve been trying to think of a more susinct way to explain my thinking about linked data.

I was told long ago that a good business deal is one that is beneifical to both parties. You should walk away from any deal which isn’t beneficial to you. I think we can say something similar about Linked Data…

Good Linked Data benefits both the producer and consumer.

If it doesn’t benefit the person or organisation which produces it then it is not sustainable, it is certain to get bitrot. Benefits can be as simple as “it gains us sales” or “we lose funding if we don’t do it”, or much more subtle; “it enhances our reputation to provide this service”.

“Benefits the consumer” is even more straght forward: If nobody wants to comsume your data, why the hell are you bothering? “We’re semantic web researchers” works as an excuse for some of the data we put out and I don’t think anybody has ever actually used. Much of our RDF is just a pretty model of our internal data-model and nobody has a use for it, in it’s current form (homespun ontology). As a result much of it had got bitrot as nobody noticed issues. (Since we realised this, we’ve been working through it slowly trying to replace our local assumptions with more generic things like DC, FOAF etc.)

What I’m trying to say here is for the next round of Linked Data/RDF adopters as we shift from exploring pioneers to early homesteaders:

“Ask not what you can do for Linked Data, ask what Linked Data can do for you!”