VoCamp (SW* -[lightness]-> SWL (!) .: Repo?)
It was RDFa/eRDF that brought me back to the Semantic Web (SW*) community with a belief that something had changed in regards to the ethos of how the semantic web should be built. I’m pleased to say that something has changed and that something is gravitas: suddenly it just feel light and refreshing like, well the web.
This attitude change reminds me of a book I once read back in college: “The Unbearable Lightness of Being” (can’t remember the content of the book, but it was the label that matters in this case) and indeed it is the feeling of ‘lightness’ in the community that has me once again believing in a new Semantic Web; what I would call the “semantic web light (swl)”.
Last I experienced the SW* was three years ago attending as part of a library community workshop that was working hard on OWL ontologies, in that case it was a pizza ontology that took up the entire day of powerpointing and mindmapping to just understand conceptually, nevermind actually learning how to do the mark-up. In the new SWL it is not a workshop, but a barcamp and instead of pizza it is whisky and instead of OWL it is RDFs; and within two days of unstructured discussions I had participated in writing two ontologies, mapped out a basic architecture and decided on a SPARQL query strategy and interface for accessing the ontologies (and had some decent drinks, food and discussion on the side). More on the ontologies created and pragmatic argument behind creating them in another post (it wasn’t just all wiskey!).
I think the event was really helped along by the obvious essentials: WiFI + AC + C8H10N4O2 + Flipcharts + NoPowerpoints = HappyDeveloper. In addition the MC of the two days (TommyH <- nice to always put a face to a handle) was superb in his open and freindly dialogue with the group making sure that everyone was able to have a say throughout the event. The event was kept informal allowing the very impressive attendee list to organically steer the optimally sized group (circa 15).
Of course, deciding to once again adopt a bit of architecture as simple as RDF requires a redesign of my own mind’s architecture for my own day to day job: “learning is a revolution to everything you knew before”, and accordingly doubts of “what a repository does in term of institutional architecture” are creeping in once again. Part of this is my ongoing struggle with mediated deposit in repositories and the lack of the community in trying to change this decrepit model. I’ve excused this down-in-the-mouth attitude for repositories as the “trough of disillusionment“, hoping I would come out of the rut sometime soon. I thought the revelation would be in regards to going further up the stack, focusing on more individualised interfaces and subject specific tools? However, I find myself moving towards a more extreme version of the web’s hourglass architecture:
<–databases, systems, services, silos, containters, etc.
<–http, xml, crud, etc.
<–resource types, content, multimedia, mime, etc.
The question I’m still trying to sort is what if we forget about the containers that hold resources and just move to semantically structured content over the basic web protocols (e.g. named graphs)? Let people create there content out on the web and leave the librarians of the world (in whaterver form that comes in) to clean up after the users to make sure the valuable resources are semantically structured (linked) with the correct ontologies, vocabularies and thesauri? I realise the moment terms like ‘ontologies’, ‘vocabularies’ or ‘thesauri’ creep into the repository conversation we are in danger of relinquishing its coolness and once again becoming library-catalogue-like. However, my suspicion for why we should care is because “the resource is the new system”!? In the very near future are we going to start realising that there are patterns for resource types, and I’m not talking about small patterns but very large patterns that are ubiquitously occurring across the web. Things like YouTube videos and GoogleDoc Spreadsheets are being widely used and structured the same way (i.e. APP). Once we start to realise this common syntactic structuring we will be able to start applying the semantic layers? I see a very cool and evolved world in dead if so. In short, I think I’m saying that the semantic web or rather the semantic web light might just be the more simplistic and straightforward architecture of the web…than the repository?
Note: I’m really looking for feedback from this post so please fire away, or feel free to give me a skype (david.flanders) if you want a more syncronous discussion (better yet see you at the pub to discuss the wiskey ontology) ;-)