finally a bnode with a uri

Cologne's 2nd Web Montag : "SemWeb and Web 2.0"

Gave a semweb talk at webmontag.
I gave a little talk on "SemWeb and Web 2.0" at yesterday's Web Montag in Cologne. A very nice (and lounge-y) event (BarCamp-like) that I'm surely going to attend again. It's an hour by train away from Essen, so still close enough.
web montag lounge, cologne
Met very interesting and (unexpectedly) also very interested folks. I only had a SemWeb-in-one-slide presentation, but it turned into a nice interactive discussion very quickly, with lots of smart questions, and a 2-hour Q&A afterwards. And we even managed to discuss the possible application of RDF technologies to Web 2.0 software ("mashup-chaining" and the like).

Gartner's Research VP Alexander Linden talked about SemWeb opportunities and problems from an analyst perspective which was very interesting as well.
Gartner Hype Cycle

Thinking about the other demos (The DOJO framework, and a promising group calendaring app called Reminderix), the atmosphere and feedback at Web Montag, and then having a closer look at Gartner's Hype Cycle, I really think it's about time for SemWeb folks to put more effort in getting the mainstream Web community involved (before the Web 2.0 hype reaches the peak). And we need more apps and demos, no matter how simple. Imagine a feed aggregator where you could hover a contributor's name and would get a nice inline profile preview, with a list of the person's last 5 posts, or articles he/she commented on. That sort of stuff. Mentioning Oracle and Adobe raises attention, but as long as there are no convincing and cool Web(!) apps out there, frontend-oriented developers are not likely to invest the time into learning RDF (We know it's not as hard as the critics say, but it's also not as simple as we often claim). We should have high hopes in SPARQL! A small list of impressions (and notes for myself) related to my talk and reactions:
  • people understand the triple (subject-predicate-object) idea and the possibilities of such a generic model, esp. when introduced via known vocabs such as dublin core
  • you lose 'em somewhere between RDFS and OWL (showing a class tree with some annotations can help)
  • they are back when you give a simple SPARQL example
  • "how do I connect my data(base) to the semantic web?"
  • "how do I find the data?"
  • "cool. now, where do I start?"
  • "what does (program) code look like?" (view-source?)
  • "where is the connection to my HTML pages?" (link between the clickable web and the semantic web)
  • "known apps?"
  • "known apps?"
  • "known apps?"
  • "how does such a system handle redundancy and missing information"
  • "which tools would I need, can I work on the lower layers only?"

written in a hurry, may add more points later. The basic thing is that these have been very practical questions, so this is a shout-out (including myself) to provide more hands-on guides and webby demos.


Comments are disabled for this post.


No Posts found