Following an earlier post on this blog, last month I found myself moderating a panel in the final session of one of the tracks at this year's World Wide Web Conference in Beijing. As I commented via Twitter at the end of the session,
"Great panel and great room... so a doddle."
The panel comprised colleague Chris Clarke, Garlik Director and CTO (and Southampton University Professor) Nigel Shadbolt, DERI's Giovanni Tummarello and David Peterson of BoaB Interactive in Australia. Each brought a different perspective to discussion of commercialising the potential of the Semantic Web, and responded well to questions from a technically knowledgeable audience.
Having expressed concerns about the relative lack of commercial engagement with the Web Conference during April's Semantic Web Gang just days before this panel took place, I was certainly reassured by much of what I heard.
Giovanni, for one, spoke with conviction about activity at DERI to move their Sindice research project onto a sound commercial footing, and flew from Beijing to California to begin meeting with potential investors. Nigel was also able to shed light on the early success of Garlik in exploiting university research, building a consumer user-base (some 60,000 registered users, of which 10,000 are already paying for additional features), and monetising both in partnership with established organisations such as the major high street banks. Garlik CEO Tom Ilube points to the next stage of Garlik's journey in an article for May's Nodalities Magazine, to which I shall doubtless return when it's published next week (disclosure: I am the editor).
Asked if Garlik's DataPatrol product would have been 'possible' without the Semantic Web, Shadbolt echoed a sentiment that I've heard elsewhere, suggesting that it would have taken longer to do and been a much harder programming task;
"unpredictable data is hard to work with, without the Semantic Web."
During questions, Tim Berners-Lee reiterated his earlier thoughts on the importance of Semantic Web technologies in facilitating 'unexpected re-use' of data, leading the panel to discuss several of the ways in which their own work has created unanticipated opportunities to push data in new directions.
W3C Semantic Web Activity Lead Ivan Herman cited a perception that universities are failing to provide sufficient skills to students entering the workforce, and asked if panellists struggle to recruit. He also pointed to SWEO case-studies on the deployment of semantic web technologies in business, and called for more examples.
Everyone on the panel was agreed in stressing the importance of building applications that solve real problems for real users. Neither users nor investors are particularly interested in being pitched with 'the Semantic Web' or 'RDF' or 'triples'; they want applications and solutions. The fact that the Semantic Web is at work behind the scenes to make those applications and solutions 'better', cheaper, more scalable or whatever is clearly important, but shouldn't be the opening gambit in conversation. Chris illustrated this point with reference to his presentation on Talis Engage from the session before the panel; a new application built in its entirety on top of a Semantic Web Platform, but intended for purchase by a conservative market (local government agencies) and use by a non-technical audience (the general public; specifically those interested in local clubs, societies and events). Chris stressed the effort that had gone into making this potentially powerful application look and behave in a conservative manner, in order to introduce its capabilities gradually to a market unused to Saas, semantic technologies and Open World opportunities.
From my (doubtless biased) perspective, we got a good discussion going between panellists and the audience, and it would be interesting to repeat the experience at a more business-oriented event such as next week's Semantic Technology or Linked Data Planet in June. The obvious omission from the panel (we did ask!) was representation from a big organisation; one of the search engines, Oracle, etc. I'd make sure to rectify that omission next time.