Speaking of get-togethers, one event I’m definitely looking forward to is the SMW users meeting, to be held in Boston on November 22. This will be, I believe, the first-ever meeting dedicated just to users and developers of Semantic MediaWiki. It’ll be a bit of a mini-conference, with visitors from the West Coast and Germany (!). Anyone who’s an SMW user is welcome to attend - you just have to add your name to the attendees list on the wiki page. There should be some very fruitful discussions about the future of SMW and its spinoff extensions, including both the technical and business side of things. Personally, I consider any event a success when I can tell people about what I do without first launching into a discussion of how categories work on Wikipedia.
Archive for October, 2008
I got back a few weeks ago from the Web 3.0 conference in Santa Clara, California. I found the conference for the most part disappointing, and I think that was mostly caused, indirectly, by the conference’s name. It seems to have become conventional wisdom, over the last year or so, that “Web 3.0″ is a synonym of “Semantic Web”, and that’s fine with me - I think it makes a lot of sense. But now, what does “Semantic Web” mean? As I somewhat knew before, to a lot of people it means two things: the transfer of data among sites through APIs and protocols like RDF (the stuff I do); and the use of artificial intelligence and natural-language processing to try to understand the meaning behind text on the web. In this view of things, the term “linked data” covers the first of those meanings; I don’t know if there’s a term for just the second. In any case, at the Web 3.0 conference the two worlds coexisted, with the first one heavily dominating; for example, the two keynote speeches were given by Peer39 and Powerset, which are, respectively, an ad network and a search engine that do natural-language processing to get better results. I’m sure there’s a lot of usefulness in that kind of text processing, but it’s not relevant to anything I do, and it doesn’t interest me. So I unwittingly signed up for a conference mostly about the many applications of natural-language processing.
There were a few interesting parts: I had some nice conversations, and there were two strange coincidences: I saw a panel presentation by a guy I hadn’t seen since we were kids growing up in Amherst, Massachusetts; and another one by someone whose company was among the first customers of Referata. I got to talk to the first but unfortunately not the second. I also talked a little to the people from Freebase, who, it turns out, had never heard of me or of the world that’s sprung up around Semantic MediaWiki, but then again they’d have no reason to; $40 million in funding (or however much it was) can really help to focus your attention. I also spoke, at a panel on marketing, where I tried to stress the importance of selling to businesses and other organizations, using the tools of Web 3.0 as a solution for so-called “enterprise application integration”. I didn’t get the sense that there was a lot of interest in what I was saying, for whatever reason (maybe because most people were there for the text stuff). I sat next to Thomas Tague of OpenCalais, and I liked what he had to say. Two of his statements stuck with me: that all this semantic technology was probably not going to build a better Google, and more generally, that semantic-web startups seem to resemble 7-year-olds playing soccer: everyone just chases after whoever has the ball (in this case, Google, or maybe Powerset).
The trip itself wasn’t totally a waste: I stayed in San Francisco, with my brother and his family, including the few days before and after the conference. I had a very nice time catching up, and helping out with the kids. I also met up with a few friends of mine, including Nick Grandy, the guy I started Discourse DB with two years ago, who now has his own startup.
The experience may have soured me on business conferences; I guess the Linked Data Planet conference, organized by the same people (Jupiter Media), might be a better fit for me, since it’s more focused on the true-data side of things, as you could guess from the name. I might also end up at the Semantic Technology Conference in San Jose in June. That’s the big one, and it also seems more data-centric. At the last “SemTech” there were at least three presentations that mentioned Semantic MediaWiki: two that mentioned Semantic Forms specifically, one from the Halo developers, and another one that may have, since it included an SMW user on the panel. So my presence there this year would either be very appropriate or redundant.
Swirrl has been launched - it’s a semantic wiki hosting site, which makes it, in my opinion, the world’s second semantic wiki host, after Referata (you could certainly make the case for Wikia being the first, since they had Semantic MediaWiki on their site about six months before Referata was released, but I think their support for it was (and is) primitive enough that it doesn’t really count). The focus of Swirrl seems to be editable tables and spreadsheets, which is different from the approach of Referata and Semantic MediaWiki, where individual pages (corresponding to cells in spreadsheets) are what’s edited. Swirrl more closely resembles SocialCalc, the application formerly known as WikiCalc, which is a wiki-spreadsheet application, though it’s not hosted, and not semantic; and Dabble DB, which also offers hosting, plus rather sophisticated editing and viewing of spreadsheet data, but without the semantic-web or wiki functionality.
Do I view any of these companies or applications as a threat to Referata? Not at all. The big hurdle for Referata and the SMW-based extensions has always been lack of awareness about structured/semantic wikis and their possibilities, as opposed to any specific competing solution. The number of people who could potentially benefit from semantic wikis but have never heard of them will, for at least the next few years, vastly overshadow the number of people who have heard of more than one structured/semantic wiki and have to choose among them. It would be great if, when describing my site, I could say “it’s a lot like Swirrl”, as opposed to having to explain semantic wikis (and in many cases, wikis themselves) every time. So I’m rooting for them to become a household name.