- From: Wouter Beek <w.g.j.beek@vu.nl>
- Date: Wed, 11 Nov 2015 23:56:22 +0100
- To: Ruben Verborgh <ruben.verborgh@ugent.be>
- CC: Kingsley Idehen <kidehen@openlinksw.com>, "public-lod@w3.org" <public-lod@w3.org>, "semantic-web@w3.org" <semantic-web@w3.org>
- Message-ID: <CANO-r+KDirwE=P_p5+E0Bj5KCr5J9pRQ-fhYqY1cGGArSGq9Wg@mail.gmail.com>
Hi Ruben, Kingsley, others, On Wed, Nov 11, 2015 at 9:49 PM, Ruben Verborgh <ruben.verborgh@ugent.be> wrote: > Of course—but the emphasis in the community has mostly been on servers, > The emphasis has been on servers and, as of late, on Web Services. > whereas the SemWeb vision started from agents (clients) that would do > things (using those servers). > Today we are nowhere near this vision. In fact, we may be further removed from it today than we were in 2001. If you look at the last ISWC there was particularly little work on (Web) agents. Now, the Semantic Web is mostly a server thing, which the Google/CSE > example also shows. > With the LOD Laundromat <http://lodlaundromat.org/> we had the experience that people really like it when we make publishing and consuming data very easy for them. People generally find it easier to publish their data through a Web Service rather than having to use more capable data publishing software they have to configure locally. We ended up with a highly centralized approach that works for many use cases. It would have been much more difficult the build the same thing in a distributed fashion. I find it difficult to see why centralization will not be the end game for the SW as it has been for so many other aspects of computing (search, email, social networking, even simple things like text chat). The WWW shows that the 'soft benefits' of privacy, democratic potential, and data ownership are not enough to make distributed solutions succeed. However, I believe that there are other benefits to decentralization that have not been articulated yet and that are to be found within the semantic realm. An agent calculus is fundamentally different from a traditional model theory. --- Best regards, Wouter Beek. Email: w.g.j.beek@vu.nl WWW: wouterbeek.com Tel: +31647674624
Received on Wednesday, 11 November 2015 22:57:56 UTC