- From: Peter Ansell <ansell.peter@gmail.com>
- Date: Thu, 8 Nov 2007 16:35:59 +1000
- To: "Giovanni Tummarello" <giovanni.tummarello@deri.org>
- Cc: "Matthias Samwald" <samwald@gmx.at>, public-semweb-lifesci@w3.org, "Holger Stenzhorn" <holger.stenzhorn@ifomis.uni-saarland.de>, "Handschuh, Siegfried" <siegfried.handschuh@deri.org>, "Klaus-Peter Adlassnig" <klaus-peter.adlassnig@meduniwien.ac.at>, "Andreas Blumauer (Semantic Web Company)" <a.blumauer@semantic-web.at>, "Fox, Ronan" <ronan.fox@deri.org>, p.roe@qut.edu.au, j.hogan@qut.edu.au
On 08/11/2007, Giovanni Tummarello <giovanni.tummarello@deri.org> wrote: > Hi Peter > > Pipes.deri.org looks reasonably confusing... If they had a graphical > > interface it would be more obvious whether there were pipes in it, as > > right now it just looks like a web (text) editor for low level XML > > > I believe that the XML is really trivial, the text on the homepage > should really be clear. > If you confidently can say you read it and something is still unclear > i will be happy to improve as needed. > > As far as the advantage of the model, pipes are handy to do specific > information aggregation tasks, e.g. you might want to do an RDF pipe > about somebody you might want to aggregate his foaf file, his DBLP and > or pubmed pubblications, his dbpedia page (if he/she is a famouse > scientist.) . These sources use different URIs.. so a pipe can > accomodate that (just put transformation blocks) > > the result is an RDF model calculated on the fly. If the person > publishes a new paper, it wil lbe reflected in theo output of the > pipe, if a nobel prize is won, it will be in the dbpedia page so it > will show in the pipe.. if a new friend is made and the foaf file is > updated, it will also be in the same RDF. So its plumbing.. yes. but > will save a lot of time rather than the alternative.. e.g. hand coding > java or whatever to do the same thing. > I kind of envisage each of the "pipes" in pipes.deri.org to be the low level implementation of the "pipes" in pipes.yahoo.com. If that makes sense. It would be nice to in the end develop a graphical mashup with the idea of "ports" like in Kepler for instance so that you can chain them together. Developing a mature set of pipe components though is definitely okay. I personally understand the XML and what it does, I was just querying about its widespread use and where that would go. Hand coding is definitely not preferable when the alternative is that simple. Keep going the way it is going as it would be great to have a few really really good semantic web applications which can be demoed to end users to save them doing for instance what they are accustomed to with compiling lists of references related to a specific set of publications and other tasks etc. Starting to abstract away from the SPARQL and XML, when the meaningful parts can simply be entered as a parameter in a text box instead, would also be good. By that I mean don't frighten people off with code they don't know about on their first go if at all possible. Cheers, Peter
Received on Thursday, 8 November 2007 06:36:19 UTC