Return-Path: <strueg at mandriva dot com> Delivered-To: anders@feder.dk Received: (qmail 43643 invoked from network); 23 May 2008 19:25:04 -0000 Received: from moutng.kundenserver.de (212.227.126.188) by 10.0.15.11 with SMTP; 23 May 2008 19:25:04 -0000 Received: from [10.1.101.3] (i3ED6C833.versanet.de [62.214.200.51]) by mrelayeu.kundenserver.de (node=mrelayeu4) with ESMTP (Nemesis) id 0ML21M-1Jzct13nbT-0000KD; Fri, 23 May 2008 21:25:08 +0200 From: Sebastian =?utf-8?q?Tr=C3=BCg?= <strueg at mandriva dot com> To: Anders Feder <anders@feder.dk> Subject: Re: Semantic GNOME Desktop Date: Fri, 23 May 2008 21:26:35 +0200 User-Agent: KMail/1.9.9 References: <1211463659.9067.3.camel@voyager> <200805231641.44138.strueg at mandriva dot com> <1211565441.9490.30.camel@voyager> In-Reply-To: <1211565441.9490.30.camel@voyager> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200805232126.36415.strueg at mandriva dot com> X-Provags-ID: V01U2FsdGVkX19geH7Jyz/iU6vQgS0uAFy0JCi4pUf2vI0wymJ m06WqmxTWSuMaZXVfVJtP9FI+7MoKpwo+JbEPwuLH3uxK6i1G4 +ZjpbaZlLFcd7yZaHbAGdP2gQ9ldqLl X-Evolution-Source: pop://anders%40feder.dk@mail.feder.dk/ On Friday 23 May 2008 19:57:21 you wrote: > Hello Sebastian, > > Thanks for your reply. I've posted an excerpt of your mail on the wiki > page. > > What are your thoughts on ontology alignment/matching? Is there a place > for this in your architecture? For instance, an application querying for > FOAF data should not be excluded from equivalent data stored in NCO, in > my opinion; the architecture could handle the alignment process > centrally and transparently, possibly via special plugins for the > purpose, rather than requiring each and every backend to implement an > alignment strategy of its own. actually that is being worked on in the scope of the semantic clipboard. I= =20 will be able to give you more details next week. > Is it possible for applications to 'subscribe' for updates to virtual > folders (or queries, otherwise) in your architecture? I see signals for > 'statementsAdded' and 'statementsRemoved', at least? Not yet. The virtual folder service is work in progress and not part of=20 the "standard" RDF API which "only" provides basic functionality. > I agree that stuff like contacts should ideally be stored only once; > some redundancy, =EF=BB=BFon the other hand, is often useful in practice.= For > instance, you will want a copy of your contacts on your PC as well as on > your PDA, so you can access your contacts even when away from the PC. sure. > Redundancy inevitably leads to syncronization issues; a challenging, > but, I think, unavoidable subject on the semantic desktop. I don't know > if there is any solutions particular suited for syncronization of > semantic data? Perhaps one could be developed under NEPOMUK? AFAIK there is no work done in that area so far. But it is an important top= ic,=20 that is for sure. This weekend I will be away. But next week I am back and maybe we could mee= t=20 on irc and chat a bit about the possibilities? Cheers, Sebastian > Best regards, > Anders Feder > > Sebastian Tr=C3=BCg wrote: > > On Thursday 22 May 2008 15:40:59 you wrote: > > > I'm looking informally at possibilities for integrating the Semantic > > > Web and the GNOME desktop. So far, I've outlined a few ideas at: > > > http://live.gnome.org/SemanticDesktop > > > > > > I would like to ask you if you have any comments on these ideas and h= ow > > > to proceed - and if you see any potential for collaboration with your > > > projects? > > > > Hello Andres, > > > > thanks a lot for your email. I think it would be a great idea to > > cooperate on the semantic desktop front. Areas I think we can share are: > > > > - The DBus interface for the semantic data storage. At the moment it is > > based on Soprano [1] which comes with a full-featured RDF DBus API [2] - > > The ontologies: here we already use Xesam [3] for everything > > file-metadata specific and the Nepomuk ontologies [4] for other stuff. > > This also means that we do not use OWL as it was considered way to heavy > > for desktop use. Thus, please consider using NRL [5] in combination with > > NAO [6] instead. - We are currently working on virtual folders which is > > pretty much what you outline, too: a "view" on a sparql query (or a more > > generic desktop query in my case). Here it would be nice to share the > > interface, so that KDE and Gnome apps could use the same folders. > > > > As for your backends: this is what we do, too. For example: > > - Strigi analyses local files and stores the data (for desktop search) > > - An Akonadi agent collects information about contacts and emails and > > stores it (actually my vision would be to store stuff like contacts only > > once, in the rdf store) > > > > Additionally applications can insert new information. There is not much > > there yet except for tags and ratings and source URLs for downloaded > > files. > > > > Ok, this just to get the communication started. :) > > > > Cheers, > > Sebastian > > > > [1] http://soprano.sf.net > > [2] > > http://api.kde.org/kdesupport-api/kdesupport-apidocs/soprano/html/sopra= no > >_server_dbus.html [3] http://xesam.org/main/XesamOntology > > [4] http://www.semanticdesktop.org/ontologies/ > > [5] http://www.semanticdesktop.org/ontologies/nrl/ (important part is > > that about the named graphs, views are not implemented anywhere yet) > > [6] http://www.semanticdesktop.org/ontologies/nao/