Re: [ANN] xOperator 0.1 (Semantic XMPP / Jabber Agent)

Giovanni,

I don't like DBin's idea, because I don't like big "all in one" applications
(I think you know about the Haystack history). Moreover, eclipse platform is
a very heavy solution for ordinal users. Users are waiting for simple useful
tools and plugins (for MS Word to open document by semantic criteria, for
WinAmp to find music by prefers and so on). And I would like to have a
general 'semantic data repository' and such tools on top of it too.

Do you agree that semantic data is more important than the applications
which process them? Why I need to use the special big application instead of
using different useful for me applications? I would like to use a Firefox
plugin to find contacts, for an example. And it's not a good idea to run the
eclipse-based big application to do that task.

I have just reformulate my idea in a other side [1]:

Semantic Web <http://en.wikipedia.org/wiki/Semantic_Web> is the greatest
technology and different useful services and tools already exist too, but we
haven't have a significant impact of them in our live. Several huge online
storages, like Wikipedia, DBLP and others, have been represented as RDF
under Open Liked Data
initiative<http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/LinkingOpenData>.
But it isn't enough, because we as ordinal users haven't ways for publishing
their semantic data directly like we can via GData
API<http://code.google.com/apis/gdata/>
.

I would like to write different useful agents on top of such API for
reading/writing RDF directly: PIM, bibliographic manager, moneys, personal
categories and so on, information from one agent will be available in
another immediately and visa versa. For an example, I would like to describe
my personal music collection and prefers and then I will show that data in
my blog through a special Wordpress plugin and see only interest for me news
in the Amazon site. Moreover, we could use another implementation of
'semantic API' in desktop and use it personally or in the Intranet.

Luckily we already have all technologies to create that API: REST
architecture <http://en.wikipedia.org/wiki/Representational_State_Transfer>,
RDF data model <http://www.w3.org/RDF/> and different RDF serialization
formats. I have just developed my own semantic API
version<http://code.google.com/p/semap/wiki/RESTfulAPISpecification>from
scratch. But, I think, that it's very responsible work, and I would
like to participant in a similar project.
 [1]
http://semapbus.wordpress.com/2008/02/24/we-need-for-something-like-gdata-api-but-for-rdf/

On Sat, Feb 23, 2008 at 4:04 PM, Giovanni Tummarello <
giovanni.tummarello@deri.org> wrote:

> Hi, my 2c
>
> dbin 2.0 is out (though we never announced it :-) (and it doesn't even
> say on the homepage! just get it out of the download page) )
>
> we removed the P2P part of dbin 0.x  and we went with the easiest and
> most robust possible model : http publishing of RDF graphs.
>
> I'll give you an example.
>
> when you download dbin you select the name of a "base graph" for
> yourself as a user. You call it, for example "giovanni.rdf" .. dbin
> will tell you. .ok "your graph is
> http://public.dbin.org/users/giovanni..rdf"  (but the base server can
> be changed, just put a PHP file in your homepage and you're done)
>
> at this point one can add instances, annotations etc.. and when one
> presses the "commit" button, its all reflected at the location above.
>
> to cooperate one can simply import other's people base graphs and
> they're "superimposed" to one's base in the local vision.
> Obviously any rdf graph can be imported and superimposed as well (e.g.
> something from dbpedia, foaf files etc.. ). A Sindice
> (www.sindice.com) search engine window is provided within dbin for
> convenience (allows locating RDF graphs around).
>
>  Pinging one's own graph to Dindice makes it possible that others can
> discover our information in the open and import it. (there should be a
> checkbox "automatically pinging sindice" but i am not sure if its on
> the last downloadable version)
>
> Obviously, one can import the outcome of a RDF processing engine
> online like Semantic Web Pipes (http://pipes.deri.org)
>
> by doing so one can create a pipe which imports base graphs from
> people and other arbitrary RDF.  Thanks to the new patching operator
> (pipes 2.0 one can even remove triples from certain  other graph by
> creating "inverse" versions of RDF modells which can then be applied
> as a "patch".
>
> Why this seems preferrable (imo) to other methods, including jabber
> etc (we had a jabber driver for the DBin 0.3 p2p engine i believe) ?
> because its http:
>
> * security and trust issues can be solved the http/dns/authentication tec
> way
> * any programming language and web platform can access the data right
> away with no extra library
> * nothing new to learn in term of paradigm and technology
> * using a publishing server (a silly php script or webdav will do) so
> it doesn't require the user to be online for his data to be available
> to others.
> * works behind firewalls and very restrictive network configurations etc.
>
> but there are shortcomings of course..  true cooperative editing of
> RDF models (that is deleting triples other have said) using HTTP is ..
> very difficult. So the best thing in this way is to share access to a
> remote DB, e.g. a sesame, but then you forge tthe speed that you get
> when using dbin (dozens of queries per second are needed to create a
> nice UI when browsing the data locally) so some copy/caching mechanism
> should be in place (etc etc.. ).
>
> Anyway .. i believe that we (in general, as community) have now all
> the basic i think also for information exchange.  But its an entirely
> different matter from actually being any close to delivering something
> that people actually want to use.
>
> Giovanni
>
>
> On Sat, Feb 23, 2008 at 9:31 AM, Dmitry Ulanov <dulanov@gmail.com> wrote:
> > Hello all,
> >
> > I think, that xOperator is an amazing project, but I have an idea. Where
> are
> > different useful projects already exist for personal usage (ala semantic
> > desktop vision): Zotero [1], DBin [2], Gnowsis [3], Soprano (NEPOMUK
> > reference implementation for KDE) [4] and others. But it's impossible to
> use
> > RDF data from one application in another. We doesn't have something like
> a
> > 'semantic system' for RDF by analogy with the file system for ordinal
> files.
> >
> > We'll take a lot of benefits if we start to use a 'semantic system'
> through
> > a language, RDF storage and/or platform independent interface. We have
> > examples of such attempts: semap's HTTP protocol [5] (doesn't
> independent
> > from the Sesame RDF storage), Soprano API [6] (doesn't have a language
> and
> > platform independence). I have just developed my own RESTful API from
> > scratch [7] too. On top of such semantic systems we'll can to implement
> > numerous useful semantic agents and information from one agent will be
> > available in another immediately. It's like the old WinFS [8] idea in a
> new
> > RDF-based point of view ;)
> >
> >  NEPOMUK [9] is a great European initiative to implement the semantic
> > desktop vision, but it doesn't try to create a general semantic system.
> What
> > do you think about it? Do you know similar ideas or projects?
> >
> > [1] http://www.zotero.org/
> >  [2] http://dbin.org/
> > [3] http://www.gnowsis.org/
> > [4] http://soprano.sourceforge.net/node/1
> >  [5] http://www.openrdf.org/doc/sesame/users/ch08.html
> > [6] http://api.kde.org/kdesupport-api/kdesupport-apidocs/soprano/html/
> >  [7] http://code.google.com/p/semap/wiki/RESTfulAPISpecification
> > [8] http://en.wikipedia.org/wiki/WinFS
> >  [9] http://nepomuk.semanticdesktop.org/xwiki/bin/view/Main1/
> >
> >
> >
> > On Wed, Feb 20, 2008 at 9:55 PM, Sebastian Dietzold
> > <dietzold@informatik.uni-leipzig.de> wrote:
> >
> > >
> > > Hi all,
> > >
> > > The AKSW research group (http://aksw.org) is pleased to announce the
> first
> > > public release of our semantic xmpp / jabber agent xOperator.
> > >
> > > xOperator combines advantages of social network websites with instant
> > > messaging. It tries to be a semantic agent for xmpp / jabber network
> which
> > > finds and shares content about resources (using RDF / SPARQL) for you
> and
> > > your jabber friends.
> > >
> > > This first release includes the following features:
> > >
> > > * basic xmpp account management (user account, proxy account)
> > > * management of query templates, datastores and namespaces directly
> with
> > >   your jabber client
> > > * simple query templates for easy starting with xOperator
> > > * query scripts in groovy for advanced queries and post processing
> > >
> > > Homepage:     http://aksw.org/Projects/xOperator
> > > ZIP/Tar/BZ2:
> http://sourceforge.net/project/showfiles.php?group_id=201552
> > > Changelog:    http://aksw.org/Projects/xOperator/ChangeLog
> > > Mailing list: http://sourceforge.net/mail/?group_id
> > >
> > > Inspired by Tim Berners-Lee's Semantic Agent the xOperator lifts
> personal
> > > knowledge exchange to a new level. Imagine sharing exactly the
> information
> > > you want with the people you trust. Need a phone number? Want to know
> the
> > > birthday of your best friends girlfriend (and in a second step where
> to
> > > buy her a present in a store near you)? Do you get tired to tell
> everybody
> > > that you have a new cellphone number? xOperator enables users to share
> all
> > > that knowledge in a trusted network. Built upon the already existing
> > > jabber network (secure connections, widely available) an agent running
> on
> > > your pc allows other users to query your RDF-database returning the
> favor
> > > to you.
> > >
> > > --
> > > Sebastian Dietzold - Department of Computer Science; University of
> Leipzig
> > > Tel/Fax: +49 341 97 323-66/-29
> http://bis.uni-leipzig.de/SebastianDietzold
> >
> >
> >
> > --
> > Dmitry Ulanov
> >  dulanov(at)gmail.com
> > http://dulanov.wordpress.com
>



-- 
Dmitry Ulanov
dulanov(at)gmail.com
http://dulanov.wordpress.com