W3C home > Mailing lists > Public > semantic-web@w3.org > November 2004

Re: 2nd CfP: ESWC2005 / preANN: DBin Project

From: Giovanni Tummarello <giovanni@wup.it>
Date: Sun, 19 Dec 2004 12:16:50 +0100
Message-ID: <41C56322.6000805@wup.it>
To: Aerostatair <links@aerostatair.com>, semantic-web@w3.org, www-rdf-interest@w3.org

Aerostatair wrote:

> Semantic Web is a very interesting sphere of research. It requires 
> total dedication and deep knowledge of linguistics and computer 
> science. Unfortunately, many of the researchers focus on one 
> particular subject, disengaging themselves from the central problem. 
> To my mind, this can lead to a fatal error in the whole research. 
> Semantic web should be viewed in complex, not as separate yet 
> interconnected aspects. We should build a proper semantic web program, 
> which will take into account every single aspect of linguistics and 
> web technologies, however this program will take months to develop and 
> evaluate. It is not something that could be developed overnight. I 
> hope scientists from all leading universities will take part in the 
> conference and contribute something of their own.  
>
>  
>

I agree so much that .. i cant hold it anymore :-)
Although not quite ready for primetime I'd like to say a few things 
about the effort we've been cohordinating for the past 14 months now.
"DBin" [1]  is basically, as you mention, an  .. "integrated semantic 
web program".

Basically all the main aspects are (or are to be) considered at the same 
time, in the same effort to show the benefit "of it all". Thus the logo: 
a guy has climbed up the "semantic web tower" (abeit a "small instance" 
of it) and is taking a look at what's interesting (hopefully something).

We started the works by designing and implementing a semantic web P2P 
algorithm (RDFGrowth)  that could at least theoretically scale "in the 
wild". We obtain this by making a large use of local resources and not 
(almost ever) bothering others in the network with your personal 
queries. (no distributed queries allowed) [2][3] .

We continue with a new trust infrastructure capable of signign RDF 
subgraphs thus maintainin a track of who said what in the same  model  
and allowing precise filtering of the data you have. Such struicture 
(called RDFTrust) contains an API which can be easily used in any 
existing semantic web project (since the signign and provenence is 
completely based on standard reifications)

We continue with a new visualization made with lightweight "scripted" 
applications for the domain of interest say "the world of beers" as in 
the demo. We call these "Brainlets".Please take a look at the 
screenshot., what you're seein is a collaborative environment where 
people discuss, add contributions, structured information, unstructured 
data (but well inserted in the structured mesh)

Anyone (witha  bit of ontological knowledge) can create a brainlet with 
a simple XML file, no coding needed. People using the same brainlet 
cohoperate in a global annotation. Annotations can be rich, messages, 
images, all allowed using http uipload services so that the actual P2P 
exchange is really just about metadata (rdf) .

We continue with.. a lot of different modules to intelligently tackle 
diffeernt aspects.. a MPEG7Audio DB [4][5] analysis module to extract 
rdf  from music  and reply to acoustic queries, a textual analysys 
module [6] to extract annotations about messages that are inserted.. so 
to create a network of  rdf annotations (rather than a classical text forum)

its all works in progress of course :-) and not much has been published 
either (unfortunately due to the usual publish/perish/do real work).
However, it is not vaporware at all, the downloadable version works 
showing a basic brainlet (With rich gui) and the P2P engine . Version 
0.2 is coming out fairly soon (realistically, end of the year) where all 
the above mentioned features (signing and trust, XML based brainlets, 
mpeg7 and textual analisys modules). Actually the're alrady in the CVS.. 
but time is needed for a real realease.

Problems? plenty, starting from my ph.d being over, funds being out ;-)  
anyway.. those who are interested please join in, its open source and 
we're very open to cohoperations.

P.S. About testing it in practice, i'd suggest people to wait the 0.2 
and the  official announcement. on these lists.

[1] www.dbin.org
[2] "/RDFGrowth, a P2P annotation exchange algorithm for scalable 
Semantic Web applications" ,  P2P Knowledge Management Workshop (Bosto, 
August 2004) /
[3] "Toward widely deployable Semantic Web P2P: tools, definitions and 
the RDFGrowth algorithm", G.Tummarello, C.Morbidoni, M. Mazzieri, 
F.Piazza , P.Puliti  ISWC04 workshop on SW technology for Mobile and 
Ubiquitous Applications, get the proper versione 
http://www.dbin.org/papers.php
[4] "MPEG7ADB: Automatic RDF annotation of audio files from low level 
MPEG-7 metadata"  G.Tummarello, C.Morbidoni, F.Piazza , P.Puliti ISWC 
2004 Workshop on Knowledge Markup & semantic annotation
[5]   "From Multimedia to the Semantic Web using MPEG-7 and 
Computational Intelligence" G. Tummarello, C. Morbidoni, P. Puliti, A. 
F. Dragoni, F. Piazza. 4th International Conferenceon Web Delivering of 
Music, Barcellona 2004
[6]   "Enabling Semantic Web Forums using Dynamic Representation 
Schemas" , L.Lella, G.Tummarello, C.Morbidoni, Semantic Web Applications 
and perspective (Ancona Italy, 2004)  
http://semanticweb.deit.univpm.it/swap2004/cameraready/lella.pdf (Note: 
very preliminar material)
Received on Friday, 19 November 2004 11:20:38 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:03 GMT