RE: ANN: Joseki, an RDF server. v1.0

Seth,

Why wait for v2.0? :-)

For something like RSS aggregation, you could do it 2 ways, depending on
what effects you want:


1/ If you want something like a site publishing changing information, you
could use a relational database backed model behind the URL.  The database
can be shared between the Joseki server and another application that is
managing the update of the database.

Then, on the server site, the application can write into the database
(transactionally) and Joseki can provide query access to it for the web
side.  Joseki access to the model is transactional where the model supports
transations.  Database update can be as simple as a short programme to read
a file of new RDF into the model (I can send you the code - I nearly put it
the Joseki distribution but I ran out of time to test it - v1.1 ...).


2/ If you want anyone to be able to update the model, then Joseki does
provides an operation to add one RDF model into another over HTTP POST.  If
you configure in an "add" processor, then any one (there is no security
except that provided by the servlet container server) can update the model
which might be OK for experimentation (Note: file-based data is not
currently updated on disk).

There is a transational BerkeleyDB Jena model as well, if you prefer to use
BDB, done by Ivan Mikushin [ivan@openmechanics.net].


In your original post, I thought you wanted Joseki to automatically combine
the models by setting the configuration file up - there is nothing to
prevent you writing a simple application that manages a database shared with
Joseki server to get live update.

Would either of scenarios work?

	Andy

> -----Original Message-----
> From: Seth Russell [mailto:seth@robustai.net] 
> Sent: 20 January 2003 22:10
> To: Seaborne, Andy
> Cc: www-rdf-interest@w3.org
> Subject: Re: ANN: Joseki, an RDF server. v1.0
> 
> 
> Seaborne, Andy wrote:
> 
> >Seth,
> >
> >Currently, Joseki would need the publisher to smush the files first, 
> >then attach the smushed one.  The configuration processing isn't 
> >designed to support what you outline.  It could - I hadn't 
> thought of 
> >doing that, that's the value of feedback - although it might 
> also arise 
> >from a mix-up in the configuration.  Today, it would just 
> get confused.
> >
> Yeah, combining RDF files from different members of the same group in 
> the same model is what some of us think the semantic web is 
> all about. 
> Aggregating RSS 1.0  feeds being perhaps an obvious case in 
> point.   In 
> fact if that one feature were present it would motivate useing the 
> server  for many applications.
> 
>  ... perhaps in v2.0 ... huh, huh ?
> 
> Seth Russell
> http://radio.weblogs.com/0113759/
> 
> 

Received on Tuesday, 21 January 2003 05:17:39 UTC