jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject Re: Jackrabbit with RDF facilities
Date Mon, 07 Feb 2005 22:12:35 GMT
Charles Severance wrote:
> [snip]
> 
> I tried to resist adding to this RDF thread but could not :)

know that feeling very well :-)

> A very small group in Sakai has been thinking about RDF and a Repository for
> a long time (gives us a headache at times).

I know how that feels too ;-)

> I won't go on at boring length here, but my high level view is that there is
> a basic role for RDF as "part of" JCR, but also a more interesting role of
> RDF as a "peer to" and "compliment to" JCR. There are things which RDF can
> naturally represent which are not natural fits for JCR.

Very much agreed.

> In one sentence, if JCR may be the flexible "file system" of the future, RDF
> may be the flexible "relational database" of the future.  But both
> statements remain to be proven in real applications at real performance
> levels.  And it seems premature to even imagine that a standard could really
> nail the requirements in the area at this point.

Careful. I was not talking about 'adding RDF functionality to JCR', I 
strongly feel this is not only premature but dangerous at this point, we 
really want to get this spec out of the door (has been in the making 
practically forever!), all I was talking about was to add RDF 
functionality to "jackrabbit" as a testbed that might yield enough 
requirements to influence the next version of the JCR spec (whenever 
that comes).

Also, if the RDF import/export/search are wrappers around existing JCR 
stores, you don't even have interoperability problems, they would work 
just like the RMI proxy thing.

-- 
Stefano.


Mime
View raw message