W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2004

Re: Amaya-like project for RSS?

From: Danny Ayers <danny.ayers@gmail.com>
Date: Sun, 28 Nov 2004 22:52:43 +0100
Message-ID: <1f2ed5cd0411281352382e2fe1@mail.gmail.com>
To: Benja Fallenstein <b.fallenstein@gmx.de>
Cc: www-rdf-interest@w3.org

On Sun, 28 Nov 2004 22:15:57 +0100, Benja Fallenstein
<b.fallenstein@gmx.de> wrote:

> I'm not the W3C, but I'm willing to give this a try :-)


> Does anybody know an aggregator library for Java (LGPL-compatible, or at
> least GPL-compatible) that gets feeds from the Web and makes them
> available as RDF, no matter what RSS or Atom they are (but not stripping
> out additional RDF if the feed is RDF-based)?

I believe Raptor's got pretty wide coverage.

I've got some feed fetcher code in Java, but I wouldn't really
recommend it ;-) It works, but it was designed for code demo purposes
rather than being bulletproof/efficient. I guess it might make a
starting point for refactoring (in one setup it actually dumps the
feed data to the filesystem as RDF/XML, for mopping up by Kowari). One
of the main drawbacks is the HTTP client itself - I've only got the
old-fashioned blocking version. In Java there's the Rome toolkit which
has the new I/O but is straight XML-based. It would give you objects
from which the graph could be built (they were talking of possibly
using my soup parser, but I haven't heard anything recently). I'm
having trouble with my Web host at the moment, but I'll put up the
latest code asap.

You can get a long way with Morten's feed -> RDF/XML XSLT, I used an
early version of that in my IdeaGraph play. It now supports Atom too.
John Cowan's TagSoup should make an excellent cleaner, but I think
you'd have to create the 'profile' (or whatever he called the schemas)
for RSS 2.0 etc.

> The interesting part, of course, is in the connections; you're looking
> at a blog entry about an interesting project and you can look at a view
> with DOAP information about the project (linked by the blogger), which
> would also show other blog entries about that project (if linked by
> their bloggers), and then you could add your own relationships (e.g. to
> other projects with similar ideas).

Absolutely. I did a flaky proof-of-concept with a graph-based display
(in IdeaGraph) using a crude client  - it worked, but I got the
impression a lot of work would be needed in the UI to make it
friendly. But being able to generate RDF/XML blog posts with a single
mouse gesture was pretty cool.



Received on Sunday, 28 November 2004 21:52:45 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:07:54 UTC