RE: [syndication] Re: SSR-Enabling an RSS 2.0 Module

["Danny Ayers" <danny666@virgilio.it>]

> Extraordinary piece of work Danny, but I'm on a bit of a semantics
> kick lately -- when you say making RSS 2.0 suitable for use on the
> semantic web, do you mean, transforming the straight XML syntax of
> RSS 2.0 into an RDF-based format that can be combined with other
> RDF/XML formats?

Thanks for the kind words. In answer to the question, yes and no. Yes in
that this can be done, but no in the sense that the format is irrelevant.
The SSR thang simply disambiguates RSS 2.0 so that it can be interpreted
using the RDF model. If SSR is used, then an inference engine could reason
with the information provided in an RSS 2.0 feed without going anywhere near
XSLT or RDF/XML.

I don't necessarily see this as enabling RSS 2.0 for
> the semantic web, because there's little true semantics associated
> with RSS.
>
> Actually, that's the real reason for this post -- do you all see
> a 'semantics' associated with syndication data, and if so, what is
> it? Probably should just start a separate thread, but Danny's posting
> seemed a natural entry point.

Ok, the definition of 'semantics' is a big pit full of lumpy stuff (try John
Sowa's material for starters). But in the limited context of computer data I
think it deals with things like the machine interpretation of data, the
description of resources (i.e. things that can be identified) and
relationships between them. Probably the biggest defining characteristic of
'semantics' is that it/they can make data *more useful*.

I would suggest that within these bounds, there definitely are semantics
associated with syndication data. An item in a syndicated feed often
identifies a document, and it contains data related to that document :
title, description, publication date etc. There isn't much of interest that
can be done with this core stuff, but it doesn't take much to make it much
more useful:

Take RSS and add the  DC and FOAF vocabularies - we can now say that an
identified document (rss:item) has been authored (dc:creator) by a
particular person (foaf:Person). We know the feed comes from another
person's (rss:channel -dc:creator-> foaf:Person) blog. We might find the
description of the poems of Wordsworth by Shelley (either!) particularly
interesting. So we ask our newsreader to highlight them if they appear.

To make it a bit more fun, call the newsreader our agent and give it the
same rule to watch for items, but now ask it to find any recent additions to
Shelley's blogroll and see if the authors of those blogs have recently
written anything about Wordsworth. If they have, send a message to your
mobile and buzz, unless you're in a meeting or bed (your schedule, expressed
in another RDF vocabulary is checked using rules defined in OWL).

Coming back to syndication - there are semantics in RSS but they are very
limited in scope. But if you view RSS more as a way of defining channels or
pipelines for potentially any kind of information, it gets a lot more
interesting. The basic RSS pieces like item (effectively the URI), title &
date just become a bare minimum you're likely to want to know about
something. Anything.

Cheers,
Danny.