Re: Would an Atom-RDF mapping be useful?

Firstly, I agree with Frank that the graph theory point is nonsense.

With regard to your points, though, Danny:

1. Sort of. My research (Semantic Web) and its software applications 
are likely to be consumers of feeds; the more flexible they are, the 
better for me. The best approach would be full extensibility, so that 
feed items could carry useful payload, such as RDF annotations (e.g. a 
review blog post which can use review vocabulary). That would itself 
encourage people such as myself to become "implementors", and thus I 
would answer "yes". If Atom isn't extensible, or doesn't map to RDF, 
then the answer is probably "no" to this question :)

2. Absolutely. If one isn't designed-in, someone's going to have to 
make one, anyway (like Dave Beckett's "tag soup" parser for Raptor, 
which does the job for RSS and the current Atom feeds), so the work 
might as well be done up front to allow SemWeb applications to consume 
Atom.

3. One of my target applications is a persistent store which consumes 
data sources, including RSS. Atom would be a good input. Furthermore, 
if the vision for Atom comes about, it will have wider application than 
just newsfeeds; extensibility and mapping to RDF will lend utility to 
both RDF applications (which can now consume Atom, providing more fuel 
for the fire) and to Atom itself (as a transformation can be applied to 
allow information integration, visualisation, and inference through 
RDF). Everyone benefits, and I think the slight leaning towards clearer 
semantics will be of use in the design process.

It seems an absolute no-brainer for me. The very minimum I would expect 
from Atom is a simple RDF mapping, with extensibility (perhaps through 
striping to give maximal RDF compatibility). The ideal, for me, would 
be a restricted RDF vocabulary (restricted in order to simplify 
parsing); it doesn't matter if it's not as pretty as a pure XML 
application, because presumably there will be tool support, but it 
allows for more straightforward RDF parsing and extensibility.

-Richard

On Jan 7, 2005, at 20:04, Danny Ayers wrote:

>
> In two recent posts to the atom-syntax list the same basic points have
> been made:
>
> "Extensibility via a mapping to RDF seems to me to add a lot of
> complexity (most people have never bothered to learn graph theory)
> without any real benefit." [1], "I have not seen any evidence that
> these RDF incantations have any relation to the needs of
> implementors." [21]
>
> Now I and others believe that the cost side of this can be kept very
> low (actually zero, unless you're writing an extension, in which case
> you'll have rules to follow). What's harder to quantify is the benefit
> to implementors of applications that might use syndication formats
> like RSS or Atom. So I thought I'd go ask where people might actually
> be enjoying similar benefits...
>
> So, quick questions:
>
> 1. Are you an "implementor"?
> 2. Would a mapping of Atom to RDF be of benefit to you?
> 3. It what way(s)?
>
> Cheers,
> Danny.
>
> [1] http://www.imc.org/atom-syntax/mail-archive/msg11922.html
> [2] http://www.imc.org/atom-syntax/mail-archive/msg11921.html
>
>
> -- 
>
> http://dannyayers.com
>

Received on Saturday, 8 January 2005 18:23:48 UTC