Re: Fwd: The new Atompub WG

Graham Klyne wrote:

>
> Hi,
>
> I'm forwarding this message because I thought the historical RDF 
> connection via RSS might make this new IETF working group of interest 
> to some folks here.  I haven't looked at any of the work in detail, 
> but my tentative understanding is that the intent is for this to NOT 
> be an application of RDF.


Thanks Graham. As you say, Atom isn't RDF based, though opinions in the 
group were very mixed on the issue. Many of the people involved are 
familiar with RDF (not least through RSS 1.0) but the bits-over-the-wire 
brigade are legion.

So the format is being designed as a 'vanilla' XML application 
(including namespaces, though extension mechanism(s) are yet to be 
decided) with the normative model & syntax described in spec prose. One 
of the primary aims is to make the specifications as unambiguous as 
possible, which should have the knock-on effect of making the data 
suitable for use (after appropriate transformation) in RDF systems. It's 
with that in mind that some work has been started on expressing Atom's 
vocabulary in OWL [1], (help from anyone around here would be 
appreciated). Note that slightly modified updates to the Atom specs on 
which this is based have recently been published by the IETF, details on 
the Wiki [2]. Whatever results from Atom/OWL is highly unlikely to be 
normative within Atom, but even as informative material hopefully it 
should be useful.

One other angle that may be of interest here is the Atom Publishing 
Protocol, which uses essentially the same (feed) format over HTTP (i.e. 
RESTfully) consistently for all post/publish operations. If anyone's 
after a challenge, it would be good to see how well this might go into 
OWL-S (there's been a fair bit of work done around WSDL).

The Wiki [2] is used for working docs, the (terrifyingly high-volume) 
mailing list [3] for discussion. I've been keeping an (unofficial) 
low-volume weblog, trying to track major developments [4] (with genuine 
Atom feed at [5]).

To become a member of the IETF Working Group all that is necessary is to 
subscribe to the mailing list [3]. There are quite a lot of people 
involved, but I think the chairs/editors have been doing a good job of 
identifying consensus, and steering things in a (roughly) forward 
direction. The mailing list volume is overwhelming, but pretty high 
signal/noise, often very informative - there's been an awful lot of 
stones turned over around existing specs (RFC 3023 is a creepy-crawly!). 
Anyhow, the more input there is from RDF people the better ;-)

Cheers,
Danny.

[1] http://semtext.org/atom/
[2] http://www.intertwingly.net/wiki/pie/FrontPage
[3] http://www.imc.org/atom-syntax/
[4] http://danja.typepad.com/fecho/
[5] http://danja.typepad.com/fecho/atom.xml

-- 

Raw
http://dannyayers.com

Received on Friday, 16 July 2004 13:44:51 UTC