(unknown charset) RSS, RDF and 'better syntax': what's done, what's to do, who to do it

[[
(copying www-rdf-interest, where some of this thread should probably live)

Context for RDF IG folk: In RSS land, we seem to have consensus that the
RDF abstract model works well for an extensible, modular (and all that
good stuff) approach to site summary / channel data formats, ie. RSS 1.0
proposal, http://purl.org/rss/1.0/   
But we're having the (now rather familiar) discussion about syntax, alternate
syntax, whether to use XSLT transforms etc etc., or learn to live with
RDF 1.0 Syntax. RSS implementors need to know what to do, what's looming
and likely on the RDF front and what's merely on the "someday" pile...  
And so Bill asks one of those probing questions that are healthy to ask,
and that keep us honest and accountable: "what's going on guys - are we
going to get a new syntax? what's the plan"... Such questions deserve
more than a casual mailing-list answer like this: we owe the world an
update on status and plans for the W3C Metadata Activity, and on the
'Semantic Web' efforts which have grown up around the RDF and XML work. 
I'll ping www-rdf-interest about this shortly. Meanwhile, I wanted to
respond to Bill's comment while it's fresh in our inboxes.
]]


Hi Bill,

A long reply to a short (but recurrent) question...

On Thu, 12 Oct 2000, [iso-8859-1] Bill de hÓra wrote:

> The W3C knows RDF1.0 model syntax isn't exactly empowering, even its
> director has admitted as much: and has offered an alternative. The syntax
> can be made better and it's beyond me why this is not being done, such is
> the importance of RDF in the W3C vision of things. 

Yes, there have been some new RDF syntax proposals, and some
implementations (eg. Sergey Melnik's work on SiRPAC, and the W3C Perl
RDF parser, XSLT stuff), but it hasn't been clear to me (with RDF
Interest Group chair's hat on) that this is viewed as top priority
within the RDF community. There is certainly much
interest in the notion of a better syntax, but nothing has yet convinced
me that we're at a stage where proposing a full-on W3C Working Group
is the next useful step. That said, I would dispute the suggestion
above that nothing is being done. Perhaps we're not doing _enough_, but
there have been some small progressive steps in recent months. My slant
on all this as of early august was aired at 
http://lists.w3.org/Archives/Public/www-rdf-interest/2000Aug/0019.html

Reading back, a reasonable number of the items I mentioned have made
some modest progress, while some (the RDF Schema  Candidate Recommendation 
report) have taken a backseat to more fundamental Model/Syntax-related
efforts. As an aside, I took this approach largely because most of the
issues raised during RDFS CR period on www-rdf-comments were against
the RDF Model and Syntax spec that underpins RDF Schema. It seemed to
me (and still does) that these were the challenges most troubling RDF
implementors at the codeface actually building stuff with the
RDF spec(s). (Maybe this was a lousy or inappropriate piece of priority
juggling on my part; I don't know, feedback welcomed). 

I asked for input from RDF IG on the syntax question at the
end of August, looking to find evidence to support a stronger push on
this: http://lists.w3.org/Archives/Public/www-rdf-interest/2000Aug/0153.html
...and did get a number of positive responses. We had a lot of RDF traffic
and announcements etc during September. Buried amongst all that there
were a few efforts toward improvements on the syntax front. See
the RDF Interest Group page for links to this stuff, ie:
http://www.w3.org/RDF/Interest/#docs

Perhaps this was a suboptimal approach, but during September I took the
approach of initiating several small steps on some inter-related
problems, rather than attempting to pick off one problem to completion.
This gives a feel for some work items that need completing (and need
additional contributors / editors etc), but gives no closure on any
particular part of the entire jigsaw puzzle. In this vein we now have
such things as...

 - a skeletal 'RDF Model without the syntax' document, extracted from
   RDF Model and Syntax REC. This shows what the RDF model looks like
   when specified separately from any particular syntax. Much shorter 
   spec for starters...
 - an RDF Issue Tracking page, with a number of contributions from the
   RDF IG list including a detailed trawl through the RDF mail archives
   from Brian McBride that must've taken hours. I've been on the
   road too much since launching that document, but see it as a
   precondition for understanding what any 'new improved' rdf syntax
   might need to do.
 - there has also been discussion of a simple 'triple dump' syntax, eg.
   see http://lists.w3.org/Archives/Public/www-rdf-interest/2000Aug/0024.html
 - work on building bridges from the RDF model to "mainstream" XML
 applications such as SOAP,  eg. see the SOAP/RDF serialisation comparison at
  http://lists.w3.org/Archives/Public/xml-dist-app/2000Sep/0002.html
  or the RDF and XLink W3C Note recently published.
  I have begun a survey / background paper on this, XGraph,
  http://www.w3.org/2000/09/XGraph/ which might feed into some of the 
  XML and Protocols discussions, for example.
 - first pass at a high level historical context overview of the RDF 
  design, see http://xmlhack.com/read.php?item=800

I'm the first to admit that there's plenty more work to do. While
www-rdf-interest isn't a working group, that doesn't stop us from
trying to get some work done there, ie. putting things in place for a
more formal W3C Working Group to do the last push effort (spec
writing, getting things on the track to REC or Note publication by W3C).

so... there's lots to do. While W3C has some resources in this area, there's
more to be done than can be done within the W3C team. In short, if folk
want this (ie. a new syntax), then we (we = W3C and the RDF/XML community at
large) need to put together a group of people to make it happen. That
group might be drawn from W3C Member organisations, they might be
invited experts to W3C, they might be a subgroup of the public RDF
Interest Group, or a more formal W3C Working Group. There
are a number of ways we could attack this problem, and different
tradeoffs associated with the organisational/process options, but at the
end of the day it comes down to putting names against deadlines and
deliverables, and finding enough folk with expertise and commitment to
do the job. I'm already entangled with too many of the RDF
documents mentioned above to spend serious amounts of time on detail of 
RDF 'Better Syntax' work, but can help put some infrastructure in place via
W3C to host such discussions. The simplest thing would be to spin off a
sublist of www-rdf-interest@w3.org. Maybe now's the time to do that...

Short version of the above, taken from
	http://lists.w3.org/Archives/Public/www-rdf-interest/2000Aug/0019.html

> So... we should step back and ask for characterisations of what we
> want from an XML syntax for RDF. What are the must-haves?  What would 
> the goals be for any effort to provide a 'better' syntax? ie. what would make it
> better...? And then we need to ask who amongst us has time to commit to
> the projects sketched above. If the time, inclination and effort are
> there amongst W3C Members and the wider RDF world, then let's do it...

Dan

Received on Wednesday, 11 October 2000 23:19:26 UTC