RE: [HTML] Re: additional GRDDL editor

Hi Bjoern,

Don't get me wrong...I am not at all defending the lack of communication
around the issues, although oftentimes answers to your questions had not
come immediately because we didn't have one, and it therefore triggered work
to find one. However, your previous statement wasn't about speed of
response, but an assertion that the whole work was being dropped. Hopefully
that issue is now clarified.

As to the SVG work, I think that is a red herring. The presence of an
element that can hold arbitrary metadata is no solution at all to the
problem. I think the combined talents of those interested in the RDF-in-HTML
story would have gone for that one years ago if it was viable ;). The short
explanation for why is "RDF/XML"...but I'll give a longer answer since as
you rightly say, the *motivations* for RDF/A have not had enough of an
airing.


THE PROBLEM

The situation we were trying to address was of course the relationship
between RDF and HTML, but we have to go down a level to get at why. We
recognised that in the HTML world, authors were constantly putting metadata
into their documents, often without even thinking about it. We don't just
mean "DC.creator" in the <meta> tag, but @class="navbar" and
@class="invoice" on <div>s, and so on. HTML documents therefore contained a
rich source of semantic information that was going to waste because there
was no way to get to it.

The RDF-world come at metadata in HTML documents from a different
stand-point; they are experts in RDF and so they know how to put metadata
into documents using <link>, or embedding RDF/XML inline, or whatever. These
techniques have been around for years, are well understood by RDFers, and in
passing I would point out that the SVG metadata element falls into that
category -- great for the experts.

However, everyone knows that pretty much no other community outside of the
RDF one has taken up RDF/XML. HTML authors certainly haven't, so the real
problem to solve is how the RDF-world can get access to the metadata that
the HTML authors provide, when the HTML authors are never going to provide
it in a form that can be processed by the RDF-world (i.e., RDF/XML). It's
sobering to recall that this issue was well summarised by Leigh Dodds over
*five* years ago [1]. You'll also find a discussion of the problem in my
early proposals that the XHTML 2 metadata story should embrace RDF, briefly
here [2], and in more detail here [3].

So the long and the short of it is that the abstract notions of RDF are
great, but the only XML syntax available is RDF/XML, and it has not gained
widespread support. Once again in passing, I predict that the SVG community
is no more likely to put RDF/XML in their documents than the HTML community,
and so we are no more likely to get svg:rects marked up as hospitals or
svg:circles as a bus-stop, than HTML authors are to mark up a div as a
restaurant. (Of course they will put basic stuff in like author, but so do
most HTML authors.)


THE SOLUTIONS

So various attempts were made to try and provide easier ways for the author
to add their metadata, so that it could later be mined.

One solution is to encourage authors to use 'patterns' and then for some
process to be run on their documents, find these patterns, and make triples
from them. A 'pattern' might be that the metadata is indicated by the use of
@class set to some value, or the use of <meta>, or even some element by
name. That's the approach taken by GRDDL.

However, we felt that it might be possible to come up with a different
approach that leverages the HTML author's understanding of <meta> and
<link>; if we could codify the relationship between <meta> and <link> and
RDF, as well as making them more powerful for the RDF-world, we may get one
'language' that serves both communities. RDF/A represents the 'workings out'
of this work, before folding it into XHTML 2.


CONCLUSION

Personally, I feel it is no exaggeration to say that the so-called Semantic
Web is a non-starter without a solution that bridges this gap, and the
metadata story in XHTML 2 is an attempt to do this. The latest draft of
RDF/A [4] does not fully represent the final answer because its main purpose
was to clarify issues for XHTML 2. But I feel we are extremely close to a
resolution of this long-standing issue.

Regards,

Mark


[1] http://www.xml.com/pub/a/2000/03/08/deviant/
[2] http://www.formsplayer.com/notes/xhtml-meta-data-02.html
[3] http://www.formsplayer.com/notes/xhtml-meta-data-03.html
[4] http://www.w3.org/MarkUp/2004/rdf-a.html  

Mark Birbeck
CEO
x-port.net Ltd.

e: Mark.Birbeck@x-port.net
t: +44 (0) 20 7689 9232
w: http://www.formsPlayer.com/
b: http://internet-apps.blogspot.com/

Download our XForms processor from
http://www.formsPlayer.com/ 

> -----Original Message-----
> From: public-rdf-in-xhtml-tf-request@w3.org 
> [mailto:public-rdf-in-xhtml-tf-request@w3.org] On Behalf Of 
> Bjoern Hoehrmann
> Sent: 19 May 2005 01:46
> To: Mark Birbeck
> Cc: public-rdf-in-xhtml-tf@w3.org; public-swbp-wg@w3.org
> Subject: Re: [HTML] Re: additional GRDDL editor
> 
> 
> * Mark Birbeck wrote:
> >> You can find comments on XHTML 2.0's meta data module and RDF/A in 
> >> the public-rdf-in-xhtml-tf and www-html(-editor) archives; I would 
> >> not be surprised if the HTML Working Group decides to reject the 
> >> SWBPD's re- quest to adopt RDF/A but I am unaware of publicly 
> >> available information to this effect.
> >
> >I don't understand how you could arrive at that point...
> 
> Considering that the HTML Working Group did not publish anything since
> 22 July 2004 and is generally known to ignore most feedback, 
> it is not surprising that it did not receive much feedback on 
> RDF/A yet. E.g.,
> 
>   http://lists.w3.org/Archives/Public/www-html/2004Sep/thread.html#5
> 
> however discusses parts of the proposal (and the lack of 
> responses to issues from the HTML Working Group...) E.g. in
> 
>   http://lists.w3.org/Archives/Public/www-html/2004Sep/0015.html
> 
> a reviewer seeks clarification on how the meta data features 
> interact with CSS or in
> 
>   http://lists.w3.org/Archives/Public/www-html/2004Sep/0013.html
> 
> a reviewer seeks clarification on how the property="" 
> attribute can serve as a replacement for blocklevel and 
> inline-level elements as done for code/blockcode, 
> quote/blockquote, div/span, etc. In
> 
>   http://lists.w3.org/Archives/Public/www-svg/2004Sep/0100.html
> 
> a reviewer fails to see the need to re-invent another 
> metadata language for XHTML 2.0, and as pointed out earlier,
> 
>   http://www.w3.org/2005/04/Tiny12DoC.html#T015
> 
> the SVG Working Group does not consider the XHTML 2.0 meta 
> data module a superior approach to their current 
> widely-adopted approach to simply use RDF/XML inside SVG 
> graphics. I've repeatedly asked for clarification on what the 
> requirements are (and many people asked clarification on the 
> requirements for and design goals of XHTML 2.0 in general), 
> how RDF/A meets them, etc. and have pointed out many 
> technical issues. In fact, it seems that even the few RDF/A 
> requirements I could find in the draft are
> 
>   http://lists.w3.org/Archives/Public/www-html-editor/2004OctDec/0042
> 
> not met at all (and the issues are further explained in
> 
>   http://lists.w3.org/Archives/Public/www-html-editor/2004JulSep/0135
>   http://lists.w3.org/Archives/Public/www-html-editor/2004JulSep/0125
> 
> and other messages). As the HTML Working Group so far ignored 
> all this feedback and as I consider any "collection of 
> attributes for layering RDF on XML languages" to have the 
> same or similar fundamental flaws, i.e., RDF/A cannot be 
> fixed to meet its own requirements, it seems most reasonable 
> to expect the HTML Working Group to reject this proposal.
> 
> So this is more a matter of the HTML community failing to 
> understand how the HTML Working Group could consider adopting 
> RDF/A in the first place.
> 
> >It's whole purpose was to try to find a solution to issues that have 
> >plagued the relationship between RDF and HTML over the 
> years. We're all 
> >pretty excited about this, so it is certainly not going to 
> be dropped!
> 
> Well, in
> 
>   http://lists.w3.org/Archives/Public/www-html/2005Jan/0089.html
> 
> I've asked you to convince "us" that RDF/A constitutes an 
> adequate solution. This would naturally include clearly 
> stating the problem, explaining in detail how XHTML 2.0 
> addresses forward compatibility and XHTML 2.0's extensibility 
> model as requested in e.g.
> 
>   http://lists.w3.org/Archives/Public/www-html-editor/2004JulSep/0042
> 
> and a discussion on how RDF/A is superior to other meta data 
> proposals.
> Even though it is unlikely that this would convince me, there 
> is a chance that some of the technical issues get resolved 
> and some other reviewers retract their objection to RDF/A 
> which would help the HTML Working Group when asking the 
> Director to advance the document on the Recommendation track.
> --
> Björn Höhrmann · mailto:bjoern@hoehrmann.de · 
> http://bjoern.hoehrmann.de Weinh. Str. 22 · Telefon: 
> +49(0)621/4309674 · http://www.bjoernsworld.de
> 68309 Mannheim · PGP Pub. KeyID: 0xA4357E78 · 
> http://www.websitedev.de/ 
> 
>

Received on Thursday, 19 May 2005 09:42:46 UTC