W3C home > Mailing lists > Public > public-rdf-in-xhtml-tf@w3.org > May 2005

Re: [HTML] Re: additional GRDDL editor

From: Ben Adida <ben@mit.edu>
Date: Thu, 19 May 2005 19:57:18 -0400
Message-Id: <fd1aeeb2da7096efefa7cfd2809bd06a@mit.edu>
Cc: <public-swbp-wg@w3.org>, "'HTML WG'" <w3c-html-wg@w3.org>, "Mark Birbeck" <mark.birbeck@x-port.net>, <public-rdf-in-xhtml-tf@w3.org>
To: Bjoern Hoehrmann <derhoermi@gmx.net>


Bjoern,

There are indeed technical reasons why RDF/XML cannot simple be used 
inside XHTML. I urge you to read documents like:

http://infomesh.net/2002/rdfinhtml/#embedNoValidate

which explain the validation issues with RDF/XML in XHTML.

There is another important issue. Even if RDF/XML validated nicely in 
XHTML, you would be forced to separate HTML content from the metadata. 
For example, if you wanted to indicate an article's author, you would 
have to both add some visible HTML byline and some unlinked RDF/XML to 
indicate the dc.author property. For maintenance and consistency, that 
is really awful.

RDF/A not only validates, it provides this ability to use existing HTML 
entities to define RDF properties. This is what I referred to as 
"merging the clickable and semantic web" in many prior messages and 
discussions. This important goal explains a lot of the design decisions 
behind RDF/A.

> This is an interesting remark; I thought it is pretty clear from the
> public-rdf-in-xhtml-tf archives that even people familiar with both
> RDF and HTML have trouble to understand when to use RDF/A's <link> or
> <meta> element, how to nest them or how to combine them with the many
> attributes to achieve the desired effect.

If you look at the latest discussions, these issues have been 
significantly cleared up. Of course, there were earlier confusions, but 
we worked through them in this public forum. We are now working on the 
last important discussion concerning bnode support. Your description of 
the state of our task force indicates that you probably haven't kept up 
with our latest discussions.

> <link> links to resources related to the document and <meta> is for
> document meta data and fake HTTP headers. Much of this is no longer
> true with RDF/A (and not widely known anyway) and beyond that I fail
> to see how RDF/A leverages much here.

I'm not sure I understand your point. We're trying to introduce new 
concepts into HTML, specifically the inclusion of RDF statements. This 
is a task that many people have been clamoring to see the W3C resolve. 
There's no way to make this happen without some amount of change. The 
<link> and <meta> elements provide the most natural entry points for 
these changes. However, if you'd rather not touch <link> and <meta>, 
you can use RDF/A attributes directly on other HTML entities.

My previous example of linking to an author becomes:

=======
This email was authored by <a rel="dc:author" 
href="http://ben.adida.net">Ben Adida</a>.
=======

To me, this looks like the best of both worlds. Clear RDF properties 
that truly respect the essence of what HTML has been and should be.

I have re-read your early message from November, which seems to mostly 
claim that RDF/A is much more complicated than RDF/XML. Certainly, 
there's an element of personal preference here, but given the direction 
that RDF/A has taken, and given the numerous simplifications it has 
undergone, I don't think those remarks apply anymore.

-Ben
Received on Thursday, 19 May 2005 23:57:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:14:59 GMT