Once more on RDFa and HTML [was Re: RDFa reliance on namespace declaration]

Hi Karl,

> > Much of RDFa will work in both HTML and XHTML. However, if you want to
> > make statements that have global relevance (i.e., use more of RDF)
> > then you will need to use namespaces, which of course requires an
> > XML-based language.
>
> When I sent this
> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2006Jun/0022
>
> You replied with
> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2006Jun/0024
>
> which is very limited and nothing new with what HTML 4.01 already
> does. You said it yourself
>
>         [[[
>         But that does not affect the key point which is
>         that RDFa's use of @rel and @rev is *already standard*,
>         and what's more *preserves* rather than overriding
>         that standard.
>         ]]]
>
> It's not RDFa it's simple HTML 4.01, or I have missed something.

I know I said it myself...that's the whole point...I was trying to
show how RDFa *builds on* HTML. But I am saying more than that; I am
pointing out that HTML already has semantics for the @rel and @rev
attributes, and RDFa leverages them. Other mark-up proposals appear to
be suggesting doing the same, yet in reality overload the @rel/@rev
attributes to include more than just a plain predicate. This is in
effect just using HTML attributes as 'carriers' for whatever data you
want, and is to ignore the semantic hooks that HTML has had for a long
time.

(An example being the use of things like "schema.cc", in eRDF, which
requires parsing in order to understand what it does, and is not a
'predicate'.)


> After you gave examples with  "about" attribute that is not valid
> HTML 4.01. Step 3 and 4 are not usable in HTML 4.01. You said it:
>
>         [[[
>         These steps take us out of standard HTML territory,
>         but are easily added with XHTML 1.1 modules.
>         ]]]
>
> It's not easy for simple Web authors :)

I assume you don't mean that web authors are simple. :)

But seriously, adding an @about attribute for those authors who want
to make statements about other resources is not difficult at all...we
don't ask the authors to produce the schemas, we produce them for
them!. And if they don't want to make statements about other resources
then they don't need to learn about @about...you can't currently do
that anyway--other than for simple cases, by using @rev--so we're not
forcing something on people.


> > Ian's point is that any XML-based language should not use QNames
> > within the *content* since the namespace prefixes can be lost or
> > changed. I've shown that in general this is  either not the case, or
> > it happens in certain circumstance that a stylesheet author can avoid.
>
> class="dc:title" is not a namespace. It's just practical :) I have
> used it in the past in my own html. And it's easy to style with
>
>         .dc\:title {}
>
> but certainly awkward for many people, then prone to errors.

First, CSS is a different issue. There are already many things that
CSS can't do, but it will increasingly support XML.

Second, RDFa doesn't *require* the use of @class.


> > In addition, it's worth pointing out that Ian's "copy-and-paste" blog
> > won't actually solve the problem, since we could have an XSLT file
> > that uses a @rel="schema.xx" statement that contradicts a different
> > "schema.xx" already in the source document. Of course to resolve that
> > we'd have to define a scoping mechanism...but we seem to be
> > reinventing XML namespaces.
>
> Agreed that is the reinvention of the namespaces. BUT as you said
> "XML namespaces", HTML 4.01 is hardly XML ;) There are two sides of
> this about this one which is about technology and another one which
> is about Web community.
>
> In terms of Web community, I often heard people who were impatient
> about using RDFa if it could be used in HTML 4.01/XHTML 1.0.
> Another side of the Web community will say, let's move to XHTML 2.0
> and then have the full potential of XML.
>
> The problem here is that there is not only one community online.
> Let's say that the community of Desktop Web browsers/services with
> APIs/Weblogs etc is more in favor of HTML 4.01/XHTML 1.0 and maybe a
> successor like Web App 1.0, when the community of Mobile, Heavy XML
> users are attracted by the benefits of XML. :)
>
> Not simple to solve.

I agree, 100%...there are many communities, and of course problems are
tricky to solve--the RDF community started discussin the issue of RDF
in HTML over 10 years ago and didn't solve it to everyone's
satisfaction.

But that's why we need to be clear about what problem we're acually
trying to solve, and to be very clear when there really is a problem.
I do not rule out adopting some mechanism other than namespaces as a
scoping mechanism in RDFa (indeed I proposed such a thing a long time
ago as part of my work on CURIEs). BUT...the suggestion that we *need*
to change from using namespaces as a scoping mechanism due to XML
tooling support is in the main wrong.

There may be other reasons to move away from XML and namespaces
(although with respect to all the smart people on this list, that's a
very big issue with lots of ramifications, so I'd tread carefully
there!), but XSLT is definitely not one of them.

Regards,

Mark

-- 
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/

Received on Monday, 19 June 2006 09:48:32 UTC