W3C home > Mailing lists > Public > public-rdf-in-xhtml-tf@w3.org > March 2008

Re: RDFa test case #1 missing @profile?

From: Norman Walsh <ndw@nwalsh.com>
Date: Thu, 06 Mar 2008 16:03:03 -0500
To: "Mark Birbeck" <mark.birbeck@x-port.net>
Cc: "Shane McCarron" <shane@aptest.com>, RDFa <public-rdf-in-xhtml-tf@w3.org>
Message-ID: <m23ar3h8fs.fsf@nwalsh.com>
/ "Mark Birbeck" <mark.birbeck@x-port.net> was heard to say:
| Hi Norm,
|
|> I get the impression that no attempt
|>  to persuade the working group to add a statement along the lines
|>  "In order to be interpreted as RDFa, a document MUST have {some explicit
|>  marker}", no matter how passionate, would succeed.
|
| There are two separate issues being raised here:
|
|  * whether a document can only be parsed for RDFa if there is some indicator
|    in the document that it contains RDFa;
|
|  * whether that indicator should be @profile, a doc type, or something else.
|
| At first sight insisting on an indicator seems to be a reasonable
| goal, and surely it couldn't break anything if we try to fulfill it?
| Except the problem is that when chemists blog about their experiments,
| and place those chemical symbols in their mark-up, the RDFa won't be
| parsed by anyone browsing to their blog, unless the author has full
| control over the web-page and can put the doc type or whatever into
| the mark-up.

Fair enough.

| That unfortunately leaves us no further on than we were 4 or 5 years
| ago, and makes the semantic web a tricky thing to realise.

Perhaps.

| To publish RDF you currently have to set up a triple store, or create
| RDF/XML files that could be linked to from your web-pages (and
| probably do content negotiation), or set up some additional XSLT pages
| on your server to do GRDDL transforms, or edit @profile in your
| documents to point to someone else's GRDDL transforms.
|
| But the HTML publishing revolution took off when people stopped having
| to configure servers, or edit and upload entire HTML pages, just to
| add a comment about what they had for their tea.

Ok, but you're not really going to reach the folks who create web pages
about what they had for tea, are you? They're filling in web forms and
using WYSIWYG-ish editors. It seems to me that the systems that don't
provide enough control to add some sort of "opt-in" marker are equally
the systems that won't allow RDFa.

As it stands today, my own blog comment system, for example, discards
extra random attributes.

Has anyone done a survey of where RDFa will actually be allowed
without any mods from the hosting providers?

| So the question is why I shouldn't be able to publish RDF as easily as
| I can write a blog, or use my CMS, or post to Twitter, or send an
| email? In short, the core tenet of XHTML+RDFa is to leverage the HTML
| publishing model, but use it to publish RDF. But we can only do that
| if we don't insist that authors must have control over the entire
| document in order to take part.

Yes, it should be easier to publish semantic web content. But the
approach that RDFa takes imposes a risk on the larger ecosystem that
concerns me.

| Which is why the RDFa taskforce took the approach that, whilst we
| _allow_ for an RDFa indicator, and even suggest that it's used, we
| don't make it mandatory.
|
|>  You may record that the submitter is satisfied, at least for the
|>  moment, albeit very reluctantly, with rejection of his request for a
|>  change. And that he's taken an "I told you so" chip which he may play
|>  later.
|
| And with respect... ;)
|
| I am totally and utterly convinced that if we *insist* on a doc type
| we'll kill this whole thing dead in the water.

Yeah, well, I agree with you there. The doc type is a really bad idea.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | Our culture peculiarly honors the act
http://nwalsh.com/            | of blaming, which it takes as the sign
                              | of virtue and intellect.--Lionel
                              | Trilling

Received on Thursday, 6 March 2008 21:15:18 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:01:55 UTC