On Mon, 15 Oct 2007, Mark Birbeck wrote:
I don't see how you can interpret my comments as saying that GRDDL
wasn't useful:
I think there was also the word "flaky" :) I know lots of companies doing
data integration via XSLT, so I'd hestitate to call it flaky.
GRDDL is a necessary hack to allow legacy mark-up to be made
'semantic'.
Unless you think I don't want legacy mark-up to be 'made semantic', I
don't see how you can imagine I mean 'hack' in any sense other than
the one you have used. :)
With great respect to everyone involved, I don't think it's a case of
"to applaud or not to applaud"; I'm sure we're all working very hard,
but hopefully we also have thick skins. :) My feeling is that we've
had RDF for a long time now, so to simply 'applaud' any additional use
of RDF as if it is _by its very existence_ contributing to a wider
goal is a little meaningless.
Disagree. The value of RDF will likely on increase as the proportion of
RDF data on the Web increased, see Open Linked Data Project.
To illustrate, Google already indexes a wide array of documents that
it finds when crawling; would it be easier to convince Google to index
the metadata in ODF and XHTML if it used the same handful of
attributes, or would it be easier to convince them if we said 'here is
the GRDDL transform for document type A, here is the GRDDL transform
for document type B', and so on.
That's one use-case, but they could always look for metadata in ODF and
XHTML using whatever format they feel is widely deployed. That's
definitely what Technorati was doing with microformats, and Yahoo! I
believe.
Obviously we don't know the answer for sure, but my comments are
motivated by the belief that (a) a generic solution could be adopted
more quickly, and (b) it would be adopted more quickly if the big
standards bodies were seen to be coordinating. :)
Agreed.
The problem is that there is *no* standard for
embedding RDF in-line in generic XML vocabularies like ODF, as RDFa is
aimed at XHTML.
RDFa was always aimed at 'mark-up languages' more generally. It just
so happens that its first major outing was in XHTML 2, and more
recently the focus has been on XHTML 1.x.
Hmm...that's fine, but the world at large doesn't know this, the SWD WG
charter doesn't know this[1], and there's no editor's drafts etc. I can
easily find with RDFa in plain vanilla XML examples. So I don't think you
can blame ODF for not being sure. So, yes co-ordinate, but be careful
about posing RDFa as the solution, particularly at this stage in its development.
However, hopefully RDFa and ODF can co-ordinate,
and regardless, and maybe Fabien's GRDDL transform from RDFa in XML to RDF
could be even be used with ODF documents if they adopt a RDFa syntax. I'd
have to look at the XSLT carefeully, and if ODF has to develop their own
GRDDL they could use Fabien's as a starting point. I think this is the
latest URI, but I'd ask Fabien personally [2].
[1] http://www.w3.org/2006/07/swdwg-charter
[2] http://ns.inria.fr/grddl/rdfa/
> >> I think ODF is on the right track here, and even if
ODF >> and RDFa converge on a sort of common syntax for doing this...
But this is the odd thing here...the little bit of syntax that I've
seen seems very close. It would almost be better to be completely
different. :)
... I have no
doubt that a simple XSLT embedded at the namespace doc for ODF that allows
one to extrtact the inline ODF RDF into RDF/XML will be very, very useful
so that other RDF processors can access this inline meta-data. I imagine
that would be not difficult and hopefully someone in ODF can write that
script quickly. And that, is GRDDL :)
Of course...but I guess that brings us back to the Google question. :)
Google, being a company based on profit, will likely adopt whatever is
used in the wild with lots of data, regardless of W3C standardization.
That could be RDFa, it could be microformats, it could be something
completely different. Regardless, I do agree co-ordination could help, but
so could a more respectful tone of voice and a lack of presumptions by all
parties, including RDFa and GRDDL advocates. After all, they all work
together quite nicely, which is to the credit of all involved.
Regards,
Mark
--
--harry
Harry Halpin
Informatics, University of Edinburgh
http://www.ibiblio.org/hhalpin
