W3C home > Mailing lists > Public > public-grddl-comments@w3.org > July to September 2008


From: Harry Halpin <hhalpin@ibiblio.org>
Date: Sun, 03 Aug 2008 23:15:15 +0100
Message-ID: <48962DF3.8010701@ibiblio.org>
To: Ian Hickson <ian@hixie.ch>
Cc: public-grddl-comments@w3.org

Ian Hickson wrote:
> Hello,
> In the development of HTML5 we have made a simplification to the language 
> that should help GRDDL usage be somewhat simpler: We have made the 
> profile="" attribute on <head> unnecessary to declare new metadata 
> keywords.
> This means that GRDDL can now be simplified to not require an explicit 
> declaration of profile="http://www.w3.org/2003/g/data-view" before making 
> the rel="transformation" keyword meaningful for GRDDL.
I would encourage GRDDL implementers to think and talk about this and, 
if so inclined, try implementing it, and I'd recommend everyone keep 
their comments on this list, and at the end of the week, and if there's 
no comments by the end of the week,  I'll put my personal opinion 
forward in more detail. But in brief...

Unfortunately GRDDL is already a REC, but if implementers all manage to 
get away from using profile, when HTML5 becomes a REC perhaps there 
would be a possibility of an errata or W3C Note on this, with a pointer 
to relevant implementations/test-cases.

I assume the concern would actually be from the microformat world, since 
this puts all microformat "names" (or any metadata keyword it appears) 
on various elements in the same global (or null, depending on how you 
look at it) namespace. I do think this rather pragmatic, as profile is 
often not used in microformats in the wild. The question is one of 
upholding norms or codifying existing practice, even if existing 
practice may have problems down the road.

I am on vacation next week, e-mail checking will be erratic at best.


> It also means that GRDDL can be simplified to not require that 
> transformations look for a profile="" before making the class names 
> supported. For example, an hCard document now no longer needs a profile="" h
> attribute declared, instead the class="vcard" keyword always has meaning. 
> Thus, a GRDDL transformation for hCard no longer needs to be predicated on 
> the profile="" attribute containing the right incantation.
> This is only relevant for HTML5 documents, and HTML5 won't be a REC for 
> many years yet, so there is no rush here. This change doesn't affect HTML4 
> documents. (Although, I would recommend making the change for HTML4 
> document as well, since many authors already omit the profile="" attribute 
> anyway.)
> Cheers,
Received on Sunday, 3 August 2008 22:15:56 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:55:03 UTC