Re: ISSUE-15 PROPOSAL: @version attribute in HTML5

On 07/08/2010 01:02 PM, Gregg Kellogg wrote:
> Okay, after reading your telecon notes, I'm not sure how to detect
> the use of a 1.0 vs. 1.1 parser. If a 1.1 parser MUST/SHOULD ignore
> @version, then it can't use it to know to switch to 1.0 mode. 

That's correct. With the latest change, an RDFa 1.1 parser would
interpret an RDFa 1.0 document via the RDFa 1.1 processing rules.

This approach is the same one that HTML5 takes - your processor does a
best effort on extracting triples from a given HTML document.

That is not to say that you can no longer run an RDFa 1.0 processor on
an XHTML+RDFa 1.0 document, you can still do that. However, if you have
a RDFa 1.1 conformant parser, it must use the RDFa 1.1 rules when
extracting triples from a document.

What we're attempting to do is resolve a couple of outstanding issues:

* @version isn't used consistently (OGP being the biggest example).
* @version can only be used by people with full control over the HTML
  document (think any CMS system, where you don't have full control
  over the entire HTML document).
* @version is not specified in HTML5 and there is active hostility
  towards versioning HTML5 documents.
* @version values differ between RDFa in XHTML5 and HTML5 and XHTML1,
  which makes implementing them in HTML polyglot documents not possible.
* Having multiple branches for parsing RDFa 1.0 vs RDFa 1.1 makes
  implementations more difficult to write and trickier to get right.

> Given
> that RDFa 1.1 is not 100% compatible with RDFa 1.0, I need a way to
> determine if a document is a RDFa 1.0 to generate correct triples
> (e.g., case-sensitivity of prefixes).

Let us know if there is a specific use case that this new change
affects. Generally speaking, we hope that this change doesn't affect how
RDFa is used in a negative way. We are making the change to make RDFa
easier to implement and less confusing between all languages.

The trade-off is that you may no longer be able to have a switch in your
processor to switch between RDFa 1.0 and RDFa 1.1 mode. That is both
good and bad for a number of reasons, but after a long discussion, the
group was unanimous in agreement that the added complexity of having a
switch doesn't surface in large positive benefits in real world RDFa
applications.

That said, if there is a use case that this approach makes impossible to
implement or difficult to achieve, let us know and we can open the issue
back up. Note: We only re-open issues if there is new evidence
presented. If we've already discussed the evidence presented in detail,
we can't re-open the issue.

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny)
President/CEO - Digital Bazaar, Inc.
blog: Myth Busting Web Stacks - PHP is Faster Than You Think
http://blog.digitalbazaar.com/2010/06/12/myth-busting-php/2/

Received on Friday, 9 July 2010 03:06:30 UTC