Re: XHTML+RDFa 1.1 'the most recent version'...

Hi Olaf,

I'll try to take a shot at answering your questions...

On 03/31/2011 12:14 PM, Dr. Olaf Hoffmann wrote:
> just because I read it in the current draft as before similar
> formulations in other recommendations, what does it
> exactly mean, that in section '3. Additional RDFa Processing Rules'
> is mentioned:
> "If the value does not match a defined version, or there is no @version 
> attribute, then the processing rules for the most recent version of 
> XHTML+RDFa must be used."?

When writing the specification, we noted that there are a non-trivial
amount of web page authors that are not including the @version attribute
on the document. There is also a group of web page authors that do not
have access to the root element of the document to specify the @version
attribute. In order to make sure that these documents are still
process-able as RDFa documents, we had to assert a RDFa version in the
RDFa Processor (since the processors can operate both as RDFa 1.0
processors as well as RDFa 1.1 processors).

The "most recent version" language is intended to signal RDFa processor
implementers that (at this moment in time) RDFa 1.1 should be used to
process documents that do not contain a @version attribute. In the
future, like when RDFa 2.0 is released 10 years from now, that RDFa 2.0
should be used to process documents that do not contain a @version
attribute.

That is to say, if the author isn't specific about which version should
be used to process the document, the latest version available to the
processor should be used. This may mean that if there are 3 versions of
RDFa that have been released: 1.0, 1.1 and 2.0, and there is a processor
that only understands 1.0 and 1.1, and there is a document that doesn't
specify @version - that the processor is still conformant if it uses
RDFa 1.1 to process the document. This is not ideal - however, it's the
best approach that the community could find to the problem of selecting
a set of processing rules to use when processing a document without
@version.

> More precisely:
> a) 'does not match' for whom - the author of the document, when writing it or
> for the reader/interpreter of the document when reading it?

"Does not match" for the RDFa processor processing the document.

> (Implication: an old interpreter may not know the newest version;
> and obviously an old document was not written taking into account a version,
> that is newer than that when the document was written)

That is correct. The old implementation is still conformant if it uses
the latest version that it knows about. Yes, there are issues with this
- like what happens if RDFa 2.0 is not backwards-compatible with RDFa
1.1? That is something for the RDFa 2.0 group to be wary of and ensure
that they have a good answer.

> b) 'most recent version' from which point of view? The creation date of the
> document? The publication date (last update) of the document? The 
> publication date (last update) of a the current viewer of the document?
> The current date? If one of this - is the most recent version the newest 
> version published before such a date?

None of the above. 'most recent version' from the point of view of the
RDFa Processor. That is, the RDFa processor can't possibly know about
future versions of RDFa - it can only know about past versions of RDFa.

Does all of that make sense?

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny)
President/CEO - Digital Bazaar, Inc.
blog: The PaySwarm Vocabulary
http://digitalbazaar.com/2011/03/31/payswarm-vocab/

Received on Tuesday, 5 April 2011 02:01:59 UTC