Re: PROPOSAL from telecon on XMLLiteral

Hi Jeni,

I'm apprehensive about changes to existing specs, too, which is why I've
tried to take this one slowly, making sure we're taking into account
what implementors are actually doing.

I've checked with both the Yahoo and Google teams, and it seems fairly
clear that they would much prefer RDFa to function this way. In fact, I
think at least one of their parsers will force plain literals no matter
what :)

Also, just to be extra clear, the *only* markup whose interpretation
would change, according to the spec, is when XML content is marked with
@property but not @datatype.

So, my thinking at this point is that this change may simply represent
the reality of implementations out there. That's why I'm not sure a true
versioning is needed.

-Ben

Jeni Tennison wrote:
> Manu,
> 
> On 28 Jul 2009, at 06:16, Manu Sporny wrote:
>> Now that you know that we're not taking this change lightly and that we
>> would ask for community feedback before making the change, do you still
>> feel as apprehensive about the change as you did before?
> 
> To an extent. I would view it as an acceptable change if
> 
>   (a) it was accompanied by a version change in RDFa (ie instead of
> version="XHTML+RDFa 1.0", people would put version="XHTML+RDFa 2.0" or
> version="XHTML+RDFa 1.1" in their documents in order for the new rule to
> take effect)
> 
>   (b) that new specification included a section on backward-incompatible
> changes that listed this as one of those changes
> 
>   (c) the location of the DTD changed, such that existing documents did
> not automatically get updated to this new version
> 
> I think it will be very hard to ascertain whether people are generating
> XMLLiterals now by accident (ie they add RDFa to their documents but
> don't check the RDF that gets generated to make sure it's what they
> intended) or on purpose (ie they author it according to the RDFa 1.0
> specification and the XMLLiterals are exactly what they do intend), but
> it wouldn't hurt to get more data about how (much) XMLLiterals are used
> in the wild.
> 
> I don't have any objections to the change from a publishing or parsing
> standpoint. For processors that wish to continue to support RDFa 1.0, it
> will obviously add a little extra complexity to check the version before
> determining which rules to use. Which is why procedurally, all I'm
> asking is that backwards-incompatible differences are kept to a minimum,
> with RDFa's development being bundled into a few, infrequent releases
> rather than multiple frequent ones.
> 
> Thanks,
> 
> Jeni

Received on Wednesday, 29 July 2009 21:36:40 UTC