Re: Comments on GRDDL (using 3rd-party XML schemas with GRDDL) [OK?]

Dan Connolly wrote:
> On Tue, 2007-07-24 at 13:24 -0600, C. M. Sperberg-McQueen wrote:
>   
>> On 24 Jul 2007, at 13:14 , Chimezie Ogbuji wrote:
>>
>>     
>>> ...
>>> On Tue, 2007-07-24 at 12:12 -0600, C. M. Sperberg-McQueen wrote:
>>>       
>>>> ...
>>>>
>>>> It seems to me to be a surprising and unfortunate lurch toward a
>>>> closed-world assumption, to require that the RDF representation of a
>>>> document originate with, or be authorized by, the creator of the
>>>> document.  I think that's a pretty severe design error, and one that
>>>> surprises me, coming from the Semantic Web activity.
>>>>         
>>> We should be careful not to conflate closed-world assumption (which is
>>> specifically a feature of inference - GRDDL has nothing to do with
>>> inference) with functional data transformations.
>>>       
>> Except, of course, that the logic of the design, as I have
>> understood it so far, relies on an inference something like:
>>
>>     The author did not point to this transformation from
>>     the document, nor did the namespace owner point to this
>>     transformation from the namespace name.
>>
>>     Therefore, the transformation is not a faithful rendition
>>     of the author's intentions.
>>     
>
> No; only that if the author didn't point directly nor
> indirectly to the transformation, we don't know whether
> it's a faithful rendition or not.

I'm not sure I've grasped the concept of faithful renditions.

I assume that most document authors have no idea what GRDDL 
transformations may be tucked away in a schema or other document 
associated with a namespace, and would be unlikely to read the code to 
determine what inferences may be generated. Do you assume otherwise? If 
not, in what sense are these transformations said to mirror the author's 
intent? Is it not primarily the document per se that best reflects the 
intent of the author, since that's the only thing we know the author 
actually looked at?

I assume that may different parties might license different sets of 
valid inferences from a given schema or document. What determines which 
of these inferences are "faithful renditions"? I understand the 
mechanics of how these transformations are found, but I'm trying to 
understand the user model.

Jonathan

Received on Friday, 27 July 2007 18:24:17 UTC