Re: CFC: Basic Roles Proposal

Hi, Ray–

On 9/1/15 1:04 PM, Denenberg, Ray wrote:
> Hi Doug;
>
> From: Doug Schepers [mailto:schepers@w3.org]
>> I don't think there's any meaningful difference between an
>> annotator's motivation in creating a particular body or target, and
>> the functional role that body or target plays.
>
> So in the example I cited, (3.1.7)  there are three roles (1)
> comparing (2)  antecedent (3)  subsequent.
>
> The intent of the annotation is to "compare" two passages. I would
> say the annotation is motivated by "comparing".
>
> The comparison is "The first passage is a clear derivative of the
> second".    The roles on the two target resources support the
> comparison by indicating which is the first and which is the second.
> I would say those are the roles that those two resources play in the
> comparison.
>
> Do you really not think there is a meaningful distinction?

You raise a good point.

I suggest that the problem lies not in a meaningful distinction between 
"motivation" and "role", but arises because Rob's example extends the 
"comparing" motive, giving it an implicit structure, where the 
"antecedent" and "subsequent" motives have a sub-relationship to 
"comparing" (which is another formulation of the extension problem Rob 
was talking about). Another UA that encountered those custom motives 
wouldn't know about this implicit relationship, or what to do with it.

To complicate it further, it's easy to imagine that there are additional 
bodies on the same annotation, including motives like "highlighting", 
"tagging", "commenting", "questioning" or so on. Do all of these 
similarly have a sub-relationship to the "comparing" motive of the root 
Annotation?

How can this be resolved, while preserving the model and allowing for 
natural extensions?

One way I could see that is compatible with my characterization of Rob's 
proposal would be to have the "comparing" motive on a body that explains 
the comparison, while the "antecedent" and "subsequent" motives are on 
the targets. Different UAs need no special knowledge to process such 
structures normally, or even to display the motives to users; UAs that 
understand the extensions can process it the same way but infer 
something a bit more than a generic UA would.

This feels clean to me. How does it strike you?

Regards–
–Doug

Received on Tuesday, 1 September 2015 17:21:06 UTC