Re: Comment on XPointer and XLink: Ranges

At 8:03 PM -0000 12/29/99, Rick JELLIFFE wrote:
>Comment from W3C member Academia Sinica
>- Rick Jelliffe
>
>I apologize if this has not made the last call deadline.
>
>Ranges should be removed from XPointer and placed in XLink.
>
>The reasons:
>    1) ranges complicate XPointer implementation;
>    2) ranges complicate explanations of XPointers;
>    3) a pointer should point to a node or a point;
>    4) a range is similar to an arc: the start and end have an affinity
>to "from" and "to"
>
>For similar reasons, I am not keen on multiple fragment specification:
>it seems a shorthand for an extended XLink of some kind.

This idea has been discussed multiple times; it seems to me your note does
not provide new reasons, or show some way to address at least some of the
many known problems that moving ranges into xlink would crate.  But
specifically:

1) granted.

2) How is it a net improvement to "complicate explanations" of XLink
instead? Especially when (see (4)) the complication would be larger under
your plan.

3) Why? This merely assumes your conclusion as a premise.

4) They may have a vague "affinity" , but they are clearly not the same thing.
As pointed out in prior mail to this list, and in the discussions in the WG
and IG, most properties of arcs do not apply to ranges.

   You don't typically traverse between the ends of a range, though
   that is the primary meaning of an arc.

   The ends of ranges do not have meaningful roles or titles as do the ends
   of arcs.

   If we put ranges into XLink, we immediately must define recursive links,
   since linking to a range (the most common anticipated user actions) requires
   making a link to a link.

   Since links can't be put into URI fragment identifiers, on your plan there
   is no way to refer to a range AT ALL via a URI. That seems fatal.

   XLink is about creating links, XPointer is about specifying locations. A
   range is clearly a location (for example, it is meaningful without even
   having a link around, just as URIs is meaninfgul not just when in <A>).

S

Steven_DeRose@Brown.edu; http://www.stg.brown.edu/~sjd
Chief Scientist, Scholarly Technology Group, and
   Adjunct Associate Professor, Brown University

Received on Tuesday, 4 January 2000 11:34:23 UTC