Re: "show" attribute and fragments

Hello James,

At 01:01 PM 9/25/00 +0700, James Clark wrote:
>Section 5.6.1 of the XLink CR contains the following:
>
>"If the starting resource is a fragment whose location set contains
>multiple nodes, then an application should treat each node of the
>starting resource as an individual starting resource and traverse to the
>ending resource as a succession of single traversals in the style of a
>programming "foreach" construct."
>
>I can't understand what this means.  Is it saying that an arc from a
>starting-resource that contains multiple nodes is equivalent to a set of
>arcs one for each node?  That seems a totally bizarre interpretation.

The wording is, unfortunately, ambiguous; the recent joint Linking/XSL Task 
Force comments[1] suggested a clarification to the ending-resource version 
of this wording, but we probably should have done the same for the 
starting-resource version.

I think the WG was thinking of non-contiguous selections (such as "all the 
section titles in the document") rather than a contiguous stretch of nodes, 
or a single range.  We felt that it didn't really make a lot of sense to 
start traversal from several non-contiguous pieces of content all at once 
(an "aggregate" starting resource), particularly if they can't even fit on 
one screen at the same time.

(It's not so unusual for a single arc specification to create multiple 
arcs; you can already do that with an XLink arc-type element that points to 
non-unique resource labels.)

>Why are multiple nodes in a starting resource handled in a completely
>non-orthogonal way from multiple nodes in an ending resource?

The WG tried to figure out with the expected behavior would be, and modeled 
that.  We discovered, on doing surveys among ourselves and colleagues, that 
the expectations for the two ends were different.

>What happens when the starting resource includes ranges or points?

If we clarify the wording to limit it to non-contiguous selections, I think 
the ranges and points just fall out correctly, since any one range or point 
is "contiguous."

>In general, I was unable to understand from the description of the show
>attribute what the intended behaviour was when the ending resource was a
>fragment of a complete document.  Should it extract the fragment and
>display just that or should it display a window on the complete document
>with the document scrolled so that the start of the fragment is in the
>window?

We were deliberately trying to avoid dictating specific presentation 
behavior, and XLink does not dictate how much to show of the document that 
contains an ending subresource.  The recent work by the joint Linking/XSL 
Task Force[2] hopes to provide a few more details about how a 
system-builder might want to control these aspects of presentation (by 
supplying parameters for how much to display, where in the XML tree to 
start processing, and what stylesheet to use).  Do you think the XLink spec 
should be more specific about not dictating these things, or that it should 
actually dictate them?

In any case, of course we'll add your comments to the issues list.

         Eve

[1]
  http://lists.w3.org/Archives/Public/www-xml-linking-comments/2000JulSep/0170.html
[2] (W3C members only)
   http://www.w3.org/XML/Group/2000/08/NOTE-xml-link-style.html
--
Eve Maler                                          +1 781 442 3190
Sun Microsystems XML Technology Center    eve.maler @ east.sun.com

Received on Monday, 25 September 2000 11:15:37 UTC