Re: Level 2 and disjoint selections

At 08:41 1998 12 29 -0800, Gabe Beged-Dov wrote:
>Paul Grosso writes:
>>What do you mean by "range selections which is one of the capabilities
>>of XPointer?"
>
>Both the string and span locations terms of XPointer "select" ranges.

Okay, you seem to be using the term "range" somewhat loosely.
Most XPointer location terms address parts of an XML tree that
may contain multiple data characters and/or multiple nodes in
the XML tree.  I thought you might be asserting that a stated
goal of XPointers was to be able to address any DOM range, and
I would take issue with that.  I now understand you not to be
asserting that.

>
>>Where do you get the equation of range selections and
>>XPointer capabilities?
>
>see http://www.w3.org/TR/WD-xptr#3.7

I did. The word "range" is not used in section 3.7 at all.  In fact,
I scanned the entire document for the word "range," and it is never 
used in the DOM sense.  (It is used three times:  one in reference
to a number being out of range and twice to all younger or older
siblings of an element.)

>>I would oppose XPointer supporting disjoint selections.  I believe
>>XPointer should be basic addressing into XML documents, and things
>>such as disjoint selection could be built on top of basic addressing.
>>Put another way, I'd be fine with XLink allowing the specification
>>of disjoint selections, but it might take a couple XPointer specs
>>(remember, XLink would support multiple end points) to describe this.
>
>I agree with you. I would very much like to see a level 0 version of
>XPointer that is only about addressing.

Sounds like we agree on what I think is the key point here.

paul

Received on Tuesday, 29 December 1998 12:45:41 UTC