Re: [xml-dev] (In)Validate My Assumptions on Linking.

[Dave Pawson <davep@dpawson.co.uk>]

Jirka Kosek wrote:
>
>> Don't get hung up on my syntax here.  The point is, the concepts of
>> presentation and relationship building need to be separated.
>
> Not always, and in many situations you can't separete these two, 
> because nature and destination of link usually depends on media for 
> which you are styling document. If you have link from place A to place 
> B in your XML file and you are then presenting this XML as two output 
> formats (e.g. PDF and HTML) actual links (destination URI) will be 
> completely different and depending on output media used.
>
> For example, link <link xlink:href="intro.xml"/> can be converted to 
> <a href="chap01.html">Introduction</a> for HTML output. For PDF 
> output, destination URI could be something like manual.pdf#page=3.
Isn't that an application of stying Jirka?
Perhaps if we listed the various properties of links, one might be 
'dereferencable' (or whatever the right
word is). If a target is page 3 in a PDF document then it won't be. 
Perhaps dereferencable given conditions
A, B and C?
  I wonder how this might be expressed?



>
> One of flaws of XLink + XPointer was assumption that link traversal 
> would occur on original XML sources. But because of lack of support 
> for direct XML rendering in applications, you must convert XML to HTML 
> or PDF if you want to deliver it to users. This means that you must 
> map original XLinks to links between HTML or PDF files.
>
>

Is this the styling aspect of links? If you know about the link (from 
the source semantics)
and within the styling application you know about the target media, then 
it should be
possible to work out the right thing to do?

Then from that, what other link properties are essential, useful, 
conditionally useful?




regards 

-- 
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk