xml-dev - Re: XLink transformations

[Ben Trafford <ben@legendary.org>]

At 01:36 PM 7/18/00 +0200, Michael Kraus wrote:
>Now the question is: Is there any chance for the browser, given the
>linkbase from above, to know that there should be a link between books
>B1 and B3? And if, which elements should be used as source and target
>of the link? the <p> element or the <b> element? In my opinion,
>there's no general solution to this problem if the complexity of
>XPointer and XSLT is considered.

         Well, you've pretty much hit the head on the nail, there. I firmly 
believe that, in that instance, the engine performing the transformation 
(assuming it's XLink-aware) should pass an error message, or at least a 
warning that the transformation has caused a loss of information. Since 
XSL-FOs, by their nature, do not discuss merging XLinks in the 
transformative process, we have a bit of a quandary. People are aware of 
this problem, at least, but as I said before, my way to handle is with 
preprocessing, where two options are given: either the engine flags an 
error, or the engine does the merging itself. I'm partial to the latter 
option, but would like to see it standardized.

--->Ben Trafford