Re: [xsl] would like to simplify my XSLT

[Abel Braaksma <abel.online@xxxxxxxxx>]

David Carlisle wrote:
> no even if the attribute had the same name, it would be a different node
> with different identity. 
>
>   

Ah, thanks for correcting me. I did a test that I thought proofed my 
right. However, I made the wrong assumptions for the test resulting in 
wrong results resulting in wrong conclusions....

Which, makes me wonder in what scenario's 'intersect' can be useful. It 
appears to me that normal grouping is more flexible.

> = allows you do do value-based intersection (of strings for example)
> if $s1 and $s2 are sequences of strings then 
> $s1[.=$s2] 
> is the sequence of strings that are common to $s1 and $s2.
>   

Yes. And that is basically what many of us showed in response to the OP. 
But don't you find '$1 intersect $2' much more readable than '$1[.=$2]'? 
(if automatic node-to-value were applied for intersect). Undoubtedly 
there are subtle reasons that the designers of XPath 2 have foreseen 
while disallowing union, intersect and except to operate on any type of 
sequences, incl. atomic types (even though there are other operators 
that yield the same result).


-- Abel