> Second, and more worrying, I get different results with XT 
> and Saxon, as shown. Can anyone tell me why?

Usually when this is reported I panic, because XT nearly always turns out to
be right. This time though, unless James comes up with some magic words
buried deep in the spec that I've missed, I think Saxon is giving the right
answer (except for the omission of the xml declaration: when the keyword
changed from xml-declaration to omit-xml-declaration, I accidentally left
the default as "yes").

Mike Kay
 
> 
> Francis.
> 
> 
> G:\xmlSchema>type t_.xml
> <?xml version="1.0"?>
> <data>
>         <a name="jaz" depends-on="bar"/>
>         <a name="spaz"/>
>         <a name="maz" depends-on="spaz"/>
>         <a name="foo"/>
>         <a name="bar" depends-on="foo"/>
>         <a name="baz" depends-on="foo"/>
> </data>
> 
> 
> G:\xmlSchema>saxon t_.xml t_.xsl
> 
>                 a=spaz
>                 a=foo
>                 a=maz
>                 a=bar
>                 a=baz
>                 a=jaz
> 
> G:\xmlSchema>xt t_.xml t_.xsl
> <?xml version="1.0" encoding="utf-8"?>
> 
>                 a=spaz
>                 a=foo
>                 a=maz
>                 a=bar
>                 a=jaz
> 


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list