RE: XSL as a better XPointer was RE: The Cathedral and the Bizarre (was: do you use pi's?)

[Dieter Maurer <dieter@xxxxxxxxxxxx>]

Jonathan Borden writes:
 > Chris Lilley wrote
 > > Ok, so now make it do the span (verse 12 *and* verse 13) 
 > 
 > 	"//chapter[21]/v[12 $to$ 13]"
 > 
 > >and for an
 > > encore, make it select "better Xpointer (less weight" in 
 > > 
 > > <foo>XSL patterns are a better XPointer <reason>(less weight more
 > > filling...)</reason></foo>
 > > 
 > 
 > 	foo/text()/region("better Xpointer (less weight")
 > 
 > 	or
 > 
 > 	foo/text()/region(21,50)
Obviously, Chris and Jonathan speak about *TWO* different XSL versions.

While Chris seems to have the current W3C XSL working draft of Dec 16, 1998
in mind, Jonathan speaks about a different XSL, maybe the XSL
currently implemented in IE5.

I, personally, hope that the final standardized version of XSL
will support many XPointer features, especially ranges (spans).
I hope, however, it will not use the syntax shown above.

I dislike e.g. the '$' around the 'to', because it is different
from the 'and' and 'or' operators. 'to' should become
an operator (named 'to' not '$to$').

I am not sure, whether I like the indexing in the form '[no]'.
This is very intuitive for programmers, but in XSL '[...]'
means filtering. Although more cumbersome, it probably
would prefer a function '[index(no)]', maybe in two
variants ('index-of-type', 'index-of-any').
I would then also prefer a function 'between' over an operator 'to'.


- Dieter


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