Re: [xsl] Can sets have order?

Subject: Re: [xsl] Can sets have order?
From: Mike Moran <Mike.Moran@xxxxxxxxxxx>
Date: Thu, 25 Jan 2001 18:23:13 +0000
Michael Kay wrote:
> 
> > Isn't it a bit confusing to say that it returns it in reverse document
> > order? My understanding is that ancestor:: is ordered most immediate
> > first i.e direct parent, then parent of parent and so on, all
> > the way up to the root.
> 
> An axis identifies an (ordered) list of nodes. The predicate associated with
> the axis is applied to the ordered list. The result of an XPath expression,
> however, is always a node-set, not a list. The node-set is unordered, but
> the nodes have an ordering, called document order. Many operations on
> node-sets process the nodes in document order. Is that clear now?

Well, no. Two sections confuse me:

1: "An axis identifies an (ordered) list of nodes. The predicate
associated with the axis is applied to the ordered list."

I would have thought that the following would make more sense, because
it seems wierd to throw away an ordering that is never seen:

"An axis identifies an (ordered) list of nodes. The predicate associated
with the axis is applied to the *unordered set*."

Perhaps it is the case that the ordered list mentioned is the path from
the root to the current node which is re-ordered to run from child to
ancestors, rather than from parent to children?

2:

"The node-set is unordered, but
 the nodes have an ordering, called document order."

It is confusing to say that node-sets are unordered, whilst nodes are.
Is this ordering intrinsic to a node? It makes more sense to think of
ordering as associated with collections and not instances (i.e. members
of the collection). This seems more sensible, regardless of  how it may
be implemented (eg a "next" pointer or suchlike).

-- 
Mike.Moran@xxxxxxxxxxx 
                   Web: http://houseofmoran.com/
               AvantGo: http://houseofmoran.com/Lite/

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


Current Thread