RE: Scalar -> node-set conversion (was How to distinguish b/n

Subject: RE: Scalar -> node-set conversion (was How to distinguish b/n
From: Dimitre Novatchev <dnovatchev@xxxxxxxxx>
Date: Tue, 7 Nov 2000 11:01:49 -0800 (PST)
David_Marston wrote:
>Sure, it's nice to be able to say "call my template with any
>arbitrary param, and I'll figure out its type and act accordingly" but
>that may be a luxury.
>I suggest that you send a request to the W3C Working Group on XPath
>that expansion of the count() specs. It would be nice to have.
.................David Marston

The real world application is called the XPath Visualiser. It is
important that whenever the user needs to get some quantitative
characteristics of his xml source (in addition to the basic function of
specifying XPath expressions that return node-sets and viewing them
hilighted in the source document), she will receive the intended result
and not an error message.

Asking the user to enter in different input fields expressions that
evaluate to node-sets and expressions that evaluate to scalars -- that
is both highly inconvenient and un-professional.

At present there's almost no serious product yet that provides an XSLT
Programming Environment. This gap will have to be filled because
there's clearly need for such products. 

Ideally XSLT will be used as their programming language, because it has
been designed to be the most suitable and powerful tool for processing
and transforming XML documents (and any stylesheet is a kind of an XML

In order for a product built upon XSLT to be available on any platform
(portable), it will have not to rely on a particular vendor's
implementation of an API, extension functions, etc.

In my opinion, these facts justify the need for the addition of a small
set of features that are still missing in XSLT, like the ability to
determine the type of a parameter.


Do You Yahoo!?
Thousands of Stores.  Millions of Products.  All in one Place.

 XSL-List info and archive:

Current Thread