xml-dev - RE: Transformation

[Didier PH Martin <martind@netfolder.com>]

Hello,

I still have this pet project I worked on last year. I named it XMScript.
The whole idea was to get procedures to be matched against a pattern and
fired when the pattern is found in the source document. XMScript supports
PerlScript, PythonScip, VBScript and JavaScript procedural constructs. As an
example:

rule("element1/element1.1/@anAttribute")
{
	javascript code here
}

The rule body syntax follows the language syntax rule. For instance, in
python the rule's body is indicated by indentation, in VBScript enclosed by
the rule---endRule constructs, etc..

XMScript engine is like a SAX based engine except that an XPath pattern is
constructed as the file stream is parsed. Each time a new XML object is
encountered, the XMScript converts this into an XPath construct and then
tries to fire a procedure matching this XPath construct.

It seems that this kind of tool may be well adapted to Steve's problem.
However, it shares the same limitations as Omnimark. Recently, someone made
the suggestion to keep pointers to the document's parts so that different
document's portion could be used for a transformation (some equivalent to
the <xsl:apply-template../> construct). I am now studying if this suggestion
leads to a more powerful transformation engine.

Anyway, this was just to share some experiments and ask if anyone has any
good idea on how to create transformation engines based not on an info set
but on the text stream.

Cheers
Didier PH Martin
----------------------------------------------
Email: martind@netfolder.com
Conferences: xml devcon 2000 (http://www.xmldevcon2000.com)
		 ICAP summit (http://www.marcusevans.com)
		 Wireless Summit NY (http:www.pulver.com)
	       xml devcon 2001 London (http://www.xmldevcon2000.com)
Book: XML Professional (http://www.wrox.com)
column: Style Matters (http://www.xml.com)