Re: XSLT wish list

Subject: Re: XSLT wish list
From: Duane Nickull <webmaster@xxxxxxxxxxxxxxxxx>
Date: Fri, 21 May 1999 18:52:22 -0700
Robert C. Lyons wrote:
> 
> Duane wrote:
>        >   The error message should include the line number within the source
>        >   document of the invalid data.
> 
>        IE 5.0 does this right now.  In fact, it pukes on anything which does
>        not conform to the proper XML syntax (as it should).  A good parser
>        should do this.
> 
> I was not referring to XML syntax errors (e.g., the source doc is not well-formed)
> or XSL syntax errors (e.g., the stylesheet contains an invalid instruction).
> Instead, I was referring to application-level errors (e.g., a person's birth date
> is greater than the current date).
> 
I don't beleive that there should be any application-level validation or
it will cause XML/XSL to lose some of it's extensibility.  The above
example of a date validator may be nice but it will instrinsically
increase the complexity of writing a good XML parser.  The tools are
there for people to write their own methods, event handlers and error
procedures.  I believe these should be used.  As far as IE 5.0 goes,  I
have been using it since the first beta and it has been both a blessing
and a curse as we found it to be very finicky.  It certainly forces me
to keep up the quality of my work ;-)

I do agree with you though that it may be nice to have a direct gateway
to SSI's through an <xsl:cgi> tag or something similar.

Another reason I am not in favor of this is the issue of
forwards/backwards compatibility.  Imagine a person gets an XML/XSL
parser written today for the current spec.  By default, such a browser
should ignore tags it does not understand and not take any action.  This
ensures that the parser (browser) will be able to operate tomorrow
without generating 100's of application-level error messages whenever it
encounters a tag/processing instruction/date etc that it does not like. 
Again,  the current spec of XSLT/XML gives authors the tools to create
their own level of additional validation above what is required by w3. 
Each individual author can set their own parameters for what to do when
errant data is encountered. 

Duane Nickull


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


Current Thread