RE: Proposed Clarification for Issues 4 and 23

> -----Original Message-----
> From: Noah_Mendelsohn@lotus.com [mailto:Noah_Mendelsohn@lotus.com]
> Sent: Thursday, June 21, 2001 5:35 PM
> To: Rich Salz
> Cc: henrikn@microsoft.com; xml-dist-app@w3.org
> Subject: Re: Proposed Clarification for Issues 4 and 23
> 
> 
> Are we sure we want to get into PI's?  One of the nice things about 
> writing an efficient SOAP processor is you don't need to deal 
> with quite all the minutia of general XML. 

I would agree if we were talking about disallowing a significant amount of
the XML syntax that has proven to be problematic in practice (CDATA marked
sections and external entities come to mind).  In other words, if we said
that SOAP messages are constrained to use the "Common XML Core" [1] or
something like that, there could be significant gains in programming
efficiency.  (I like to tell the story of an experienced programmer who knew
nothing about XML at the beginning of a project, wrote a very fast "Common
XML Core" parser using straight C in a couple of weeks ... and it took him
most of a year to get a full XML 1.0 + namespaces parser working). 

But unless we are going to seriously whack out the "minutia" of XML, we
might as well stick with XML 1.0 + namespaces rather than make an issue out
of PIs. Or is there more XML minutia that SOAP disallows that we haven't
talked about?

[1] http://www.simonstl.com/articles/cxmlspec.txt