> From: vdv@d...
> To: xml-dev@l...
> Date: Wed, 1 Feb 2006 19:25:38 +0100
> Subject: Re: [xml-dev] Minimal XML Specification
> 

> 
> Minimal XML, Common XML, SML and friends are pretty much dead in their
> initial forms but they are still strongly alive in the desperate
> attempts of some of us to keep XML simple!

Yup.  It's worth emphasizing that anyone can safely *produce* a subset of XML and encourage others to do likewise.  Any XML parser is automatically a minimal parser as well, and fairly small XML implementations are pretty ubiquitous parts of most infrastructure.  
 
There is to some extent a well-defined subset of XML in widespread use -- that which excludes DTDs.  SOAP forbids DTDs for performance reasons, and since accepting DTDs from untrusted sources opens up the possibility of a denial of service attack, recent Microsoft APIs move to disable DTDs by default, see http://msdn.microsoft.com/library/default.asp?url=/library/en-us/xmlsdk/html/344c9f1e-b724-44bb-89ff-9a04b2d02638.asp. Nobody claiming to support "XML" can refuse to accept DTDs altogether, of course, and obviously DTDs have perfectly good use cases for hand-authored XML, so this subset is unlikely to displace XML altogether.  
 
Somewhat similarly, people who use the Namespaces spec in a very minimalist manner ( aka "sane" - http://lists.xml.org/archives/xml-dev/200204/msg00170.html ) are likely to suffer a lot less pain than those who go crazy with all its features.   There are lots of attempts to find a sane subset of the Schema spec, but I don't know of a nice compelling summary like this.  OK, plenty of people argue that the null set is the largest sane subset of the Schema spec but that's another permathread....
 
IMHO, the only way XML will be kept simple is if enough people vote with their feet against the cruft. No committee can figure out whether what I think is cruft is more important than what you think is a feature, but maybe the wisdom of crowds can.  
 
 
 
 Express yourself instantly with MSN Messenger! MSN Messenger