RE: XML and stateless protocols

["Hollenbeck, Scott" <shollenbeck@xxxxxxxxxxxx>]

> While contributing[1] to a TAG discussion, I realized that the point I
> was making was likely relevant to this work.
> 
> When using XML in the context of a stateless protocol, be it the
> protocol itself (e.g. SOAP), or simply as content transferred by an
> existing protocol (e.g. XML/HTTP), care must be taken not to make the
> meaning of a message depend on information outside the message itself.
> 
> XML 1.0 provides external entities, which are an easy way to make the
> meaning of message depend on something external.  Using 
> schema languages
> that change the Infoset, like XML Schema, are another way.
> 
> I think it would be a good idea to discuss this in the draft.

How is this unique to XML, and thus of specific concern for the draft?  One
could take the general statement of "care must be taken not to make the
meaning of a message depend on information outside the message itself" and
apply it to _any_ protocol design.

-Scott-