Lists Home |
Date Index |
It doesn't seem that you are responding to what Simon wrote.
Don Box wrote:
> The inability for an XML entity to "contain" another XML entity
> (along with a text declaration, doctype decl, etc) has less to do
> with SOAP than it has to do with XML core.
That's exactly the point. That's why XML is a lousy envelope format. It
wasn't designed to be an envelope, and SGML was famously bad at
packaging issues long before XML came along.
> There are several ways to do it (independent from SOAP), but few people
> are thrilled with the results. Relating this to SOAP's prohitibion
> on DTDs is a red herring IMO.
Doctor: it hurts when I try to put use XML as a packaging format. ;)
> As for yet another layer to add headers, the world should be used to
> it by now. HTTP introduced its own header mechanism despite the fact
> that TCP supported header extensions back in the 1980's. By now,
> people should be used to seeing protocols that attempt to avoid
> unneccessary dependencies on adjacent layers.
Simon said that it DOES (or at least might) make sense to define a new
layer for XML-based headers. But he also said that it makes no sense to
require this to be in the same XML document as the payload. This makes
no sense because headers are _by definition_ supposed to be distinct
from the payload in terms of namespaces, schemas and perhaps even
notation and encoding. They are in no sense the same "document".