RE: Summarizing the last 192 discussion

Is this a "Just Say No" approach? There's a difference between
"not promoting this" and "pretending like this isn't rampant".
It's fine to "not promote this", but it's a bad idea to design
the SOAP binding to HTTP such that it doesn't work when such
(widespread) practices are encountered.

SOAP-over-HTTP is promoted as a way of taking advantage of
the existing deployed infrastructure of HTTP. Designing SOAP-over-HTTP
while ignoring the realities of the full range of that infrastructure
means you're developing a binding that is unreliable: you make
it difficult to predict the situations in which it works or
doesn't work. And the support calls will go to the folks who
deploy the end systems, not the protocol designers.

> I think we have to be careful not to encourage "smart" implementations
> and be very crisp on that we see it as a buggy implementation. The Web
> has very bad experience with HTTP applications that try to be smart
with
> respect to content types, URI rewriting, content modifications etc.
The
> problem is that after some time, the smartness turns into a serious
> interoperability problem because people sort of start to rely on it in
> some cases but not in others. I would strongly encourage us to stay
away
> from promoting this as much as possible.

Received on Sunday, 31 March 2002 13:49:18 UTC