W3C home > Mailing lists > Public > www-ws-desc@w3.org > May 2002

Re: Issues 6a, 6d, 41. Define encoding into a request URL

From: Paul Prescod <paul@prescod.net>
Date: Tue, 07 May 2002 16:33:52 -0700
Message-ID: <3CD86460.6FA2737C@prescod.net>
To: www-ws-desc@w3.org
I don't think that the <http:urlXML use="literal"/> idea is bad but I
also don't think that it is strictly necessary. It's attractive because
it would answer a common HTTP FAQ: "how do I handle recursive data", so
I won't fight against it.

But the better answer to that question is that URIs are URIs are in
almost every circumstance flat (segmented but flat) and as easy to read
as possible. If you need to send deeply nested XML in a URI then you
probably are not using HTTP as it was designed. Therefore I am not
convinced that the inability to encode complex types into a request URL
is a significant problem worthy of much effort. If it is important to
you that every binding be able to express every message, then okay.
Otherwise, I wouldn't bother. As long as people who want to use WSDL
with HTTP can do so, I don't mind it if the WSDL spec encourages them to
use HTTP best practices (flat, human readable URIs).

 Paul Prescod
Received on Tuesday, 7 May 2002 19:33:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:20 GMT