W3C home > Mailing lists > Public > www-ws-arch@w3.org > November 2002

RE: Roy's ApacheCon presentation

From: Anne Thomas Manes <anne@manes.net>
Date: Wed, 20 Nov 2002 15:46:04 -0500
To: "Mark Baker" <distobj@acm.org>, "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>
Cc: <www-ws-arch@w3.org>
Message-ID: <ECEDLFLFGIEENIPIEJJPGEMDCNAA.anne@manes.net>

Mark Baker said:
>
> You snipped out the HTTP/WSDL bit.  Do you agree that HTTP defines an
> interface in the same way that a WSDL document does?

No, I don't agree. The HTTP interface represents almost no semantic meaning.
A WSDL interface can represent enormous semantic meaning. See Ugo's last
message regarding the law of conservation regarding application semantics.

HTTP has no knowledge of the contents of messages. SOAP/WSDL has extensive
knowledge of the contents of messages. Therefore with SOAP/WSDL I can
generate code to process my messages. I don't have to write all the code by
hand.

In many circumstances, I prefer to put some semantics into my interface
because it makes the application development process that much easier.

That's not to say that I don't like the REST approach, but I haven't seen an
argument that works for me as to why it's so much better than the "abuse of
POST" approach. I understand that REST takes better advantage of the power
of the Web, and you can do things like create book marks and pass links, and
cache results, etc. And these might be very desirable features in some
circumstances. But from my perspective, most "real" Web services (filing my
taxes online, submitting an insurance claim, ordering a thousand widgets,
managing logistics, posting an email correspondence to my CRM system at
Salesforce.com, etc.) will require SOAP headers for security, management,
reliability, message coordination, etc. The REST approach can't address
these requirements.

Anne
Received on Wednesday, 20 November 2002 15:43:53 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:10 GMT