RE: Interest in standardizing Batch methods?

I see no evidence of a BATCH method or a Request-Id header in my
copy of 2616.  Could you indicate what sections of 2616 you are
referring to?

Thanks,
Geoff

-----Original Message-----
From: Erik Seaberg [mailto:erk@flyingcroc.com]
Sent: Wednesday, January 09, 2002 1:53 PM
To: w3c-dist-auth@w3.org
Subject: Re: Interest in standardizing Batch methods?


Juergen Pill <Juergen.Pill@softwareag.com> writes:

> This BATCH method would open the door into a BATCH language, e.g. execute
> request #2 only if request #1 results in a 200/OK response code.

RFC 2616 already has a (non-XML) encoding for a series of requests or
responses, and each request could have an ID that dependent requests
can check the state of:

	BATCH * HTTP/1.1
	Host: www.example.com
	Content-Type: application/http
	Transfer-Encoding: chunked
	
	141
	MKCOL /foo HTTP/1.1
	Host: www.example.com
	Request-Id: <cid:mkcol23952639587@client.example.com>
	Authorization: Digest response="..."
	
	PUT /foo/bar HTTP/1.1
	Host: www.example.com
	Content-Type: text/plain
	Authorization: Digest response="..."
	If: <cid:mkcol23952639587@client.example.com> (["2xx"])
	
	hello world
	
	0

Microsoft's approach of applying a common request to each of several
resources also handles the basic performance problem and is probably
simpler to implement, though moving it from custom bodies up to HTTP
(for example, use "*" as the Request-URI and put a list of resources
in a "Request-URIs:" header) would make it more reusable.

Received on Friday, 11 January 2002 09:25:58 UTC