W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1995

Re: Connection Oriented HTTP conflict

From: Lou Montulli <montulli@mozilla.com>
Date: Mon, 13 Nov 1995 18:04:41 -0800
Message-Id: <30A7F939.41C6@mozilla.com>
To: "Roy T. Fielding" <fielding@avron.ICS.UCI.EDU>
Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Roy T. Fielding wrote:
> 
> > Did your proposal include multipart/mixed responses for
> > keep-alive cgi scripts?  That seems to be what everyone
> > is implementing.
> 
> No, and I'm not aware of *any* implementation that does this.
> Is somebody holding out on me?
> 
> Right now, CGI scripts just ignore keep-alive and close the connection.
> For the "official" HTTP/1.1, they can use the chunked transfer encoding
> instead of a multipart, if desired.

Netscape navigator and the netscape server both use 
"multipart/mixed" as per Alex's proposal.  I'm not
aware of any server implementer's even wanting to
support chunked transfer encoding.  I for one
will strongly fight against adding yet another 
encoding form.

> 
> > I'm not happy with using the "multipart/mixed"
> > name.  I would prefer "multipart/x-http-response" or something
> > like it so that we don't have name space collision with email
> > messages.
> 
> That isn't a namespace collision -- multiparts are just multiparts:
> a mechanism for sending multiple bodies in a single message.  The UA
> should treat them identically no matter where they come from.
> 

There is a namespace collision.  the user representation of
a multipart/mixed message is well specified and does not
conform to the usage in Alex's proposal.  "Multipart"
encoding is the right thing for HTTP but it should use
a different sub name.

:lou
-- 
Lou Montulli                 http://www.netscape.com/people/montulli/
       Netscape Communications Corp.
Received on Monday, 13 November 1995 18:09:20 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:35 EDT