home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Collection of Internet
/
Collection of Internet.iso
/
infosrvr
/
dev
/
www_talk.930
/
001228_weber@eitech.com _Tue Jun 1 21:37:30 1993.msg
< prev
next >
Wrap
Internet Message Format
|
1994-01-24
|
2KB
Return-Path: <weber@eitech.com>
Received: from dxmint.cern.ch by nxoc01.cern.ch (NeXT-1.0 (From Sendmail 5.52)/NeXT-2.0)
id AA04634; Tue, 1 Jun 93 21:37:30 MET DST
Received: from lks.lks.csi.com by dxmint.cern.ch (5.65/DEC-Ultrix/4.3)
id AA18307; Tue, 1 Jun 1993 21:59:02 +0200
Received: from eitech.eitech.com by lks.lks.csi.com (5.65/6.930123)
with SMTP id AA25050; Tue, 1 Jun 93 14:57:47 -0500
Received: from bd.eitech.com ([192.100.58.27]) by eitech.com (4.1/SMI-4.1)
id AA05939; Tue, 1 Jun 93 12:53:47 PDT
Message-Id: <9306011953.AA05939@eitech.com>
X-Sender: weber@eitech
Date: Tue, 01 Jun 1993 12:59:32 -0800
To: www-talk@nxoc01.cern.ch
From: weber@eitech.com (Jay Weber)
Subject: HTTP ACCEPT field and MIME types
X-Mailer: <PC Eudora Version 1.1a5>
I was looking through a draft HTTP 2.0 spec (may be as old as 7-Dec-92)
and noticed a potential problem with the ACCEPT request field. One puts
a comma-separated list of MIME types as the field value. The problem is
that MIME types can involve parameters, and some parameters may be
crucial to a client's acceptance of the type. For example, an old
example from the MIME spec is:
Content-Type: application/octet-stream;
name=foo.tar.Z; type=tar;
conversions="x-encrypt,x-compress"
This example is extreme, since application/octet-stream is usually a
no-brainer for clients, but the presence of these parameters makes it
pretty UNIX specific. You can image other examples, such as a "lossiness"
parameter on image/jpeg, that would impact portability of the type.
Anyway, if we include parameters in MIME typing, then a comma-separated
syntax doesn't work too well. An easy fix would be to allow multiple
ACCEPT fields in the header, with one MIME type per field. I don't see
an HTTP stance on multiple fields, and this treatment of multiple
fields would be consistent with RFC 822 (the To: field, for example).
How about it (Tim)?
In general, HTTP 2.0 is pretty cool. Is an RFC coming soon (or here)?
Jay Weber
EIT