W3C home > Mailing lists > Public > xml-dist-app@w3.org > January 2001

RE: Binary attachments to XP

From: Mark A. Jones <jones@research.att.com>
Date: Thu, 25 Jan 2001 06:30:42 -0800
Message-ID: <3A703892.86805222@research.att.com>
To: xml-dist-app@w3.org
> RE: Binary attachments to XP
>
> Date: Thu, 25 Jan 2001 01:46:25 -0500 (EST)
> From: "Larry Masinter" <LMM@acm.org>
> To: "Frank DeRose" <frankd@tibco.com>, "Mark A. Jones" <jones@research.att.com>
> Cc: <xml-dist-app@w3.org>
> Message-ID: <NDBBKEBDLFENBJCGFOIJIENHEEAA.LMM@acm.org>
> Subject: RE: Binary attachments to XP
>
> For some applications, MIME multipart is not adequate, because it makes
> it necessary that both the sender and the receiver be able to buffer
> arbitrary amounts of data.
>
> There's a reason why web clients open multiple HTTP connections.
> And while one can imagine GET transactions being multiplexed over
> multiple connections, there's no way to use multiple connections
> to POST multiple parts of the same request.
>
> Either you need a bi-directional multiplexing protocol (e.g., BXXP),
> or else some kind of packaging that allows interleaving of chunks of
> parts (e.g., that would allow you to send the embedded image data in
> chunks in between two chunks of XML data.
>

Larry,

I have thought a bit about these same issues and I agree that there are cases
that seem to require an interleaving of chunks of data and XML data to avoid
buffering arbitrary amounts of data.  We also have draft scenarios in which we
want to support fire-and-forget types of clients, so interleaving would be preferable
to the BXXP-style solution.  Since we also seem to be chartered to use existing
solutions, do you know of any?

Mark A. Jones
AT&T Labs - Research
Shannon Laboratory
Room A201
180 Park Ave.
Florham Park, NJ  07932-0971

email: jones@research.att.com
phone: (973) 360-8326
  fax: (973) 360-8970
Received on Thursday, 25 January 2001 09:30:25 GMT

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