- From: Mark Nottingham <mnot@mnot.net>
- Date: Tue, 2 Jun 2009 11:02:07 +1000
- To: Roy T. Fielding <fielding@gbiv.com>
- Cc: Julian Reschke <julian.reschke@gmx.de>, HTTP Working Group <ietf-http-wg@w3.org>, Larry Masinter <masinter@adobe.com>
I haven't followed that discussion (indeed, I can't subscribe to the HTML-WG mailing list, but that's another topic). To me, the biggest difference is that a UA is the end node; i.e., it is the 'origin client' to compliment the 'origin server.' A proxy is a client, not a user-agent. Gateways make this a bit tricky, just as they do with an origin server. E.g., a gateway that implements ESI makes independent requests, but doesn't have a user sitting there. However, it pretends to be a UA. On 02/06/2009, at 10:50 AM, Roy T. Fielding wrote: > On Jun 1, 2009, at 12:59 PM, Julian Reschke wrote: > >> Hi, >> >> based on an off-list suggestion from Larry, I just added <http://trac.tools.ietf.org/wg/httpbis/trac/ticket/166 >> >. >> >> Larry Masinter wrote: >>> Brief summary: >>> The HTTP document uses two terms "user agent" and "client", >>> and in some cases it is careful to distinguish between them, >>> but not always. I would like to see a careful >>> review of each appearance of the terms to insure that >>> "user agent" is used in situations where there is a user, >>> and "client" used in cases where the initiator of the action >>> is either a user agent or another agent with no explicit user. >>> I think this will help the document, and the task relatively >>> limited and editorial. >>> Some instances may require discussion. > > As I mentioned on the HTML list, user agent does not mean there > is a user watching the screen. UA means a client that initiates > requests (as in, decides to start an action based on user controls > or user configuration). > > ....Roy > -- Mark Nottingham http://www.mnot.net/
Received on Tuesday, 2 June 2009 01:02:44 UTC