W3C home > Mailing lists > Public > www-talk@w3.org > July to August 1995

Re: "Hits" pragma (objection and alternative)

From: Reinier Post <reinpost@win.tue.nl>
Date: Thu, 17 Aug 1995 12:24:31 +0200 (MET DST)
Message-Id: <199508171024.MAA05883@wsinis10.win.tue.nl>
To: burchard@cs.princeton.edu
Cc: brian@organic.com, koen@win.tue.nl, www-talk@www10.w3.org, bne@bne.ind.eunet.hu
You (Paul Burchard) write:
>
>Talking to several people privately, I've been convinced that  
>"bundled request" reporting information would most naturally be  
>placed into a Forwarded header, instead of a Pragma.

  Excuse me for jumping in like this, but ...

Instead of using a header, ordinary HTTP transactions can be used to
query proxy servers for state information.  Particular types of request
will yield particular excerpts from the logfiles.  The standardisation
required would be in the URLs used to access this information.  URLs like

  http://proxy.server/getlogs?format=ncsa1.4+type=access+from=01/01/95
  http://proxy.server/getlogs?format=ncsa1.4+type=proxy+from=01/01/95
  http://proxy.server/getlogs?format=ncsa1.4+type=cache+from=01/01/95

might return the specified section of the logs (as far as they are available)
in the specified format (if the format is supported), grepped for the
host making the request.  (This requires special attention, if the requests
themselves are forwarded through proxies.)

Something like

  http://proxy.server/getlogs?report_implementation

might be employed to make the server return a free-format page containing
list of supported formats, the range of data for which logs are available,
an email address for further information, whatever.

Servers, and clients as well, have the right to view the exact imprints they
make in proxy logs.  Why invent a special format allowing proxies to censor
this information?  Just serve the logs!

Advantages:

  - user (server-side) initiated: 'piggybacking' this information onto
    document requests has several obvious disadvantages;
  - not restricted to proxies: this mechanism allows all servers, clients,
    and proxies to query each other for information;
  - open format specification: by using logfile formats,
    any selection of information can be used, provided that a logfile
    format is invented to support it; why invent a new, fixed, and overtly
    specific standard, if logfiles already contain the information?
  - doesn't require a change to HTTP, which is supposed to be a simple and
    standard protocol;
  - not restricted to HTTP servers and clients.  use it on your ftp site :)

A Forwarded: header is necessary, if only to prevent looping, but
I don't think it's the right vehicle for this type of information.

-- 
Reinier Post						 reinpost@win.tue.nl
a.k.a. <A HREF="http://www.win.tue.nl/win/cs/is/reinpost/">me</A>
Received on Thursday, 17 August 1995 06:26:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 October 2010 18:14:18 GMT