You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Rob Hartill <ro...@imdb.com> on 1996/05/14 16:43:52 UTC

Problems with Content Negociation (fwd)

not acked.

is this answered by mod_negotiation  ?



Date: Tue, 14 May 1996 12:32:59 +0100 (BST)
From: Nick Smith <na...@ant.co.uk>
Subject: Problems with Content Negociation
To: apache-bugs@apache.org
Cc: Nicko <ni...@ant.co.uk>, chris.lilley@sophia.inria.fr
Message-ID: <Ma...@linda.ant.co.uk>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; CHARSET=US-ASCII
X-Organization: ANT Ltd, Cambridge UK. http://www.ant.co.uk/
X-Mailer: ANT RISCOS Marcel [ver 0.99o]

This is a problem with the Apache server (at least 1.0.3, and I haven't 
seen any mention of this problem in the latest beta release notes), and
applies to any hardware platform.

All WWW clients (Lynx, NetScape, Mosaic, Fresco, etc) send out */* somewhere
in their HTTP_ACCEPT. eg, NetScape 2.0 sends;

  */*, image/gif, image/x-bitmap, image/jpeg

the Apache interprets the */* with the same priority as the other, explicit
image types, and so the highest priority in the server .var file will be
sent (eg, image/png) even though the client doesn't understand such files.

It can be argued that this is a fault of the client - it should send
something like;

  text/html, image/gif, image/jpeg, */*; q=.2
  
which is what HotJava sends, but this doesn't address the problems of all
the millions of existing browsers out there.

I propose that Apache be updated to default */* to a low q factor, if no
other q factor is explicitly provided by the client.  Explit MIME types such
as image/jpeg presumably default to q of .5, so they would be supplied in
preference to something like image/png.

Apparently the CERN server does this - Chris Lilley of W3C said recently 
in the png image mailing list:

> So, I would say that when assigning q factors to client requests that
> didn't have them, use 0.5 for fully qualified media types
> (type/subtype) but have perhaps 0.2 for partial ones (type/*) and
> very low values for the unqualified case (*/*).

Cheers,

Nick.
-- 
Nick Smith, ANT Ltd, Cambridge UK // Home page: http://www.ant.co.uk/~nick/

----- End of forwarded message from Nick Smith -----

-- 
Rob Hartill (robh@imdb.com)
The Internet Movie Database (IMDb)  http://www.imdb.com/
           ...more movie info than you can poke a stick at.