Re: distinguishing browser types

   From: Brian Behlendorf <brian@organic.com>

   I like to think we've gotten close with Apache.... certain presumptions 
   were made (an implicit text/html; q=0.5) to account for broken 
   (excuse me, all :) browsers.

Uh, Brian... once again, the current Apache code does nothing special
with text/html, though I did consider it.  (The current accommodation
to current browsers is in the "same-quality" tie-breaking rules only;
this is sufficient to send "text/html; level=3" only to browsers that
ask for it, so long as "text/html" is also available.  Unfortunately,
setting qs on the variants means we're back in the quality zone, and
forced to do it wrong, which is why qs values are currently a
misfeature.  Defaulting quality on */* to 0.1 would fix this,
however).

Aside from the tie-breaking rules, which are specifically outside the
scope of the HTTP spec (as of the current draft), and the business
with CacheNegotiatedDocs, it really is as close to a reference
implementation of the current HTTP draft as I knew how to make it.

Defaulting quality on */* to something other than one is getting
increasingly tempting, however...

rst

Received on Thursday, 27 April 1995 17:30:41 UTC