Re: LINK TYPE=override/type

Rasmus Kaj (kaj@cityonline.se)
Fri, 23 Jan 1998 14:48:20 +0100


To: liam@htmlhelp.com
Cc: www-html@w3.org
Cc: kaj@cityonline.se
From: Rasmus Kaj <kaj@cityonline.se>
In-Reply-To: Your message of "Fri, 23 Jan 1998 08:36:25 -0500"
 <3.0.5.32.19980123083625.0096f700@undergrad.math.uwaterloo.ca>
Message-Id: <19980123144820F.kaj@cityonline.se>
Date: Fri, 23 Jan 1998 14:48:20 +0100
Subject: Re: LINK TYPE=override/type

>>>>> "LQ" == Liam Quinn <liam@htmlhelp.com> writes:

[ About sometimes ignoring the http content-type ]

 LQ> In general, yes, but there are cases in which guessing can be
 LQ> more appropriate than simply ignoring the badly served file.  For
 LQ> example, if foo.css is linked with <LINK REL=StyleSheet
 LQ> HREF="foo.css" TYPE="text/css"> and the server returns a
 LQ> Content-Type that is definitely not a style sheet (e.g.,
 LQ> text/plain or application/octet-stream), then I think that using
 LQ> the TYPE attribute would be more appropriate than ignoring the
 LQ> file.  However, if <A HREF="foo.css" TYPE="text/css"> is used

How about saying that a specific type (i.e. application/octet-stream)
means 'unknown', so when a useragent comes by something that is
application/octet-stream it tries to guess (by LINK TYPE or looking at
the file or anything) the correct content-type.

This seems to me to be the way that application/octet-stream is used
today anyway (i.e. to say 'I couldn't find the right type for this
content').

// Rasmus

-- 
kaj@cityonline.se --------------- Rasmus Kaj - http://www.e.kth.se/~kaj/
 \               CityOnLine IB Production AB - http://www.cityonline.se/ 
  \------------------------- Unite for Java! - http://www.javalobby.org/