[10417] in www-talk@info.cern.ch
Re: HTTP header suggestion/request
daemon@ATHENA.MIT.EDU (Rens Troost)
Sun Nov 10 12:26:47 1996
To: MegaZone <megazone@livingston.com>
Cc: masinter@parc.xerox.com (Larry Masinter), kdyer@draper.com,
www-talk@w3.org, koen@win.tue.nl, rens@century.com,
sdorner@qualcomm.com
In-Reply-To: Your message of "Sat, 09 Nov 1996 20:33:22 PST."
<199611100433.UAA25381@server.livingston.com>
Date: Sun, 10 Nov 1996 12:20:13 -0500
From: Rens Troost <rens@name.net>
Resent-From: www-talk@w3.org
>
> You can develop a complex interrelation - when does Content-Type take
> precedence, when does Content-Disposition?
Appended is what I do. In brief; trust C-D for storage names,trust
C-T for viewer selection. Im not sure this kind of thing is
appropriate to promulgate in an RFC, though. ANy thoughts?
to generate a filename:
if C-D provided
use suggested name
else
if C-T has associated suffix
generate name with appropriate suffix
else
dont suggest a name
endif
endif
to launch a viewer:
if c-t has registered viewer
launch appropriate viewer
else
prompt user
endif
-Rens
