Re: Using Content-Encoding and Content-Disposition together

In a previous episode Klaus Weide said...


:: <http://www.w3.org/Protocols/HTTP/1.1/draft-ietf-http-v11-spec-08.txt.gz>.

:: This is currently being served with headers (among others)
:: 
::   Content-Encoding: gzip
::   Content-Type: text/plain
:: 
:: If one wanted to add a Content-Disposition header, should that be
:: 
::   Content-Disposition: xxx; filename="draft-ietf-http-v11-spec-08.txt.gz"
:: 
:: or
:: 
::   Content-Disposition: xxx; filename="draft-ietf-http-v11-spec-08.txt"


I'd argue strongly that the second one is the right way, though the
19.6.1 definition of Content-Disposition doesn't help much. If it
isn't the second one, than there is no way (well, cd is just a hint,
but you get the idea) to save an unencoded version to disk that was
transferred encoded.. a typical reason for saving the resource to disk
is for interoperations with other software on the client machine, and
then if you've got no other local software that knows how to the
decoding, you're SOL.. that's going to be a problem for platforms like
windows and mac where the web UA may be the only thing going that
understands 'deflate' on a typical installation..

even using the second option as the right way doesn't prevent the UA from
saving a compressed version for an entity that was transferred encoded
(a checkbox in the save-as dialog indicating that would be a good
implementation in my mind..)


:: Current usage (typically larger files, which will be saved in
:: compressed form) suggests the first alternative.  If

fwiw - netscape for unix implements it's save-as for the above
resource as plain text. (the second option, though it isn't driven by
a content-disposition: header)

-P

Received on Saturday, 2 August 1997 11:13:36 UTC