Re: content negotiation filename extension

[Simon Spero <se...@tipper.oit.unc.edu>]

I'm in a lot of pain at the moment, so this may end up short.

For MDMA-3,4, in the basic file-system object store, the way I'm handling
multi-typed objects that really exist, as opposed to having been
dynamically generated, is through the directory-to-* converter stack.

When an request is issued that should return a file, and the URL given
names a directory, the variant list is from files in the directory of form
'index.*' or '<directory-name>.*'. Information about base types and 
encodings comes from the extensions and the meta-base; I've been playing 
around with adding a new meta-info handler that will also look at extra 
extensions of the form attr=value.html.gz; I haven't had a chance to play 
with apache much, but I'll take a look and see how easy it would be to 
fit this in if anyones interested.

Simon

Re: content negotiation filename extension

[Brian Behlendorf <br...@wired.com>]

On Mon, 20 Mar 1995, Simon Spero wrote:
> When an request is issued that should return a file, and the URL given
> names a directory, the variant list is from files in the directory of form
> 'index.*' or '<directory-name>.*'. Information about base types and 
> encodings comes from the extensions and the meta-base; I've been playing 
> around with adding a new meta-info handler that will also look at extra 
> extensions of the form attr=value.html.gz; I haven't had a chance to play 
> with apache much, but I'll take a look and see how easy it would be to 
> fit this in if anyones interested.

I think that trying to determine meta-information from file name 
extensions can only go so far - the system as described by roy 
and implemented by rst seems much more powerful.

	Brian.en.html.gz.uu

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@hotwired.com  brian@hyperreal.com  http://www.hotwired.com/Staff/brian/