Re: A broken browser

Martin J. Duerst (mduerst@ifi.unizh.ch)
Mon, 20 Jan 1997 14:45:58 +0100 (MET)


On Mon, 20 Jan 1997, Koen Holtman wrote:


[regarding the asymmetry between 

> >    Accept-L   Content-L
> >     en         en-us
(matched according to HTTP 1.1)

and
> >     en-us      en
(not matched according to HTTP 1.1)


> When I got involved in HTTP, the above asymmetry already existed.  I have
> justified why we did not extend the matching rules to make it more
> symmetrical, but I do not know the reason why the first matching rule was
> there already.  Maybe because all other accept/content pairs also have this
> kind of matching.

That may explain the present state. But while e.g. for Content-Type,
it makes sense to say that you accept text, or an image, but all actual
documents are text/html, text/plain, image/gif, or whatever, this
is different for Language. It is perfectly reasonable to tag a document
as "en", and it is perfectly reasonable for a user to set preferences
for "en-us", and it would make perfect sense for a server to match
the two if it matches them the other way round.

Regards,	Martin.