Re: HEAD and 301/302/303 conflict?

At 1:18 PM 9/13/95, Daniel DuBois wrote:
>To reaffirm: HEAD requests NEVER produce an entity body; anywhere the spec
>implies that there might/should/could be a body, the HEAD rule overrules it?

Um, make that HEAD requests *should* never produce ... In my recent survey
of 2000 different Web sites
(http://www.proper.com/www/servers-survey.html), I found that about 1.5%
*did* produce a body. Some of these were HTML that said that HEAD was not
supported, some of them were error messages reporting page not found (note
that the requested page for all the searches was "/"), and some were
clearly the same thing a GET would have returned.

I guess that CERN 3.0 allows you to disable HEAD in its configuration file,
since some of the "Method HEAD is disabled on this server" messages came
from sites running CERN 3.0. Maybe Henrik can shed some light on this...

--Paul Hoffman
--Proper Publishing

Received on Wednesday, 13 September 1995 15:17:06 UTC