Re: Caching Servers Considered Harmful (was: Re: Finger URL)

Daniel W. Connolly (
Mon, 22 Aug 1994 16:01:01 -0500

In message <>, Brian Behlendor
>*Many* times I've gone looking for specs related to HTTP/HTML issues,
>scouring CERN, NCSA, using any and all robots, only to come up short - a lot
>of what's there is poorly maintained, out of date, irrelevant, or pipe dreams
>that never saw their way to implementation.

Yes. One of the things I have strived for in the HTML 2.0 effort is to
make it clear what's really out there and standard, and what's just
a proposal or an idea.

> Saying "it's in the hypermail
>archive" isn't a help either, since short of reading a large chunk of the
>messages I don't have a way of finding the information I need.

Now there's an opportunity where somebody could do a GREAT service to
the community: edit an archive of the www-* mailing lists: provide
fulltext searching, collect stuff by threads, generally organize it
for easy access. The existing archives are good, but the data has
exceeded heir navigation techniques (pick from a great big list) .

> Finally,
>hardly any of these docs contain a last-modifed date, so I have no idea
>how current they are.

Aha! This is not the fault of the CERN folk. It's your fault, or the
fault of the developer of the client you use. My client tells me the
modification date of,
along with other semi-handy info:

Server: CERN/3.0pre6
DATE: Monday, 22-Aug-94 21:54:11 GMT
CONTENT-TYPE: text/html
Last-Modified: Tuesday, 03-May-94 17:37:05 GMT

> This is one thing I'm hoping W3O will make a
>priority of fixing.