Re: Caching servers

Leigh Blue Caldwell (901732ca@udcf.gla.ac.uk)
Sun, 12 Feb 1995 09:48:15 -0800

Bob Novick wrote:
> On Fri, 10 Feb 1995, David H Dennis wrote:
>
> > My view is that a site should have a right to know how many times its
> > pages have been accessed, even if they're loaded from cache.
>
> This is the crux of the anti-caching argument.
> Certainly it would be nice to know how many people, and even who they
> are, that visit your site. But it is not a _right_, only something that
> technology may make possible due to the nature of electronic communications.

Here's an idea. Next time you're speaking to your clients, mention:
"Not only have the same number of Internet clients been reading
your page, but it has been downloaded to the Prodigy Caching
Server! That means another 300,000 people have direct access to
it!"
To gullible clients, this may sound like 'Cashing Server' and they
will be convinced they are about to get 300,000 online orders for
their products :)

Seriously, though, the main effect is another level of education of
the market - "See this line, with the prodigy.com address - that
was a cached download, and an estimated 4,000 people have read it
from this particular data transfer" or whatever. It's a shame to
lose some information about precisely who is doing the downloading,
but I imagine someone will build the capability soon enough into
their caching server to interpret a request of:
CACHEDATA 95-02-01 CommonLogFormat/1.0
and return the cache requests it has had to documents with URLs from
the server sending the request. This wouldn't work for all servers,
for example those whose real name is something like 'newton' and has
just been aliased as 'www'. But there are obviously ways to get
round issues like that, for example checking domains only.

Leigh.