Re: "Hits" pragma

Roy T. Fielding writes:
> "Balint Nagy Endre" <bne@bne.ind.eunet.hu> mentions:
> 
> >I presume we shall incorporate the hits pragma in the ongoing version of the
> >HTTP draft as a must requirement.
> 
> The protocol does not require any Pragma directives.  If it is a requirement,
> it is made based on the system semantics (e.g., WWW) rather than at
> the protocol level.
> 
> I am generally against making anything a requirement if it hasn't
> already been implemented and demonstrated to be a defacto requirement.
Agree, generelly, but in particular, I mean simply that making the hits pragma
optional may lead to misuse of no-cache pragma. It's not too hard to stuff in
the pragma by anybody who owns the document. For example using CERN httpd/3.0
one must create the file $MetaDir/<document-name>$MetaSuffix containing
"Pragma: no-cache".
(Yeah, it works this way, and the pragma also works, using CERN httpd/3.0,
I checked right now. I presume using other implementations the situation is
quite the same.)
I found too many redirections now, which (I guess) are used to fool caches.
(At least CERN httpd/3.0 doesn't cache redirections, which is a good practice
according to rule '200' responses are cacheable only).
I'm sure, if we miss the appropriate point to enforce feedback of hit counts,
then we will lose in transferring the same over the net again and again,
just to have those hits counted.
And of course, I mean HTTP/1.1 under 'ongoing version', not HTTP/1.0.
Generally, if we help information owners gather the same statistics from caches,
what they may colllect about direct request, then it will be easier to treate
cache content as part of the network infrastructure versus local copies of
copyrighted material. And this is a really important question. If the caches
will have bad publicity in this question, we all will lose, except WAN link
providers. But the original goal of founding the internet wasn't to make
WAN link providers rich.

I don't want pression you, rather I want hear other opinions too.

Andrew. (Endre Balint Nagy) <bne@bne.ind.eunet.hu>

Received on Monday, 14 August 1995 16:25:39 UTC