httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <field...@kiwi.ICS.UCI.EDU>
Subject Re: META revisited
Date Tue, 25 Feb 1997 00:36:56 GMT
>A long, long time ago, I brought up the question of HTTP-EQUIV META tags. I
>was told at the time that this is a client-side thing. However, this does not
>appear to be the case. Firstly, the HTML 3.2 spec clearly says that this tag
>should be used by servers, and secondly, Netscape at least do not process it at
>the client end (as far as I can tell).

Well, I created META for MOMspider, and I wrote that part of the HTML
specification, so I guess I can answer this one. ;-)

HTTP-EQUIV META tags are a mechanism for the author of an HTML document
to provide HTTP metainformation to the server and/or client.  The intention
is that the information be provided in response to a HEAD request, which
is why it isn't a client-side thing.  The reason is so that a spider can
collect that metainformation without retrieving the entire document.
There are many other ways to do the same thing -- META was just the one
way to do it in HTML.

The intention was *not* to have the server parse the file upon every
request.  The information is simply provided -- how the server obtains
it for later use was none of our business.  The sensible thing is to
maintain a table of metainformation for each resource, or simply an
internal cache of entire responses; the former is how WN does it and
the latter is how a fast HTTP server should be designed.

It would be sensible for Apache to implement it for SSI-parsed files,
since we are already doing byte-parsing for those, but nothing else
until we have a built-in caching mechanism.  However, I wouldn't bother
until 1.2 is out the door, since we provide other mechanisms for
accomplishing the same task.

.....Roy

Mime
View raw message