Re: Automatic Refresh

> - if the cookie is set, do a reload via javascript (not META) after a
> set timeout period

Firstly you meant (the non-standard) HTTP Refresh header, not META; 
Refresh works as a real HTTP header including on non-HTML resources,
at least for major browsers.  Conceptually it only works in META
because it works as a real header.

Secondly, there is no reason that I can see for involving scripting
here (except possibly as a counter counter measure for users who
turn off the Refresh mechanism - but such counter counter measures are
likely to be rapidly adopted by the abusers if the counter measure is
adopted by users).

There is not strong case for cookies either, except to allow an anonymous
user to use the main URL on a later visit but have their preference
remembered.  Within a session, two alternative URLs would be as good.

Generally, though, HTML is not well suited for this sort of periodic 
update, even with some of the features in IE that allow the update to
be done without he screen jumping.  There might be a legitimate case
for scripting here to create a client side application that fetches
just the data, although non-script alternatives must be provided.
Note that NS4 would require misuse of form fields to avoid a complete
regeneration of the display.

Also, pages that auto-update like this should never contain any form
controls, with the possible exception of submit buttons, as one of the
most irritating features about this sort of technique is that it can 
catch you out when you are halfway through changing parameters (the typical
case is that you are changing filtering paramters).  Even links and submit
buttons are vulnerable, but not as vulnerable as multi-step processes.

Received on Sunday, 4 July 2004 03:42:18 UTC