Re: IFrame Resets

[Strongly off topic: how to, browser behaviour, *browser* object model;
invalid HTML; accessibilty violations.]

> You could try adding the date and time as an argument when calling an update

This a workaround for misconfigured web servers (incorrect cache control
information), or limitations in the browser object model preventing a
proper refresh++.  It's likely to cause pollution of the browser cache,
although public caches no longer cache this sort of URL== because of
abuse.  (POST was always uncacheable, but GET and HEAD were supposed to
be cacheable for all resources; frames and iframes always generate GETs.)

Sites SHOULD NOT depend on scripting for their basic function, nor on
iframe - the purpose of the iframe element content is to provide alternative
access to the information, not a place for a "you are using the wrong
browser" message.

Off topic features:

How to question.
About the behaviour of a specific, if unnamed browser.
About the behaviour of invalid HTML (form actions are #REQUIRED, maybe others).
About the browser object model (even W3C document object models are off topic,
but the W3C doesn't, as far as I know, have a browser object model standard).

(It is possible that it could be reformulated as a document object model
problem, although I don't know if the href property is defined for iframe
in the W3C models, and if it is, whether any behaviour is specified.
The frames collection is not part of the W3C model.  The specifications
are the right place to answer this, although they need to be cross checked
against those for real world browsers.  I suspect any DOM implementation
would also optimise out the null change in the resource.)

++ Netscape's (proprietory) browser object has had refresh (conditional and
unconditional) capabilites as methods of the location object since NS
3.0) - I can't guarantee that they work properly for iframe Windows,
or in the IE clone of the object model.

== Excessive use of uncacheable contents means that some ISPs with connectivity
problems use more aggressive caching than the HTTP standards allow.

Received on Thursday, 9 January 2003 01:48:00 UTC