W3C home > Mailing lists > Public > www-html@w3.org > November 2005

Re: Loading order

From: David Woolley <david@djwhome.demon.co.uk>
Date: Mon, 7 Nov 2005 07:40:09 +0000 (GMT)
Message-Id: <200511070740.jA77e9Q08766@djwhome.demon.co.uk>
To: www-html@w3.org

> Well, I experienced situations where the javascript code inside a
**document loads in frame A could not operate correctly until another

Frames have been effectively deprecated for over half a decade!

**document was loaded in frame B. That is, javascript stuff operating
**cross-documents should be activated ONLY when alla documents are loaded.

I'd put that down as just another undesirable effect of using frames, along
with the primary problem that you cannot link to a frame configuration.
> Currently the <body> tag specifies onLoad for a single document. Is
**there a way to specify onLoad for <frames> too?

I suspect that no consideration was given to this simply because frames were
considered dead when the specification was being written.  iframes are alse
effectively deprecated, but there is a body element that contains them.

I think you will find that early GUI browsers started loading all
resources when they were first encountered.  I believe the current
standard is to not run more than two HTTP connections, for good
neighbourliness reasons, and resources are allocated to a connection as
it becomes free.  A complication is that better HTTP implementations will
use pipelining, so multiple requests can be in flight on a connection.

Also, some resources may be cached in the browser and not need loading or
verifying.  In a well designed site, script resources ought to be cached,
and may well be part compiled.

I'm generally sympathetic with the idea that browsers should be free
to optimise load order.  I admit, though, that I am not sympathetic to
the general pandering to people who don't read specifications which is
resulting in demands for identical presentation, in a medium that was
intended to be extremely platform independent and therefore tolerate
widely different display capabilities and user interface conventions.
Received on Monday, 7 November 2005 07:42:16 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:06:12 UTC