[thelist] how "heavy" do you make each page?

[aardvark]

> From: "Luther, Ron" <Ron.Luther at COMPAQ.com>
[...]
> I always kind of smile when I see these '7' or '27' second rules of
> thumb.  Don't get me wrong - I see the value and 'considerate-ness' in
> having a light footprint.  Also, hanging around here has educated me
> on the pay-per-connect-minute access that is the primary access
> available in some parts of the world today.

actually, that rule mimics a number (8 seconds) as put forward by
Tog and others doing UI testing for applications... if an application
or OS did nothing for 8 seconds after the user expected some form
of feedback, the user would reboot...

that number bears out quite well on the web...

when i've run focus groups or done user testing, i can almost set
my watch to the time the majority of users get impatient...  i
always have a few users in these studies who say that the 2-3
second wait is too long, but at 8 seconds, i can see everyone
starting to lose their focus as an ever-so-mild bit of frustration
creeps into their eyes...

> But I work on an intranet.  I've got users that want to see
> 'everything' and see it 'all at once'.  That takes time.

well, that's a completely different issue, then...

> I've got some report 'pages' that take roughly an hour to come up.
> That's sixty minutes on a client machine with *lots* of memory sitting
> on a big fat corporate pipe pulling information off a tuned Oracle
> database on a internal (and local) corporate server.  [Moral: Don't
> try working at home!]
>
> I dunno - I'm not picking on anyone - but I guess I just wanted to
> point out that when you need to return 500,000 records of data
> weighing in at 30+ Mb on a single page ... '7' seconds isn't really an
> option.

but that's a very specifc case... back in the day, a Wang running a
medical claims payment report might take 8 hours, and couldn't do
anything else in the meantime...

now, i suspect that the only reason 1 hour is acceptable is
because your users are forced to sit there as you stream that down
the pipe...  otherwise, you might have already come up with some
ingenious methods to cut down on that lag...