[Tfug] Progress indicators
bexley401 at yahoo.com
Wed Jan 14 09:45:49 MST 2009
> Poag's Law: All progress indicators lie.
Law of Wall Street: All liars progress.
> Most users are satisfied just knowing that tangible work is
> being done, beyond just a spinning busy pointer or
Yes. This is especially true as some of these indicators
are simply "the spinning dial process is still running"
(e.g., under Slowaris, the indicator has nothing to do
with the process that is doing the real work in some
> hourglass. For example, if they're downloading a file,
> show them how many KB/sec it's coming in at, rounded to
> 2 or 3 decimal places and updated 10-20 times per second, so
> they can see it flit around randomly.... Staring at a
> static, averaged speed reading that never changes is no fun
> for anyone.
It also doesn't convey any useful information. Except for
tech-heads who may want to know what portion of "bandwidth"
is being used, it doesn't tell you how much progress has
been made nor how much remains. And, therefore, *when*
you can expect to be finished.
> One of the things I loved about my Commodore 64 was when
> intros that pirate groups slapped onto the front of releases
> would decompress and execute. Since memory was pretty
> scarce, usually screen memory was used...so the screen would
> randomly fill with garbled characters and flash all sorts of
> crazy ass colors..You at least knew something was happening,
> versus if it were done "cleanly" by not corrupting
> the screen.
The POST memory tests in many (arcade) video games pushes a
64K byte pseudo-random number into memory, verifies it, then
repeats (a few times). On a ~2MHz processor, this takes
a few seconds. As a result, you "see" the random number
being generated (we called the pattern a "carpet") as it
floods the screen.
> Better to overinform the user into a state of amusement
> than underinform them into a state of suspicion.
But, how do you provide *useful* information? Something
the user can benefit from?
More information about the tfug