Progress and Delay Indication for Varying Performance

11 Mar 2008 - 12:05am
6 years ago
3 replies
372 reads
Chris Ryan
2004

There has been a fair amount written over the years about how to give
good feedback for delays of varying lengths. However, what if your
application has somewhat unpredictable performance (I won't get into
the reasons here; suffice to say it is not sloppy implementation),
where a given operation may complete quickly, or at other times result
in a significant delay? Is there anything that could, or should, be
done to establish this variability with the user without their giving
up due to a perception that things may be even more variable than they
actually are? On the other hand, might simply tailoring the type of
feedback to the length of delay on the fly confuse users as they will
never be sure what to expect?

Any ideas?

Thanks,

Chris

Comments

11 Mar 2008 - 9:56am
Oleh Kovalchuke
2006

Use two progress indicators in a sequence: replace the first with the second
indicator, when delay is long.

Oleh

On Mon, Mar 10, 2008 at 10:05 PM, Chris Ryan <
interactiondesign at redrooffs.com> wrote:

> There has been a fair amount written over the years about how to give
> good feedback for delays of varying lengths. However, what if your
> application has somewhat unpredictable performance (I won't get into
> the reasons here; suffice to say it is not sloppy implementation),
> where a given operation may complete quickly, or at other times result
> in a significant delay? Is there anything that could, or should, be
> done to establish this variability with the user without their giving
> up due to a perception that things may be even more variable than they
> actually are? On the other hand, might simply tailoring the type of
> feedback to the length of delay on the fly confuse users as they will
> never be sure what to expect?
>
> Any ideas?
>
>
> Thanks,
>
> Chris
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> Unsubscribe ................ http://www.ixda.org/unsubscribe
> List Guidelines ............ http://www.ixda.org/guidelines
> List Help .................. http://www.ixda.org/help
>

--
Oleh Kovalchuke
Interaction Design is the Design of Time
http://www.tangospring.com/IxDtopicWhatIsInteractionDesign.htm

11 Mar 2008 - 11:25am
Oleh Kovalchuke
2006

You are right, it is unfailingly annoying, when the system fails your
expectations. That's why the second indicator should have an apologetic
explanation: something like "It seems to take longer than usual...", and to
provide some options to opt out from the process (perhaps to run it in the
background).

The time perception for the second indicator is contextual. I don't have
enough information about the user, the process being performed, the first
progress indicator.
Oleh

On Tue, Mar 11, 2008 at 8:54 AM, Micah Freedman <i at micahfreedman.com> wrote:

> > Use two progress indicators in a sequence: replace the first with the
> second
> > indicator, when delay is long.
>
> Hmm... that sounds a little annoying to me -- watching the first
> indicator nearing it's completion and being surprised and disappointed
> when the second one appears. In fact, I'd much rather have it start
> slow, and then jump to the end than go to the end and start all over
> again. In other words, is there a way to rig it so that it show
> progress as if for the longest possible time?
>
> Some questions: can you know how long it's going to be? Is it two
> distinct times, as you've suggested, or is it that it can take
> somewhere from a few seconds to a few minutes (or whatever) to
> complete, and it's random how long?
>

--
Oleh Kovalchuke
Interaction Design is the Design of Time
http://www.tangospring.com/IxDtopicWhatIsInteractionDesign.htm

11 Mar 2008 - 12:35pm
Chris Ryan
2004

To re-orient the question somewhat, how much variability are users willing to put up with? Does anyone know of any research on this?

Syndicate content Get the feed