Response time

[Walter Lewis]

On Thu, 30 Nov 1995, Michael Engle wrote:

> Does anyone know of studies (formal or informal) on how long users of
> networked computer systems are willing to wait for a document page/image to
> be displayed?  I realize that this a a very complex issue, that user
[snip]

The key element, in what I appreciate is a highly complex matrix of 
issues, is how well the user interface communicates to the client the 
relative amount of time they will have to wait.

Dumb terminal screen with no clues (2-5 seconds)
Dumb Gopher client with *some* animationi to indicate "work-in-progress": 
significantly longer
WWW clients with x of y characters (or visual clues like percentage 
bars): significantly longer again --although frustration starts to build 
on table intensive pages where the entire table must arrive before the 
client can build it on-screen--particularly when the client displays part 
of the document as it arrives.

I have long maintained (in considerable company no doubt) that the 
success of Windows and Macs are directly proportional to the clues that 
the system is *doing* something.

Once you're past that hurdle users make a lot of other decisions about 
wait time/network response time/size vs value to them of document, etc.

If they know what they are in for determined users will wait significant 
fractions (and multiples) of *hours* to retreve information online (costs 
being fixed).

Walter Lewis		  		     walter.lewis at sheridanc.on.ca
Deputy Chief Librarian/Systems
Halton Hills Public Library, Georgetown, ON Canada L7G 2A3 (905) 873-2681