W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 2004

Re: General Usability DetecVS tor

From: Maurizio Boscarol <maurizio@usabile.it>
Date: Fri, 14 May 2004 12:19:53 +0200
Message-ID: <40A49D49.1030802@usabile.it>
To: Gregg Vanderheiden <gv@trace.wisc.edu>
Cc: w3c-wai-gl@w3.org

Gregg Vanderheiden wrote:

>> I would be interested in your examples.
>>
>>   
>
> Ok, I'll take the time to make a simple table to express myself in a
>
> screen-readable mode..
>

Here I am...

... This is what I meant talking with Gregg about the matrix including 
different tasks for different groups of users.

http://www.usabile.it/examples/usermatrix.htm

I don't think it's nothing new, but perhaps it can be useful to remember 
that in usability we can have general evaluation in which more 
task-specific and user-specific consideration should be made. Even 
comparing the usability problems in the same task for different user in 
two version of design can be useful. The user configuration can be 
split: the same type of user can have a 640 x 480 monitor, as well as 
1024 x 768... for certain tasks and sites this can make a lot of 
difference (I have a webmail where the logout is *impossible* to see and 
to click in 640 monitor, even with scrolling...).

The problem remain in numerals. Numbers can be used to extimate 
difference in performance between groups? Yes, but should be done with 
experimental constraints. Not so easy to do, very expensive and 
time-consuming.

Extimating time on the basis of 3-8 user per group can lead to big 
mistakes in evaluation. So I propose analytical method, also with 
numbers, but not based on user measured performance (like the task 
analysis).

As an example, I compared two version of  alistapart page design, using 
some measures to extimate how ease to read an article was. (only in 
italian, sorry: http://www.usabile.it/242003.htm; on the same topic, but 
on a different problem: 
http://www.usabile.it/archivio/dicembre2003.htm#contrasto ). A 
before-after comparison can always be done in similar way, even if what 
is to measure can be difficult sometimes to agree.

Don't know if this can help, anyway. Let me know, if you want.

Best,

Maurizio Boscarol
Received on Friday, 14 May 2004 06:21:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:30 GMT