Re: [Usability] how to make usability test?

[Jason Hoover <jasonhoover verizon net>]

On Mon, 2005-06-27 at 20:55 -0400, Eric Larson wrote:

> I don't think we need to create our own rules and guidelines as there
> are already such guidelines available. But making these publically
> avaiilable and in some cases more gnome specific could be very helpful.

Maybe having a few links to them from the usabilty page in external
resources would be helpful. That whole page needs cleaning up, a few
links are just missing amongst other things. 

(On an off-note, has anyone ever taken a look at the KDE style guide?
Makes me really fond of the HIG. Cancel and Close meaning the same
thing, eugch.)

> This is a great idea! But, I think it doesn't address things like
> ethnographic studies and many observational studies. The problem in both
> scenarios is the amount data collected and how to process it in a way
> that allows retrieval of information relevent to the questions being
> asked. I know a lot of people won't be running large scale research
> projects on Gnome usability but it would be useful to have a way for
> hackers to test their software on a person and record the results in a
> useful way. This kind of testing is very common in things like web
> development where it is simple to go to a cafe and have a few people
> give comments on a design. There is no reason the Gnome community
> couldn't have a central place for similar tests that would create an
> easy way to look through all the data people gather. 

The example I pointed out allows cross-tabulation, so an ethnographic
study wouldn't be so difficult. Just have a few cross-questions like
"How often do you use your computer?" and "Describe your level of
computer proficiency" with a few basic options.

That way you could break it down into clusters of user types. Also,
surveys could be made that are simply not published on gnome's front
page, but on various websites.

As for micro-ethnographic and observational studies involving 3rd
parties, there's already something like this here:

http://developer.gnome.org/projects/gup/usertesting.html

The problem is that different studies gather different data, and have
different focuses, and make different reports. I suspect however, that
it would be appropriate to formulate a standard for gathering reports.

Asking study conductors very nicely for a simple CSV sheet of their
results would be more than enough to start with, in case someone so
inclined ever wanted to put something larger together.

> In any case, I think this is a great idea and will be of huge use to the
> gnome and free software community in general. 
> 
> Eric

Heh, thanks I guess. I'm a lazy raging pessimist, though. It'd need 2 or
3 people to get it up and running, and make it look right, maintain, and
adjust to suit Gnome. If done right, though, it could probably give all
of Gnome a good idea of "Where to go from here". But of course I think
that, it's my idea.

-Jason

(I hope evolution replies to this right...[holds breath])