Re: Call for constructive user criticism.

[Bruce Stephens <bruce cenderis demon co uk>]

Havoc Pennington <hp@redhat.com> writes:

> On 1 Aug 1999, Bruce Stephens wrote:
> > I agree that a partnership would be good.  Why not go for the
> > University of Michigan, in the form of the Linux/UNIX Independent
> > Group for Usability Information, <URL:http://www.luigui.org/>, which
> > might have been created for this sort of thing?
> > 
> 
> luigui's site is down, and their mailing list has had about 5
> messages in the last several months, so I don't think they're very
> active.

Well, the site was there when I looked!  I agree that the project
seems a bit dead, but perhaps that's a matter of timing---perhaps
there'll be things happening in the coming academic year.

> The problem with HCI departments is the same problem we had with
> gnome-gui-list: lots of pie-in-the-sky "let's rewrite the whole damn
> thing" goo, not many small, concrete, workable improvements to the
> existing codebase. This is what we need. I don't think small tweaks
> to an essentially windows-icon-mouse UI are going to interest HCI
> researchers.  I could be wrong.

The person that announced the project (Nathaniel Borenstein) seems a
pragmatic guy, judging by his book ("Programming as though people
mattered: friendly programs, software engineering, and other noble
delusions").  Although the Right Way to develop programs is presumably
to involve HCI practitioners and end users right at the beginning,
presumably HCI students need to learn how to evaluate and suggest
improvements to working designs, too.  Presumably they also need to
learn to engage in dialogs with developers, to help choose the most
cost-effective improvements.

[...]

> The challenge of this initiative is to get concrete, specific
> suggestions for enhancing apps. If it loses that focus, it's
> worthless.

I think the suggestions already made are good ones: fix bugs, and
improve documentation.

For really specific questions, user surveys can work OK.  Real
experimentation is better, but user surveys can work.  RMS used a few
for Emacs, for example, and I think they went OK.