[412] in Athena User Interface
Re: Beta, system, and usability testing, p.s.
daemon@ATHENA.MIT.EDU (Bill Cattey)
Mon Sep 11 19:22:49 2000
Message-ID: <QtjKZ6pz0001Nsttkw@mit.edu>
Date: Mon, 11 Sep 2000 23:22:46 +0000 ()
From: Bill Cattey <wdc@MIT.EDU>
To: aui@MIT.EDU, beland@MIT.EDU
In-Reply-To: <200009081613.MAA02150@Press-Your-Luck.mit.edu>
---------- Forwarded message begins here ----------Suggestion to the team:
Let's all try and be a little more respectful of each other's suggestions
about what is proposed as THE user interface, and how we reach out to
others as we work this issue.
I've been watching people learn to use emacs as an MIT employee for in
excess of 15 years, but I still temper my usability suggestions with the
awareness that I am generalizing from my own personal experience.
One thing I came away from that experience with was the notion that, no
matter what I designed, there would be SOMEONE who found none of my
effort helpful, and some of it at odds with their ability to make sense
of the program.
Everyone please remember that usability testing is about making a better
guess about where to put controls so that the majority of users find the
ones they need most easily, but are not confused or intimidated by the
ones they don't need.
It may be that the majority of the intended audience of the new AUI
would know that there's Notepad and Microsoft Word under windows, and
understand intuitively, from the get-go that there are different editors
for different tasks. In that model of the user-universe, it would make
sense to offer StarOffice, gEdit, and Emacs as separate editors, and to
make something like gEdit the default.
Personally, I don't like gEdit because every time I tried to use it it
core dumped on me so I quit trying. But it's possible that it's working
better now.
At this point, I'd be content to leave Emacs as the default because with
all it's failures in usability, it's the one editor we best understand
how to teach about, and it seems to do the best job of not core dumping.
gEdit offers the possibility that the toolkit will supply a nice,
intuitive editor component for use in many applications. I'm still
remembering the pain of Motif's failure in that regard to feel
comfortable recommending gEdit. But I KNOW I'm over-generalizing from
my past experience.
Again with the experience thing: I'm uncomfortable with a multi-line
prompt. But it's only because I'm making a guess for how people
interpret the output of the computer, and because I can think of too
many situations where bad coding prevents a multi-line prompt from doing
the right thing.
By the same token, I think that long directory path output may be a problem.
So everyone is entitled to disagree with my opinions about what we
should do, but I would really like people to try and not call each
other's idea "dumb", ok?
-wdc
