21 Mar 1998 ............... Length about 3,000 words (20,000 bytes).
This is a WWW document maintained by Steve Draper,
installed at http://www.psy.gla.ac.uk/~steve/HCI/tw.html.
You may copy it. How to refer to it.
up to workshop main page)
Day 1 topics (HCI teaching workshop)
by Steve Draper
This workshop is about how we do teach HCI and how we should teach it.
The general plan is bottom up rather than top down in two dimensions of scale
and abstraction: from small pieces of a course (e.g. what small student
exercises do we use) to the structure of whole degrees, and from the specific
experiences of our teaching so far by induction to abstract principles. We
hope to go away with some general ideas that might direct new practice in
future, but we think it will make it more "real" to start by grounding that in
sharing what we really do and what we have actually experienced. Having said
that, a leavening of suggestions from those who haven't yet designed HCI
courses may be useful. Plans without experience of the practical problems may
be open to many criticisms, but it is also true (at least of me) that too much
bruising from practical problems can lead to doing some things without
believing in them because it seems to difficult to fight them. An interplay of
principles and experience of problems should be fruitful.
While a provisional timetable is available on another page, here the
discussion topics are discussed in detail. It would definitely help
discussions if you prepared and brought a brief description of your own under
each of the following headings. If you put them on the web in advance and send
me the URL, I will link them in. If you don't have something to describe under
a heading, bring questions or proposals for how you would do that activity, or
reasons why you wouldn't want to.
Many HCI courses have a large exercise as a central feature involving
implementing, testing, and revising a user interface. Do you do this? If so,
what are the details of how you do it? For instance, I discovered recently
that another teacher requires students to submit two sets of code (from before
and after user testing and revision), while I only require them to describe
problems found and changes made. Details like this affect how students pace
themselves in a large exercise, and hence the resulting quality. I for
instance have learned to demand an early report on initial work, just to make
them start work in time. Other important background details are who the
students are, especially their level of programming skill, and what language or
environment they use for this exercise. What do people do for non-programming
students? (I am considering a minimal manual exercise, which involves the same
prototyping cycle, but for text rather than code.)
My own exercise is "ExA" in the table this link goes
See also Terry Anderson's answers to all this.
What other exercises do people use, and how useful do they find them?
For instance, I have an exercise for teaching non-psychologists
basic questionnaire and interview techniques. Work by Phil and others on the
has developed some exercises on design notations like UAN and statecharts.
Pointer to all my exercises.
UAN (User Action Notation)
See also Terry Anderson's answers to all this.
Currently in many universities we are asked to provide learning aims and
objectives for courses and for lectures but not for exercises. But what do we
hope to achieve in providing exercises?
My own best current attempt at my rationale for the central prototyping
exercise I put on is: "To give them the experience of trying to do a good
interface design, and then see users fail on it; and furthermore, to find that
making changes to improve the interface in the light of this is fairly easy."
This is not really "to give them practice in a technique", but more of trying
to change their attitudes through an experience. If taken seriously it also
leads to problems in assessment, since I don't care about the quality of their
design but in whether they identified with it and then witnessed it failing. A
student with above average intuitions could fail to have this experience
through being too good at design: would produce good "work" here yet go out
believing that user testing wasn't important -- a fatal error.
What are your aims for such a central prototyping exercise, if you run one?
(Or: what are your reasons for not having one?)
Currently conventional requirements on teachers are to produce a set of
testable learning objectives for a course, indeed for each lecture.
In addition, and easily consistent with that, are feelings that:
But what do we each think really matters in HCI? In my view, it is something
like an attitude change that is the one important thing I would like to achieve
for each student: something that could be described as "user-centered", but
which is nothing to do with mouthing such a slogan, nor even being able to
debate the good and bad senses of the slogan, but to do with their reaction to
any given design problem.
- A course "ought to" cover salient or fashionable topics (e.g. multimedia)
- It's nice to teach one's own research and theoretical interests, whether or
not they are useful; and perhaps famous theories like GOMS, again whether or
not they are useful.
- Computer scientists like to teach methods because it makes them and the
students feel scientific and professional. Actually of course real science is
based on evidence, and there is almost no evidence, or effort to produce
evidence, that these "methods" lead to better usability.
What in your view is the most fundamental aim of an HCI course?
(My own current compromise over these conflicting issues is in
the official description of my HCI module.)
A related question is to ask, for each topic we teach in HCI, whether it
is actually useful to students:
- Is it useful to students in their design and implementation work?
Do they or could they use it in design exercises?
- Is it useful in the jobs they may take in HCI?
- Is it justified by its relationship to your main overall aim for the course?
Much HCI "theory" fails this
test, even though it interests us and is straightforward to teach and
The MANTCHI project (for links, see below) is exploring collaborative HCI
teaching at a distance, with some exercises developed and offered by a remote
expert. A feature is that this is done reciprocally: teachers exchange
these. The question is, does this save enough work to be worthwhile? Are
there other compensatory advantages? The problem is that some work always
has to be done by the local teacher, and even more than expository material
exercises need adapting to local conditions. Still, that is true of using a
textbook or any material at all. In favour is the basic fact of university
organisation: experts (specialised researchers) are scattered round the
country, but have to teach "complete" HCI courses not just their speciality.
Most of us know of topics we think would be good for students, but would like
someone more experienced to do, or at least to design for us. As
communications get radically better (faster internet, WWW, video
conferencing), personal involvement of the remote expert also becomes
possible without travel (and the uneconomic time penalties that involves).
What do you think about this? Is it likely that HCI teaching could become much
more collaborative, with units authored in rotation round many sites; and your
teaching duties being partly local management of these imported units plus
offering one of your own to a number of remote sites?
Some links to MANTCHI
Which books do you use, and why? and how? We hope to assemble
mini-reviews that will together constitute advice on what to buy/use, how to
use them (i.e. which bits are valuable in teaching and in what way), and more
generally what it is that textbook users need and want.
In addition to textbooks, journal articles and other print-based material,
many of us use a variety of other teaching and learning resources,
including video courses and video-delivered demonstrations of systems),
multimedia courses, web-delivered lectures and seminars, web-based course
materials, etc. Are these sorts of resource especially important for HCI
teaching, as opposed to other topics? How do you exploit these resources?
Are there resources which you don't now use which you would like to use? If
so, what are they and what is preventing you using them now? Are there new
sorts of teaching/learning resources in HCI which you anticipate becoming
relevant to your teaching in the near future?
We ask you to prepare for the workshop to make the discussions go
- Please email me
stating your degree of experience in teaching HCI.
Example 1: I have taught HCI for 10 years, running a whole course; but apart
from a very few guest lectures and industrial courses, it has always been to
the same audience (a conversion MSc).
Example 2: Phil Gray has taught HCI for about 10 years, but at 5 levels (4
undergraduate, and MSc level), although only to computer science students,
plus industrial courses.
(On the first day, much of the time will be spent in small group discussions,
and my experience is that these will be most interesting if we mix people as
much as possible, so that their past experience is as different as possible.)
- We shall produce a report from this workshop (see
Please come prepared to volunteer to write a short report on a session.
- The topics for the first day are described above.
Please come ready to summarise your practice or view on each of these eight
topics to other participants.
- (The second day, coordinated by Chris Johnson, will consist of "debate" type
discussions suggested by your position papers. They will be started by
individuals who have agreed to this in advance.)
Terry Anderson's answers to all this.
up to workshop main page)