Sections 17.2.1 and 17.2.2
give two reasons for using formal methods in HCI: communication
and analysis. These are focused on the sort of mathematical
models found in chapter 17. However, there are other
sort of 'formal' modelling in HCI: dialog notations
are formal models of the syntax of the humancomputer
conversation, hierarchical task analysis is a formalisation
of the task structure, some cognitive models are effectively
formal models of the user's mind.
- Are communication
and analysis reasons for using these other sorts of
- Can you think of other reasons why you
would or would not use formalisms?
- Try to use these
pros and cons to formulate issues in the choice of appropriate
forms of formal model and analysis.
answer available for tutors only
Very similar arguments apply
to dialog notations as to more mathematical notations.
HTAs are certainly used for communication, but perhaps
not so much for analysis. Cognitive models are aimed
more towards analysis (e.g. learning complexity with
TAG, timing with KLM).
Some additional cons:
- difficult to learn
- time consuming (and
- only understood by a few and
hence not so good for communication
- can save time (see 'It works!'
page 561 of Human-Computer Interaction)
- helps discover potential usability problems
at design stage (see section 16.6)
- analysis can
be automated (see box 'automatic analysis' on page 586)
Some issues in choosing an appropriate formalism
- Who is involved?
- Do they know
any particular techniques?
- Will they repeatedly
use the same techniques or is this a one-off need?
complex is the problem?
- Are there likely to
be difficult issues that it is hard to solve using informal
- How costly will mistakes be?