Cognition: Memory, Attention, and Learning - Foundations for Designing User-Centered Systems - page 132

available, or could help support users by noting the interconnected nature of most
systems. Multiple choice exams can, with their questions, remind you of com-
ponents of the answer. For example, a question that asked you ''Which of the
following is a type of memory?'' implies that there are several types. This type of
test could help create false memories.
If used directly, this effect can help users. Queries to users that provide context
to the question or even a default action can help users.
5.2.5 Implications for System Design
What we already know about human memory provides numerous implications for
improving the lot of the user. Users will be able to learn interfaces that are similar
to other interfaces or other knowledge structures more easily than completely new
interfaces. Users are limited in the amount of new information that they can
perceive, comprehend, and learn. A fairly direct implication is to use words that
users know, and use the words consistently to strengthen the chances of later
successfully retrieving these words from memory.
It has long been known that retrieving names from memory is faster than
naming objects. This suggests that, instead of displaying icons, we might be better
served by displaying words (Chilton 1996 ). The approach, of using names instead
of icons would also help the visually impaired, who rely on software to translate
computer interfaces into verbal representations. Raskin ( 2000 , Sect. 6.3), provides
further interesting examples and arguments encouraging the use of words instead
of icons. Figure 5.3 shows icons used for translation during international travel, as
well as some icons from interfaces that are more difficult to interpret.
Our understanding of memory can and should influence password choice. This
knowledge can inform more than ''experience and common sense,'' which is
sometimes used. Users will want to choose passwords that are hard to guess, and
systems may enforce this. However, many passwords that are hard to guess are
arbitrary (that is, not meaningful to users) strings of letters, digits, and punctuation,
such as ''ouibou94!3''). Thus, some strong passwords are hard to recall when users
need them. There have been surveys on what passwords people use when they are
not given any guidance (the results are pretty scary from a security standpoint).
What needs to happen more is helping users generate mnemonics for remembering
the passwords, rather than writing them down on paper or installing them in a
computer-based passport (unless you trust the computer company selling the
passport, which many state attorney generals do not). The issue of recallable yet
strong passwords is one of the topics routinely dealt with under the auspices of the
topic of usability of security. Some analysts have argued for word-based pass-
words, such as 'islandcarekeyboard', which is long and thus hard to guess, but
easier to recall as each chunk is several letters rather than a single letter.
There are several poorly understood but important questions relating to the use
of memory with respect to system design: