The paper gives a succinct description of pattern languages
and a nice overview of their development.
Pattern languages have a formality that yields
good reuse and knowledge transfer properties;
these features are desperately needed in a field such as HCI,
where one often feels lost among vague claims, fuzzy descriptions,
and battles of aesthetic opinion.
The paper provides good examples in multiple domains
(music, interaction, and software)
to support the usefulness of pattern languages.
WEAKNESSES
The proposed hypertext model
is limited to a hierarchical structure:
no sideways cross-references are allowed,
only downward references;
cyclic chains of references are not representable.
The section on applying patterns
in Nielsen's usability engineering life cycle seems weak;
certainly patterns are applicable to some of the 11 phases
that Nielsen identifies,
but the author seems to be really
reaching to try to find a relevance for patterns in every single phase.
In some situations,
the pattern language approach may prove too rigid
for design and communication.
The paper may be a bit quick to assume that pattern languages
are the right answer to every question.
zesty
ca)