Thanks to Stephen's note, I have read his paper with Simone Gristwood and
the PhD thesis by Bruce Archer.
Terry will really like the thesis. I did too, but I have an important
reservation and distinction to make.
Boyd Davis, S., & Gristwood, S. (2016). The Structure of Design Processes:
Ideal and Reality in Bruce Archer’s 1968 Doctoral Thesis. Paper presented
at the Proceedings of DRS 2016, Design Research Society 50th Anniversary
Conference. from https://drs2016.squarespace.com/s/240-Boyd-Davis.pdf
Archer, L. B. (1968). The structure of design processes. Royal College of
Art.
http://researchonline.rca.ac.uk/2949/1/Bruce_Archer_Structure_of_Design_Processes_1968.pdf
It is incredibly formal, filled with mathematics (mostly algebra) and a
fascinating bit of history: the thesis is typed on a typewriter, with
numerous hand-written annotations, some corrections, and a lot of
handwriting to take over the equations where the 1968 typewriter could not
cope. (It must have taken a month to type -- the thesis is almost 200
pages of lots of equations).
I include excerpts from the thesis below. There may be various typos, but
they won't change the major point I wish to make. (The typos are because I
copied and pasted from the PDF, and the ancient typewritten characters did
not get translated very well into computer fonts. So ended up as !. Also,
where I say P1. the 1 should be a subscript, but i couldn't figure out how
to insert it into my browser.)
Still, I include a lot of the test to illustrate the way the thesis went.
Yes, very logical. But will it help design the controls for a washing
machine: Nope. Will it help select a font? Nope. Will it help make the
form of the product more beautiful? No.
Would it help design the washing machine itself so that it perfumed the
task of cleaning clothes effectively? Yes, it probably would. That is
important, of course, but that is Engineering Design.
Note that the department which granted the PhD was called (both written and
stamped, with what I imagine to be a big, red rubber stamp):
School of Industrial Design (Engineering)
I have nothing against Design Engineeri9ng: it is very important. It simply
is not what many of us on this particular list do (I may be wrong, but
those of you who are engineering designers are mostly quiet).
I myself have two degrees in engineering and am a member of the U.S.
Mathematical Academy of Engineering. How nice of them to allow a view of
us human-centered folks into the esteemed academy. (One of my students made
it to the Royal Society of Engineering! Hey!). But Engineering design is
not what I do (nor what she does, that is Abi Sellen, my former student)
The mathematics, by the way, is mostly empty symbolism. Consider this on
page 93:
21 THE PERFORMANCE
corresponding to a given *proposal* is the set of states of *properties* which
are exhibited in the outcome, or which would be exhibited in the outcome if
the proposal were to "be implemented.
{P1(x), P2(x) ..... Pn (x)} signifies a performance x of a set of
properties P1,P2....Pn.
{P1(x), P2(x) ..... Pn (x)} = {f D1(i), D2(i) ... Dn(i)} where f signifies
some function of proposal i.
basically empty of content, but to folks who don't know math, very
impressive.
It is all very logical, very methodical, very dull. And for the kind of
design I do, very irrelevant. (I suspect Archer would agree, probably with
all those statements).
The real problem is that it is engineering design. And yes, engineering
design is very important and critical. However, it is not the design most
of us talk about. There is basically zero discussion about human needs,
emotions, or for that matter how people actually would use the results of
the designs. Engineering designers, such as Archer, do mention these words
now and then, but with no real depth.
Early Herb Simon would have liked the thesis (and yes, he is referenced).
The Herb Simon in his later years would have felt it insufficient. If you
look at the references, you will find lots of formalists. Indeed, my Ph.D.
thesis advisor, R. Duncan Luce, gets four references for his work in Game
and Decision Theory (Luce's PhD was in Mathematics from MIT).
Of course, Archer makes use of solution spaces. Just read his summary:
Section 11. (Page 84 of the PDF. The ms. pages are not numbered!) Within
this framework, a design problem is defined as follows;
11 Summary and conclusions
11.1 The arguments and hypotheses set out in this thesis are based upon the
observation and analysis of a number of case studies, listed in the
Appendix. Concepts developed in neighbouring disciplines have been employed
to illuminate the nature of the design activity. It is concluded that
design is a specific example of a general pattern of problemsolving
behaviour,
11.2 All problem-solving activities are goal-directed,- The correct
identification of the goals, as well as the discovery of means for
attaining them, is an essential part of the activity.
11.3 Most design problem-solving activities are carried out within the
framework of larger endeavours, and some, at least, of the objectives of
the design problems reflect the goals of the encompassing problems.
11.4 More than one person or body of persons may be in a position to
nominate the objectives of a design problem, and more than one person or
body of persons may be empowered to contribute decisions in its solution.
The group of people so empowered constitute a coalition formed for the
pursuit of common, allied - or even, in part, opposed - ends. In such a
situation the procedure is a dynamic one, where each decision is made, and
any objective may be reviewed, by common consent, according to the aims and
bargaining powers of the participants.
11.5 Many design problem solving activities are carried out in opposition
to competitors. It can be .argued that in the absence of human competitors,
a project is conducted 'against nature'. In every case, therefore, the
design activity is dynamic and spread over time,, each decision being taken
in the light of the actual or potential counteractions open to competitors.
11.6 Within this framework, a design problem is defined as follows;
*Given (by the arbiter(s))*
A set of properties which are required to be exhibited in the end result; a
specification of the ways in which varying states of the various properties
will be regarded as providing varying degrees of satisfaction, including
any limits to the acceptable ranges of states of the properties; a
specification of the relative importance which is attached to the
satisfaction of the various requirements; the set of decision variables
which are to be accessible to the decision maker(s), and which govern or
partly govern the states of the properties! the limits to the ranges of
states of the decision variables which are to be open to the discretion of
the decision maker(s); and the set of context variables which are NOT to be
at the discretion of the decision maker(s), but which partly govern the
states of the properties.
*It is required (of the decision maker(s)) *
to select a set of states of the decision variables such that the
satisfaction of the requirements, weighted according to their importance,
is optimised:
*Provided that *
the states selected for the decision variables are mutually compatible; the
states selected for the decision variables lie within their respective
limits of discretion; the laws governing the relationships between decision
variables, context variables and properties are complied with; the
resultant states of the properties are mutually compatible; and that none
of the states of the properties lies outside their respective limits of
acceptability
11.7 The function of conceiving solution hypotheses, or design ideas, has
hitherto been regarded as the principal 'human' contribution to designing.
Advanced automatic data processing techniques, however, offer the
possibility of exploring mechanically all the feasible combinations of
va.lues for given decision variables. Creativity in this sense might well,
therefore, become a less valuable human facility as time goes on,
11.8 The functions of nominating objectives and rating them for importance
remain essentially human and largely subjective. Social science techniques
can be employed to discover the value systems of individuals or classes of
people, and logical tools can be employed to compute with the data
obtained, but those are only means for making explicit the human and
subjective.value judgements about goals, which must be made by the people
concerned. Perceptiveness in discerning worthwhile objectives, and
judgement in rating them for importance, might well become more valuable a
human facility as time goes on. A major deficiency in the present state of
the art of designing is the almost complete lack of data on the value
systems extant in the community, or employed in design practice.
On Sat, Feb 24, 2018 at 10:12 AM, Stephen Boyd Davis <
[log in to unmask]> wrote:
> Dear all,
>
> Sorry to come so late to this. For those interested in a historical
> perspective, here https://drs2016.squarespace.com/240 Simone Gristwood
> and I discuss Bruce Archer's changing views on the potential to systematise
> (or even scientise) the design process.
>
> Specifically we note Archer's early apparent belief that the requirements
> could be largely finalised in advance of making -- and his emerging
> realisation that "any effective design procedure must ... permit radical
> reappraisal of the problem at any stage." (Archer 1968: 6:17).
>
> Archer's 1968 doctoral thesis is here: http://researchonline.rca.ac.
> uk/2949/ .
>
> You will see (as I mentioned on this list a couple of years ago) that
> Archer makes use of the "solution space" idea, most visibly in Figure 2.18
> on page 117. He says in Section 2.24 (p.17), "The limiting states of the
> properties concerned may be similarly set out according to the conventions
> of co-ordinates (fig 2.16). The spaces marked off by these limits indicate
> the field of mutually acceptable degrees of fulfilment of the co-existing
> objectives. Any solution whose mutual states of the associated properties
> lie within this field is an acceptable solution."
>
>
>
-----------------------------------------------------------------
PhD-Design mailing list <[log in to unmask]>
Discussion of PhD studies and related research in Design
Subscribe or Unsubscribe at https://www.jiscmail.ac.uk/phd-design
-----------------------------------------------------------------
|
