[Fwd: [self-interest] prototypical analysis and design]
lourenci at lsi.usp.br
Wed Nov 22 15:42:16 UTC 2000
Well I am overstressed and I may be coopting what I read to tune with
what I expect but let us examine this passage from Günther's PhD
thesis (page 49-50)
Class-based systems. Traditional class-based systems do
not effectively support behaviour evolution.
The class-based model cannot easily express changes
in the structure or behaviour of an object. Required behaviour
evolution has to be hard-coded into application programs.
As seen in the previous section, manually coded behaviour
evolution is a possible workaround for the limitations of the
class-based model but not a satisfactory solution.
The essence of various design patterns suggested in this
context are more or less faithful simulations of delegation. These
simulations do not achieve the full functionality of delegation (3.6.3)
Because the functionality missing from the language is
simulated by a set of cooperating classes, these classes
tend to be tightly dependent on each other in ways that are
not grounded in the application but just in the technical details of the
pattern. Such additional dependencies and assumptions built into
the design require consistent changes in a complete hierarchy
of classes even for small unanticipated changes int he application
logic, e.g. addition of a method to a class of addition of a new
class (see 184.108.40.206 and 3.6.3). Thus the application of design
patterns for simulating BEHAVIOUR EVOLUTION CAN
IMPEDE REUSE (OF EXISTING CODE AND DESIGNS)
AND COMPLICATE PROGRAM MAINTENANCE, achieveing
\the opposite of what is widely regarded as a main benefit of OOP.
Therefore, extension of traditional class-based object models
by a mechanism for unanticipated behaviour evolution is
I invite Günther to throw light in this passage. But I interpreted
as delegation is better than design patterns.
And I put forward that both Christopher Alexander and I recognize
the failure of "A pattern language" to grasp the essence of
the nature of architecture and urban design. The granularity
of the patterns is too coarse. The granularity of my ecodesign
model is fine. I would say it will be very hard to build a finer
ecodesign model. It grasps the macro as well as the micro aspects
involved in the design process. I mean it deals with modeling
geoengineering and climate as well as with comfort zone and
the ergonomics of the behavior of the human being.
If architecture is not translated into forms it is not architecture
And there is no geometric modelling in A Pattern language.
My geometric modelling consists of the theory of tilings,
(based on the graphical artist Maurits C. Escher , I extended
the theory of tilings with the notion of prototiles, tiles of
different formats - so I can reproduce through this concept
exactly what an organic designer would do. I mean I can
layout the behavior of the human being around the furniture and
crystalize it with this concept), symmetry groups of the plane
and the dotless plane. Similarity and conformal symmetry
groups are fractals!!!The two most importants models to
model phenomena are catastrophe theory and chaos theory
(as far as I know fractals are within chaos theory context).
Through the relationship of the subgroups of the crystallographic
groups of the plane as well as the notion of metamorphosis
so well grasped by Escher in his xylogravures I arrived at the
essence of the free plan.
I mean I am building a knowledge based system in a seamless
process. The domain dependent model follows the geometric
modeling and I hope through delegation I can achieve a
computation model that is "isomorphic" to the previous two.
I love Alexander's work although I have read his book only
after I had finished my master's dissertation in 1988. And I
was already involved with Lionel March the head of the
Graduate School of Architecture and Urban Planning at UCLA.
Lionel was Alexander's colleague at Cambridge.
He introduced in his PHD thesis in the sixties the notion
'that architects since Leonardo da Vinci have applied
symmmetry groups of the plane to architecture.
In this century Frank Lloyd Wright, the greatest formmaker
of this century as well as Le Corbusier applied symmetry
groups to their designs.
However they applied the traditional notion of tilings.
The outcome of their work is not a free plan, a fundamental
paradigm of the Modern Movement with respect to the
inner layout of an apartment.
Let's feel how Alexander is:
I sleep in a room that has windows in two sides. It is
simply gorgeous. My house is located in a villa that
was built in 1903.
So let's see what Alexander writes on page 745.
prepare to knit the inside of the building to the outside,
by treating the edge between the two as a place in its
own righ, and making human details there:
159. light on two sides of every room
160. building edge
161. sunny place
162. north face
165. opening to the street
167. six-foot balcony
168.connection to the earth
one reads on page 747
...once the building's major rooms are in position, we have to fix
its actual shape: and this we do essentially with the position
of the edge. The edge has got its rough position already from
the overall form of the building -
Wings of Light (107), positivive outdoor space (106)
Long Thin House (109), Cascade of roofs (116).This
pattern now completes the work of Wings of Light (107),
by placing each individual room exactly where it needs to
be to get the light. It forms the exact line of the
building edge, according to the positin of these individual
rooms. The next pattern starts to shape the edge.
Well The two projects I designed to illustrate my design
model curiously if they don't have windows in both sides
light falls from the top of the roof, so one can see the
moon, the sun, the flowers with different colours all around
I love windows at the corners. And I love to let light fall from
I also designed a project for a small church. When I applied
the formula to see if I had calculated the right amount of
area necessary for daylight illumination and natural ventilation
they fit the needs perfectly well.
Indeed Alexander and I were anxious to work together.
However Berkeley was not a center of OOP in 1990 -1992.
The only interesting researcher was STephen Omohundro
the author of the Sather Language, an optimization of the
Eiffel language from International Computer Science
Institute and so the Brazilian agencies made up their minds
not to send me to Berkeley.,
Lionel March and the CAD group finally recognized I was
an expensive student and UCLA could not cope with my
needs, not only in terms of facilities as well as human
as far as I know Günther (I know him) does not claim that delegation is
better than design patterns.
I reckon he says that with delegation most design patterns are easier to
> Kniesel in his thesis wants to show the superiority of delegation but
> I am still reading through his thesis and the papers I gathered and I
> have no conclusion.
I'd be very interested in the list of papers you gathered for this topic
and maybe even your opinion on them.
I am currently looking at delegation as a means to modeling as opposed
to class based modeling how it is usually done. I am be happy to share
findings with you and the list but at the moment I am trying to get a
handle on prototyping, i.e., find out when you would prefer it when
Dr. Thomas Kuehne
0178 4314387, http://www-agce.informatik.uni-kl.de/~kuehne
The difficulty in doing research is to find the right questions
so that all the answers come easily. -- TK
| Albertina Lourenci |
| PhD in Architecture and Urbanism |
| post-doctorate researcher |
| Laboratory of Integrated Systems University of Sao Paulo |
| Avenida Professor Luciano Gualberto, 158 Travessa 3 |
| CEP: 05508-900 |
| Sao Paulo Sao Paulo State Brazil |
| Voice: +55 011 818 5254 |
| Fax: +55 11 211 4574 |
-------------- next part --------------
An HTML attachment was scrubbed...
-------------- next part --------------
An embedded message was scrubbed...
From: Thomas Kuehne <kuehne at informatik.uni-kl.de>
Subject: Re: [self-interest] prototypical analysis and design
Date: Tue, 21 Nov 2000 08:33:18 +0100
More information about the Self-interest