Thursday, October 31, 2002

the problem of representation

i read Cillier, and i agree at quite many points. especially those arguments about modernist approach to complex system; we need theory before we program anything. i think i had tried to do this in the whole first year of my study, trying to identify the rules. through Cillier's i say that configurational theory of space is a modernist approach, a simplification of spatial problem. the theory may not be suitable as the basis for generative project. the unlimited layout/configuration possibilities and indefinite numbers of possible activities had ruled out space as simple problem.

however the big problem is that if architecture is complex, within connectionism perspective; it means we need as much complex program to model it. i personally believe that architecture is complex but not as complex as natural beings. architecture is artifact, from the beginning man had succesfully creates architecture, thus it must not be as complex as living things. so where is it? where to sufficiently analyse this?

the problem of observation

paul asked if we could define the observation level, if we would see the model we've tried to develop as a co-evolve, autopoietic-kind of system. i remember to read that philosophy is about organisation and interpretation of experience as objective as we can, which the first step to do this, is to define the observer. experiencing means that the observer is the part of the whole observation proces itself, thus what can we actually do is to say to what extent can we separate the observer from the observation. i summarise this below from the same reading:


1 from Immanuel Kant, Critique of Pure Reason, its the nature of human mind to observe things as spatial and temporal and having certain categories superimposed into the observation.

2. from Heisenberg, Physics and Philosophy, the observer always affects the observation

thus, in light to the above information, i would argue that an autopoietic project is actually a very subjective project, and/or autopoietic models is a medium for philosophical utterance. the basic argument in autopoietic theory is that the behaviour is not emerge from interactions between elements in the model, but from what has been observed by the observer. accordingly, it would be relatively easy to argue that while we can see the apparent similarities of Hillier's thesis (ie. configuration of space influence and can be influenced by configuration of people) and that to autopoietic models, that the project should go on to model spatial configuration under Hillier's terms, autopoietically.
then the argument about modelling spatial configurations will roll on to the point that, architecture problem is not simplistic, it is actually complex, having to embed in it the very complex things of all, human beings. here, i have to be brave enough to argue that Cilliers approach, ie. the post-structural approach to complexity, while complex phenomenon needs not single method to understand it, but in the end, choices have to be made when we come to actually create the representation of it. this will bring us to the problem of representation

Sunday, October 27, 2002

Anderson's Problem Solving and Problem Worrying: summary

Problem Worrying is an antithesis to problem solving, instead of having confidence that every problem can be identified, and thus can be solved; one needs to try to understand more about the problem. what it needs to be done in design is a deeper understanding of the problem, continuously worry with the problem. with this cumulative knowledge of the problem, he/she then constructs the environment to support the success of human purposes, thru relating reformulation of problem to form.
Anderson model of architecture design process:
1. generalised understanding of the problem
2. production of various formal proposal
3. study of the implications of proposals
4. successive reformulation of problem and proposal
5. selection of form for its appropriateness to formulated problem
Anderson acknowledged that problem solving techniques which utilises a highly sensitive feedback mechanism within an appropriate domain is actually close to the idea of problem worrying.
other post move to here

Problem Solving and Problem Worrying continues...

Among them is Christopher J Jones, with his book Design Methods published in 1970. This book replies to Anderson extensively, that a problem solving approach is the way forward. William J Mitchell follows Jones's path and thus produce The Logic of Design, 1990. This book actually summarised design as a computable activity, a process with finite number of procedures (using shape grammar rules) to finally produce architecture design. Up until now there are no widely use and known design software that applies Mitchell theory of architecture design by computer. what have been acknowledge by designers in common are expert system types of design software like 3D studio. yet theoretical development continues.

John Gero, having quite many research interests and publications, had dwelled on topic such as optimization (one of important problem solving technique) in architecture design, and recently (2002 research publications) directing his research to situated design which employs multi-agent system. The concept of situated design is compatible to, (if not part of, need to confirm on this) Peter Senge's philosophy of learning organisation. this brings us back to Anderson problem worrying, that architecture design is more about learning, understanding more what the problem is. the lesson learned from these research shows strong parallelism if not mirroring, the development of computing science. somehow, architectural scientist opt to map architecture ideology to the advancement of tool, ie. computers. Within his text, Anderson did not show any particular attention to the development of machine, while around that time (1960's) the idea of intelligent machine, has been quite strong although not influential.

Friday, October 25, 2002

this post move to here

Thursday, October 24, 2002

acknowledgement to Levik move to here

Problem Solving and Problem Worrying, Stanford Anderson

The problem of Problem Solving approach for architecture, originates in the nature of architecture design. According to Anderson, architecture design is about structuring environments so as to facilitate the achievement of human purposes. Purposes are problematic because they are incompletely known, not all aspects of them can be listed, and they can not be extrapolated from known purposes. Some purposes can be stated, but lots of others can not, and impossible to state. Moreover, purposes always altered by environments created to facilitate them. The problem of design is to structure environments with yet incomplete list of purposes that somehow relate both ways to the environments itself. But the design must go on.

Architectural methods also the origin of design problem. There is no convention on how architect should design, no apparent method. What happens when architect design is a series of tentative foresight taken place, to follow by critical examinations of that foresight, then actions based on these examinations that made up the design. Thus, it is obvious that architecture method is definitely undetermined in advance. The process is dynamic and very much dependent to relation between the problem and the form which the architect is trying to define. Thus, the best approach to architecture design is to understand more about the problem and the relation between problem and form, not to solve it. This is Anderson's idea of problem worrying.

He made an interesting point, while not fully investigate problem solving applications in architecture. I seen this as a valuable source for anyone who wants to embark on design automation project. He states that there are problem solving approaches to architecture that close to the idea of problem worrying, i.e. ones that involved a highly sensitive feedback mechanism and operating in "appropriate" domain. Anderson wrote this article in 1966, the time when some famous architecture scientist is about to emerge. Scientist in a sense they urged for methods, and justification. Among them

Wednesday, October 23, 2002

this post move to here

Tuesday, October 22, 2002

acknowledgement to Zaeem move to here

notes on 17 oct 02 uel circle


half way in re-organising my paper files, apparently i got lots of data to support design automation. if we want to give some valuable criticism to this field, we need at least equal amount of data. the data contains sound theoretical back-ups and lots of research articles about design automation: implementation of problem solving approach in architecture. so how come i didn't mention this in last seminar? i keep forgetting things. prolly robert can teach me his remembering techniques.

mm mmm and i'm looking forward to add bits of bloglinker code from Zaeem, so i can easily keep all these in the form of links rather than print outs. i never open any saved links in my pc, or in IE favorites: bcoz its so boring that way, not much help with my study. part of it bcoz i believe, i learn visually. i can't see the meaning, if i don't see them all.

i definitely need to review anderson's paper, problem worrying. i read it before, but it just doesn't make any sense since all i read was about the possibility to automate design and how those people had done it. thus, when robert argued so strong about it in the last meeting, he finally showed me that i may think of it as a probability rather than a fact; and with paul and robert both referring to anderson, i just feel that i had all the spirit again to search and search. this study suddenly becomes interesting, again, as once before.

Monday, October 21, 2002

this post move to here
neither the human purposes nor the architect's method are fully known in advance. Consequently, if this interpretation of the architectural problem situation is accepted, any problem-solving technique that relies on explicit problem definition, on distinct goal orientation, on data collection, or even on non-adaptive algorithms will distort the design process and the human purposes involved. Stanford Anderson, Problem-Solving and Problem-Worrying.
problem: is this mean an adaptive algorithms will do? the latest from holland is cgp-v. generating complexity with exact formula. mmm mmm and classifier is its predecessor, important moment in adaptive system development.

Saturday, October 19, 2002

why blog move why i blog