|
Author
|
Topic: Tinkering Design
|
Micah Sparacio
Member
Member # 6
|
posted 30. November 2003 08:56
As usual, my thoughts are going to be dispersed throughout this thread in a relatively sloppy manner (off the top of my head).
Tinkering Design
Doubts have been cast regarding the weak relationship that intelligent design theorists insist can exist between designed objects and their designers.
One example that I can think of off the top of my head is gedanken's "motive" thread, where it is argued that a stronger relationship exists between the designer's motive and the designed product.
While I won't re-address that issue, I am interested in developing a thread around the notion of a "tinkering designer."
The tinkering designer concept often has in it several implications that I'd like to enumerate. Not all of these assumptions are contained in everyone's concept (and some contradict others):
1. All design implies tinkering
2. Tinkering is a less-efficient method than single instant design
3. Tinkering implies improvement (rather than mere change) over a previous state of design
4. Tinkering implies that the designer is not significantly better at designing (discovering an optimized state) than, say, a simulated-annealing algorithm
5. Tinkering reduces the significance of cognition in design to not much more than trial and error
Anyway, that's a start. I'm interested in what people think of these five statements and any other insights that might be out there.
To be less than cryptic, my main interest in this thread has to do with design methodologies, and whether the intelligent designer has design options that lay outside the "tinkering" model and if not, "whether tinkering implies bad design" or "whether tinkering can be viewed as a positive design virtue"
[ 30. November 2003, 08:58: Message edited by: Micah Sparacio ]
IP: Logged
|
|
Evan
Member
Member # 164
|
posted 30. November 2003 10:02
Tinkering is how people design. One of our tinkering skills is called "thinking", where we play around with internal and abstract representations of the world, trying to get a feel for what might work. Then we start to externalize our thoughts - this very post is an example of tinkering as I reproduce the flow of my thoughts in externally represented words to see, somewhat experimentally, what I think about this topic.
All designers do this - they make sketches, write rough drafts, make models, etc.; see what they look like and how they work; and then move the tinkering back "inside their head" until it's time to fiddle around with the external world.
Design is a tinkering process - I think it's a most apt phrase.
IP: Logged
|
|
Steve Petermann
Member
Member # 884
|
posted 30. November 2003 11:35
Micah wrote:
quote:
To be less than cryptic, my main interest in this thread has to do with design methodologies, and whether the intelligent designer has design options that lay outside the "tinkering" model and if not, "whether tinkering implies bad design" or "whether tinkering can be viewed as a positive design virtue"
From my experience "tinkering" is an inherent part of the design process. Whether "tinkering" implies bad design or not depends on scale. Designers by their nature have limitations. They have varying levels of skill and experience. Designs work in the real world and as such are subject to complex environments. This means that there can be an enormous number of causal relationships both inside the design and in interface with the environment. The more complex the design and the environment, the more causal relationships.
Good designers become very good at understanding and predicting causal factors. They also utilize wherever possible existing designs where causal factors are already known. In fact, all designs are built on prior designs. Invariably new designs are just novel combinations of prior lower level components.
Experienced designers have both a good intuitive sense about causal relationships between components and the environment as well as good analytical skills for analyzing these causal relationships. Even experienced designers, however, do have a limit to their ability to deal with complexities. Very complex designs create a matrix of causal relationships that may surpass the intuitive and analytical capabilities of the designer and her tools. This limitation is why tinkering is necessary.
Good designers know these limitations and wherever possible will test their designs before implementing them. If the tests are designed well, then any flaws in the design can be identified and corrected by tinkering.
One test of how good a designer is, is how much tinkering is needed. Designers have to have skills at knowing and predicting causal relationships. The better they can do that, the less tinkering is needed. One should, therefore, expect tinkering even in good design processes.
The other factor that may require tinkering is a change in the design goal(or specification). Often when designs are implemented it is recognized that the initial goal was either not thought out fully or must be modified based on new information. In this case tinkering is necessary to meet the new goal. Good designers also recognize this possibility and will build flexibility into their designs so that tinkering can be done without major redesign.
IP: Logged
|
|
Evan
Member
Member # 164
|
posted 30. November 2003 11:43
Those are all quite good points.
I would like to return to a point that I made, also: thinking is also tinkering. Thinking is internalized, abstract manipulation of images and symbols that substitutes for actually working in the physical world. All the things Steve said about design in the external world is true in some analogous sense about the conceptual aspect of design.
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 30. November 2003 16:01
Micah raises an interesting question namely if the tinkering designer is a 'bad designer' or has a positive design virtue.
In fact I would like to extend this by pointing out that a tinkering designer can include both intelligent designers as well as regularity and chance processes. Thus the appearance of tinkering design in nature leaves us with some interesting questions as to the nature of the designers.
But would that not require some knowledge about pathways, methods, processes?
Steve adds some interesting observations: Designers by their nature have limitations. They have varying levels of skill and experience. Designs work in the real world and as such are subject to complex environments.
Thus an understanding of the limitations of the designers may be helpful in understanding the relevance of tinkering design in nature. What are some of the limitations observed in tinkering design? What do these limitations state about the nature of the designer?
Evan observes that All designers do this - they make sketches, write rough drafts, make models, etc.; see what they look like and how they work; and then move the tinkering back "inside their head" until it's time to fiddle around with the external world.
Can we observe these evidences of tinkering? Is there (indirect) evidence that may help us understand the nature of the tinkering design(ers)?
[ 30. November 2003, 16:05: Message edited by: Pim van Meurs ]
IP: Logged
|
|
Micah Sparacio
Member
Member # 6
|
posted 30. November 2003 17:06
Pim - I'm going to be blunt -> please stay out of my thread.
your method of sabotage is disheartening and deflating to those of us who try to have interesting/civil conversations.
As a response to your post, I don't consider natural selection to be a designer in any meaningful sense of the term "designer". Maybe it has the causal power to mimic agent design, but it certainly is no "agent", and the notion of a designer implies "agency". So, regardless of the causal powers of natural selection, for the purposes of this thread, I'm interested in the tinkering nature of agent designers (not disembodied processes).
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 30. November 2003 17:18
I apologized for having misunderstood Micah's usage of the term designer.
I may have been misled by Micah's comment Doubts have been cast regarding the weak relationship that intelligent design theorists insist can exist between designed objects and their designers. which suggested to me that his comments of tinkering design(ers) were to address the weak relationships.
I conclude from Micah's response that Micah was not talking about designers but rather about intelligent designers.
Given the risk of confusion when using terms like design and designers especially when discussing biology, I propose we use the term intelligent to denote the subset of design(ers) restricted to intelligent agency.
With these limitations in mind, I would like to continue my contributions to this thread in my usual civil manner and ask this interesting question:
Given the observations supporting a tinkering design, what can we infer about the nature of these designer(s)? Surely reverse engineering may be helpful to understand the motives, pathways chosen?
As far as Micah's questions about his five examples
1. I am not sure that all design requires tinkering though I would agree that tinkering implies design.
2. Is tinkering less efficient than a single instant design? How does one establish this?
3. Tinkering implies improvements over the past implementation. Or does tinkering implies changes? Need these changes always be improvements?
4. I am not sure about this one
5. that's an interesting thought to pursue further.
[ 30. November 2003, 17:30: Message edited by: Pim van Meurs ]
IP: Logged
|
|
Evan
Member
Member # 164
|
posted 30. November 2003 19:31
Here are some more direct responses to Micah’s statements:
Micah: 1. All design implies tinkering
Yes, I agree and responded to this in my first post.
Micah: 2. Tinkering is a less-efficient method than single instant design
I think the wording of this statement prejudges some issues. First, I don’t think I know of any nontrivial example of “single instant design.” People (who are the only “intelligent designers I know of, excluding lesser skills in other animals) can be more or less efficient at tinkering, and bring other skills to the process in varying degrees, but I don’t see tinkering as being a method that can be compared to some other mode of design, especially “single instant design.”
I would be interested to hear if Micah has some thoughts on this.
Micah: 3. Tinkering implies improvement (rather than mere change) over a previous state of design
Not necessarily. Designs have both functional and aesthetic components, and are done to achieve both internal and external goals or satisfactions. Artists go through different periods, for instance, which at times they see as growth but at other times they see as just trying something new - and sometimes new turns out worse.
Micah: 4. Tinkering implies that the designer is not significantly better at designing (discovering an optimized state) than, say, a simulated-annealing algorithm
Not necessarily, although we don’t know nearly enough about how we design (where thinking comes from) nor about the ultimate capabilities of artificial algorithms to really know much about this.
Micah: 5. Tinkering reduces the significance of cognition in design to not much more than trial and error
No, because thinking is also an internalized tinkering procedure.
Also, I don’t think it’s correct to call this tinkering aspect of design “trial-and-error,” although that may part of it at times. When a composer writes and plays the first version of his song, he then has the opportunity to hear it from the ouside-in, so to speak; he gets to experience it as an external object. From that experience he makes judgments and gets new ideas for his next version. In fact, it is a very common experience for a designer to wind up with a product dramatically different than what he first envisioned. I don’t think it’s correct to say these stages in the design process are “errors.” Rather, they are the means of creating a feedback loop whereby the designer and his design mutually effect each other as the design evolves.
Micah: To be less than cryptic, my main interest in this thread has to do with design methodologies, and whether the intelligent designer has design options that lay outside the "tinkering" model and if not, "whether tinkering implies bad design" or "whether tinkering can be viewed as a positive design virtue"
To repeat,
a) Do you think a designer has options outside the tinkering model? Do you think “single instant design” is such a model?
b) No, tinkering is neither bad nor good - it just is a fundamental part of design.
P.S. in regards to PvM’s posts in this thread, at this point I don’t see the topic being about the designed object itself, but rather about the designer and the design process. I also don’t see this thread as being about biology as far as I can tell. Micah has asked question about how intelligent designers design, and it seems like both his questions and Peter and my responses have been about human beings designing. I’m not sure how we could be talking about any thing else, as these are the only examples of designers going through the design process that we know about.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 30. November 2003 22:10
Tinkering design is a form of evolution, at least as practiced by humans. It's not Darwininan evolution in any normal sense of the term. But you have mutation (coming up with new ideas/objects) and selection (tossing the stuff that doesn't work). Tinkering lets you combine good ideas that worked in a variety of disparate areas (e.g. a car combines the good idea of levers and gears; internal combustion engine; glassmaking; electric lights; quantum mechanics (lasers in your CD player), etc. etc.), which isn't something that's normally accessible to living organisms (at least not modern multicellular eukaryotes, as far as we know).
Because of the tinkering nauture of human design, some things that we unconsciously associate with design may be aspects of a generalized evolutionary process rather than aspects of the generalized design process.
As to Micah's five implications:
(1) Most human design involves tinkering, but I don't think one can make much of this. There are many other possibilities (exhaustive search, direct prediction, etc.).
(2) I'd agree that tinkering is less efficient. With instant design, you have only and exactly what you need. Tinkering can't do any better than that, and may do worse depending on what it is that you need.
(3) I'm not sure why tinkering has to involve an improvement. One can also tinker for fun, and end up with something worse--or one can fail at successful design and end up with something worse at the end of the tinkering process.
(4) If multiple passes of modification are required, then this says that the designer isn't able to predict the outcome of a modification before making it. However, the designer may still do much better than, say, simulated annealing, as the designer may be able to rule out nearly all possible modifications as unhelpful, leaving only a handful that need to be tried.
(5) I wouldn't say that tinkering reduces cognition to the level of trial and error; cognition can be important in constructing the trials. There are many less useful ways to construct trials, so it isn't fair to be too dismissive of the value of cognition here.
Overall, I'd say that the intelligent designer does have options other than tinkering, but not if that designer is unable to predict the outcome of the tinkering. If you are a designer with limited capabilities, tinkering is a positive design virtue as it allows you to extend your capabilities. If you are a designer with unlimited knowledge and capabilities, tinkering implies bad design since you knew how to do better the first time and yet had to come back and fix things. (Unless there was some reason to have a bunch of substandard designs around, or what was optimal kept changing and there was some reason to keep redesigning things instead of making them self-reconfigurable to be optimal in all situations.)
Edited to get the thread owner's name right. Oops.
[ 07. December 2003, 01:35: Message edited by: Rex Kerr ]
IP: Logged
|
|
Stuart Harris
Member
Member # 152
|
posted 30. November 2003 22:34
Micah,
I'd add that tinkering may imply:
1. A lack of complete foreknowledge of the outcome of a design decision even though a known goal is sought,
2. or, a desire to explore and know the different ways to achieve the same goal. I think when we tinker we are not just trying to reach a goal, we are also experimenting with different ways to come to the goal. Evan said that thinking is tinkering. If so, tinkering is both an exercise in epistemology and teleology: getting to the goal and understanding how we know the various ways to get there.
3. The desire to give "telic freedom" to the designed object. By this I mean that tinkering implies imperfection which can be improved upon and moved forward by the designer or by the designed object. This could be viewed in both Darwinian or Theistic terms. The designer of the Universe sets the laws such that universe can tinker away and design itself in a Darwinian way within the restrictions and freedoms of those laws. Or, the designer gives a theistic free will to the universe. It can also be seen in engineering terms: a designer may write a self-improving program or simulation software to tinker it's way to a known goal.
IP: Logged
|
|
The Pixie
Member
Member # 548
|
posted 01. December 2003 07:31
I would just like to comment on a point by Evan
quote:
All designers do this - they make sketches, write rough drafts, make models, etc.; see what they look like and how they work; and then move the tinkering back "inside their head" until it's time to fiddle around with the external world.
All designers make sketches and rogh draughts, and this could be called tinkering, but very few of those ideas come to fruition. Maybe Ford do a lot of tinkering with the design of a car, but if you look at the cars they actually manufacture, all you will see are the successes.
If you think you can see the tinkering of an intelligent designer when you look at the diversity of life on this planet, this would imply that the sketches and rough drafts here are just trial runs for the "production" world elsewhere.
Pixie
IP: Logged
|
|
Steve Petermann
Member
Member # 884
|
posted 01. December 2003 09:59
There is another tinkering type of design style that might be more analogous to a biological designer. In this case, however, the tinkering is not done because of the epistemic limitations of the designer, but rather because the designer has *enhanced* epistemic capabilities.
A design goal need not entail a rigid set of detailed specifications. Instead the design may be intended to be a "bounded" design. This type of design is more open to unexpected novelty(beneficial or non-beneficial), having degrees of freedom within the design process.
In this design process the designer sets up some rough initial constraints for the design including the ability to evolve. Then the designer lets the self-design process proceed but tinkers to keep it within certain bounds. The designer in this case has the ability to detect trends in self-design that go too far afield of the design's telos. Tinkering, on the fly, keeps things within certain bounds but also allows for some element of freedom within those bounds. A human analogy to this might be a parent who watchfully lets a child do some exploring "on its own" but quickly steps in if the child strays out of safe boundaries.
Theologically this could be taken at least two ways. It could be viewed as an interventionist type of design process where the designer and natural law are viewed as causally independent. However, if the initial rough constraints(natural law) were not viewed as ontologically and causally independent of the designer, while the tinkering would appear interventionist, in actually it would be just a different type of intelligent activity. In other words the design agent would be intelligently applying both stability and novelty to the process.
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 01. December 2003 12:00
Stuart Harris wrote quote:
2. or, a desire to explore and know the different ways to achieve the same goal. I think when we tinker we are not just trying to reach a goal, we are also experimenting with different ways to come to the goal. Evan said that thinking is tinkering. If so, tinkering is both an exercise in epistemology and teleology: getting to the goal and understanding how we know the various ways to get there.
The conclusion in the last sentence doesn't follow from the notion that "thinking is tinkering." As I understood Evan, he was suggesting that (among other tactics) humans use mental models to try out potential changes - we run the mental model and evaluate the potential changes in the light of the outcome of the behavior of the mental model. There's no more (nor any less) "epistemology and teleology" in that process than there is in a process where the designer uses physical models (e.g., wind tunnels), mathematical models ([i[e.g.[/i], equations describing turbulence, or the actual object (e.g., a bunch of full-size physical wing forms). While the media are different, I do not see how the process - winnowing candidates on the basis of prospective performance - is epistemologically or teleologically different. All entail the same epistemological (knowledge) and teleological (goal-seeking) considerations, except insofar as playing around with actual objects can suggest occasionally design modifications that are inaccessible to internal cognitive processes because of lack of knowledge of the behavior of the objects. Highly experienced designers - those with a good deal of background in the field in which they are designing - have acquired cognitive schemata that allow them to skip steps in the process - they don't have to actually build physical models of some candidates to know that some variants simply won't work. Inexperienced designers depend more on the prototyping and physical modeling process.
Of course, the experienced designer has a potential trap: if a variant that his cognitive schemata rule out a priori on the basis of his mental modeling is actually (given omniscience and unlimited time and resources) a step on the route to a better solution, he'll never know it. Our cognitive processes have evolved (!) to be good enough but not perfect: we have evolved cognitive processes and strategies that enable us to operate effectively in the world not as optimizers but as satisficers.
RBH
[ 01. December 2003, 12:05: Message edited by: RBH ]
IP: Logged
|
|
nobody
Member
Member # 145
|
posted 07. December 2003 13:43
Micah says:
quote:
1. All design implies tinkering
I'm a little uncomfortable with that. I'd rather see it say "All human design implies tinkering."
Just my two cents worth. I don't think God needs to tinker.
IP: Logged
|
|
Evan
Member
Member # 164
|
posted 07. December 2003 14:20
Nobody writes, "Just my two cents worth. I don't think God needs to tinker."
The history of life shows lots of evidence of "tinkering" - small changes, odd adjustments, compromises, etc.
Do you think, then, that this is evidence that God is not the designer?
IP: Logged
|
|
|
Printer-friendly view of this topic