Dear All,
As I get this list in digest form, I'll respond to a couple of points in the same post
> Date: Thu, 26 Jan 2006 09:32:40 -0000
> From: "Rust, Chris" <[log in to unmask]>
> Subject: Re: Design Automaton
> I would say that if we have an artefact that can identify and solve =
> wicked problems then it is designing
Yes, 'solve' certainly; I'm very tempted to agree, and I'm getting less concerned that the 'artefact' isn't conscious.
Initially I looked at the 'strong' and 'weak' camps in AI and drew a parallel with this 'artificial designer', a parallel with which I am now less happy. The 'strong AI' camp seem to be saying that an appropriately programmed computer really is a 'mind' and can be attributed with a range of cognitive states, whereas the 'weak' camp seems to take the view that the system is only mimicking a 'mind' but that we might gain insights into though processes by studying such systems.
My parallel 'weak' view of 'artificial designers' was that the study of them might offer some insights to designing, but that they were not themselves truly designing. I am now inclined to believe that, if an artificial system is solving design problems (in all their 'wickedness'), then they *are* actually designing, not mimicking designing. (I have not yet checked this out this with my inclination to reject the 'strong' view of AI, but I am not overly worried in this regard, there probably isn't a conflict.)
I am still thinking about the implications of 'identify' in your statement above. I wouldn't have thought to put it in.
> If an artefact was designing I'm not sure I'd call it an "automaton" =
> since that implies a routinised function:
Fine with that. I only used 'automaton' because it was the word used in the post that stirred my thoughts on this originally. Anyway, I'd much happier call it an 'artificial designer' or similar.
> Date: Thu, 26 Jan 2006 09:37:49 -0800
> From: Monica E Cardella <[log in to unmask]>
> Subject: Re: Design Automaton
> by suggesting that sometimes it is not the problems
> themselves that vary, but instead the way that the designer
> treats the problem. Sometimes an expert designer will take a
> "simple" problem and treat it as a more complicated problem
> or a complex problem. So a novice desginer might be solving a
> simple problem and an expert designer might be solving a
> complicated problem, while both are solving the same problem.
I disagree with this. If the expert designer treats the problem in such a way that it changes the nature of the problem, then it must be a different problem. If they treat a problem as complex when it could be seen as simple, it must be that they have brought additional issues into the problem domain. They have re-drawn the boundary of the problem they are working on, and hence it is a different problem. For example, the novice asked to design a simple beam bridge may do just this, whereas the expert may spot that some other form of bridge may be lighter and stiffer, possibly cheaper, and maybe more aesthetically pleasing. In doing so, the designer gives the client a better bridge, but has changed the boundary original problem significantly.
This also doesn't preclude two other possibilities, that the expert designer may bring a wider range of processes to the design activity, and may, through experience, be able to see (and hence search) a wider solution space. Neither of these appear, of necessity, to change the problem. (Though nothing prevents the expert doing all three thing simultaneously.)
I am not sure that there is anything inherent in this that could not be embodied in an artificial designer. (Though confess I'm unlikely to work out how to do it anytime soon!)
Best regards,
John Shackleton
|