Path: utzoo!attcan!uunet!lll-winken!ncis.llnl.gov!helios.ee.lbl.gov!pasteur!ucbvax!ucsd!orion.cf.uci.edu!uci-ics!venera.isi.edu!smoliar From: smoliar@vaxa.isi.edu (Stephen Smoliar) Newsgroups: comp.ai Subject: Re: Thought/Emotion/Feeling Keywords: Devils and Angels, Bugs and Features, Chaos and Order Message-ID: <7345@venera.isi.edu> Date: 23 Jan 89 15:58:13 GMT References: <1380@tank.uchicago.edu> <43583@linus.UUCP> <7301@venera.isi.edu> <43767@linus.UUCP> Sender: news@venera.isi.edu Reply-To: smoliar@vaxa.isi.edu.UUCP (Stephen Smoliar) Organization: USC-Information Sciences Institute Lines: 59 In article <43767@linus.UUCP> bwk@mbunix (Barry Kort) writes: >In article <7301@venera.isi.edu> smoliar@vaxa.isi.edu.UUCP >(Stephen Smoliar) gently chides me for being disatisfied with >with my right brain. Stephen writes: > >>In article <43583@linus.UUCP> bwk@mbunix (Barry Kort) writes: >>> >>>I, for one, would love to clean up my inefficient and erratic >>>right-hemisphere and install some decent code. >>> >>This reminds me of a remark from a letter (74) by Rilke expressing his >>reaction to the goals of psychotherapy: "If my devils are to leave me, >>I am afraid my angels will take flight as well." Granting that you HAVE >>the sort of right hemisphere you want to "clean up" (and I tend to agree >>with Minsky that one should be suspicious of such brain division), what >>makes you think cleaning it up will improve its performance? One of the >>nice things about THE SOCIETY OF MIND is the way it makes cases for the >>necessity of indirect connections among the agents, as opposed to the more >>direct links one might find in an efficient and reliable dataflow network. >>That architecture may be "inefficient and erratic;" but it's robust! How >>many pieces of efficient and robust hardware do you know? > >My suspicion that there is room for improvement comes from the >observation that most of the decent code in my left hemisphere >was stolen from von Neumann architectures. (I admit it. It was >my computer who, more than anyone, taught me how to think.) >I am looking forward to making friends with the new class of >parallel architectures. I am hoping they have some good ideas >on how to implement combinatorial logic, intuitionist logic, and >theory construction. > There is, of course, the possibility that your "friendship" may only further enhance your left hemisphere! Certainly, anything along the lines of "study" of such archiectures would seem to have more to do with the left than the right. Probably the only way in which the right hemisphere could "get involved" would be through your accumulating concrete experiences with such archiectures. On the one hand, your left hemisphere would benefit from your learning how to achieve your objectives with new equipment, while, perhaps, your right would, in some way or another, assimilate the new experiences of using that equipment. Nevertheless, I would still question whether or not those experiences would have anything to do with "cleaning things up." More likely, they would just add a bit more metaphorical spice to an already savory stew. >As to efficient and robust hardware, I nominate the now defunct >Bell System Telecommunications Network and the Northeast Power >Grid, which hasn't crashed since 1967. (Internet should be so solid.) > I don't know if I would call these systems efficient. They certainly meet demanding performance requirements, but their robustness makes their performance far from optimal. We sometimes forget the occasions when we have dialed a number and nothing happens; this is because, out of relfex, we know to simply hang up and try again. Clearly, the system has failed us; but it is robust enough that the failure is not catastrophic. Furthermore, the failure is small enough that we can live with it. Power surges are a similar example of sub-optimal performance. All we can ever expect of robust systems is that they be "good enough" (Simon's "satisficing"); similarly, all you may be able to expect of your right hemisphere is that it does what it does.