[OT] Re: [NCLUG] Linux can do THIS too??!!!
Marcio Luis Teixeira
marciot at holly.colostate.edu
Sun Oct 27 05:39:34 MST 2002
On Sunday 27 October 2002 11:29 am, Evelyn Mitchell wrote:
> One of the things I learned studying Academic philsophy is that any
> Platonic Ideal we believe we're (moving | progressing | evolving ) towards
> is just a projection of our beliefs.
I believe that one of Pirsig's point was that Quality has absolutely nothing
to do with a Platonic ideal. I don't think he would have been sympathetic to
Plato, in so far as Plato defines the ideal (for instance, Plato would have
said that "the set of all points equidistant from a central point" is a
Platonic ideal for a circle).
The way I see it (and this is my own interpretation, not Pirsig's) quality is
some sort of gradient over some unknown "fitness" function. We cannot see the
global maximum of this function, so instead we follow the gradient. Those of
you who know AI would call this the "hill-climbing" algorithm. There's no
guarantee we will reach the global maximum, and we might as well get stuck in
a local maximum. Anyhow, I think that Plato tends to talk only of *the*
global maximum (the global maximum) while Pirsig seems to be talking about
the gradient. I think this is why he calls it "dynamic Quality" -- the
gradient changes as you move around on the plane of the function. I don't
think it's much of a leap of faith to believe that "hill-climbing" is
employed by evolution.
> I wonder. There was a thread on Dave Farber's IP list this week about the
> lack of historical perspective of most contemporary academic Computer
> Scientists. They'd much rather reinvent a solution (often badly), than to
> do research into published solutions from the past. In general, I've found
> computer scientists to be quite intellectually lazy when it comes to
> knowing the past of their own field.
Yes. As a computer scientist, I admit that I hate reading philosophy or past
works from late computer scientists. The reason is that most of the material
was written long ago and has an incredible amount historical context that
gets in the way of the reading. Descartes, for example, has some really good
ideas but I can't stand reading it because he lapses off into religion,
probably because 1) everyone was religious back then 2) he didn't want to
have the church to come after him. The problem I have with philosophy is that
in order to make sense of it you have to spend a lot of time understanding
the historical context. If you find that stuff facinating, you'll enjoy it
and have a blast with it; if you don't, you'll suffocate and gasp for air.
The same is true when you read literature from computer scientists from the
1970s. In computers, things become obsolete much more quickly. When they
write about how probable it is that computer will have tens of thousands of
bits of memory by the end of the millenium, I find it so tiresome to read
that whatever relevant points they may make gets lost in the noise.
So, I would read philosophy, but only if it was modernized and written by
contemporary authors. I wouldn't mind reading from computer scientists
either, so long as they at least had at used a computer that didn't have a
punch card reader. I wouldn't mind reading Shakephere either, if someone
would please translate it into modern english and remove all the obscure
references to Elizabethan times.
Yes, I know many of you will criticise this viewpoint as being backwards and
uncultured, but that's the way I see it. I just don't have the time or
insterest in learning the cultural context that I need to make sense of
historical material. I grant that Pirsig may be a bad philophopher from an
academic standpoint, but at least his book is timely and I'm not likely to
ask myself whether his views would have been any different had he lived in
this decade. Ten years from now, I'll probably find his book irrelevant and
will hope that someone else has written something with a modern context, even
if the underlying ideas are exactly the same.
Marcio Luis Teixeira
More information about the NCLUG