Re: Epistemology 201: The Science of Science

From: aeo6 (aeo6_at_cornell.edu)
Date: 03/01/05 

Date: Tue, 1 Mar 2005 10:10:13 -0500

Lester Zick said:
> On Fri, 25 Feb 2005 20:15:04 -0600, Albert <albertwagner@cox.net> in
> comp.ai.philosophy wrote:
>
> >Wolf Kirchmeir wrote:
> >> Albert wrote:
> >>
> >>> Wolf Kirchmeir wrote:
> >>>
> >>>> Albert wrote:
> >>>>
> >>>>> Wolf Kirchmeir wrote:
> >>>> [...]
> >>>>
> >>>>>> Statistical averages are not what's meant by "good approximations"
> >>>>>> of unsolvable problems.
> >>>>>
> >>>>> Only in math. In common usage statistical averages are often
> >>>>> referred to as good approximations of the world.
> >>>>
> >>>> Really? By whom? I can't think of an example. Not even our
> >>>> newspapers, which are not noted for the brilliance of their analysis,
> >>>> make that mistake.
> >>>
> >>> It is not a mistake, except under mathematic's narrow definition. If
> >>> you want examples then read some economics. Economists all fancy
> >>> themselves mathematicians.
> >>>
> >>>> Maybe you mean survey results? Well, a survey is a measurement, so of
> >>>> course its results are an approximation. And survey results aren't
> >>>> averages, they are merely -- statistics. (How good a measuremnt is an
> >>>> entirely different question.)
> >>>
> >>> So averaging is a disallowed operation in statistics?
> >>
> >> No, it's just an operation on numbers, is all. It used in many areas,
> >> not just statistics.
> >
> >That was not your implication in the statement I replied to.
> >
> >> the average score on a test can be computed, for
> >> example, but whether it's a meaningful number is another question. All
> >> course-mark algorithms use averaging of one kind or another, which
> >> raises the question just how useful a course-mark really is.
> >>
> >> Besides, there are many different kinds of averges, as you know, I
> >> think.
> >
> >Yes.
> >
> >> To that the average family has 1.2 children is arithmetically
> >> correct, but doesn't tell you the size of the typical family. For that
> >> you need the mode, which tells you what the most frequent number in a
> >> data set is. It's 2 children, IIRC, ie, when you survey a large enough
> >> sample (around 500-100) families, the most common number of children is
> >> 2. The median is slightly higher IIRC, because there are enough families
> >> of 3 or more children to raise that number. BTW, if you leave out
> >> childless families, the average is close to the mode.
> >>
> >> People aren't dumb enough to believe that arithmetic averages make sense
> >> in every situation.
> >
> >LOL. You should pay more attention to politics and advertising copy.
> >
> >> But because they haven't been taught about the other
> >> averages, they regard statistics with a mixture of scorn and awe.
> >
> >I find 'scorn' for applied statistics more prevalent among those
> >who know statistics well but have no agenda. No other discipline
> >is so apt to use its knowledge to lie.
> >
> >> If you want a good intro/review of mathematics as it applies in the real
> >> world, find John Paulos's "Innumeracy." His "A Mathematician Reads the
> >> Newspaper" is also worth reading. Besides clarity, his books exhibit a
> >> good deal of wit.
> >
> >Thank you.
> >
> >> Anyhow, don't blame mathematics and mathematicians for stupid
> >> applications of mathematics.
> >
> >I don't think I did. I only said: "In common usage statistical
> >averages are often referred to as good approximations of the
> >world." Such differences in definitions are useful to know so
> >that one can communicate effectively. When talking to
> >mathematicians you use the mathematical definition. When talking
> >to non-mathematicians you use the common definition. Or at least
> >be aware of the differences and explain your context up front.
> >This inability or unwillingness on the part of Mathematicians to
> >be aware of common definitions and to compensate when talking to
> >non-mathematicians is responsible for 95% of the arguments on
> >this thread.
> >
> >After 35 years in programming, in constant contact with end
> >users, I am very much aware of the effects of jargon and private
> >definitions on outsiders, and of the need to explain things in
> >terms understandable to outsiders.
> >
> >This is purely anecdotal, but in my whole career working with
> >college graduates having a variety of majors, I noted that
> >mathematicians were singularly the most ineffective programmers,
> >due primarily to their inability or unwillingness to step out of
> >their bubble and actually grasp just what the problem to be
> >solved really was.
>
> Amen! God save us from academic programmers. The worst recommendation
> in professional programming was a computer science degree.
>
> Regards - Lester
>
Whoa, hold yer horses!!! A computer science degree is a little different
from a pure math degree, and either way, having been in the business and
programmed for 20 years, and seen a variety of other people's code, I
have to say that there are a whole lot of people out there banging out
code that don't know what the hell they're doing. Granted, there are
some smart young folks that pick up ideas well enough on their own to
wing it and create good databases and applications, but to solve serious
problems in computer programming can require some anaylsis that isn't
apparent unless you're taught about it. It's kind of like saying that
the worst engineers have studied physics.

For my part, I am glad to have learned what I learned srudying computer
science. 

-- 
Smiles,
Tony

Relevant Pages

  • Re: Epistemology 201: The Science of Science
    ... Lester Zick said: ... >>mathematicians you use the mathematical definition. ... God save us from academic programmers. ... > in professional programming was a computer science degree. ...
    (sci.cognitive)
  • Re: Epistemology 201: The Science of Science
    ... Lester Zick said: ... >>mathematicians you use the mathematical definition. ... God save us from academic programmers. ... > in professional programming was a computer science degree. ...
    (sci.math)
  • Re: Why (or not) use single letter variable names?
    ... On 10 Dec 2004, AngleWyrm wrote: ... >Programmers and Mathematicians have different objectives and uses for their ... with extraction of meaning in the documentation. ...
    (sci.math)
  • Re: Why (or not) use single letter variable names?
    ... "Dan Piponi" wrote in message ... Programmers and Mathematicians have different objectives and uses for their ... the documentation, wherein the variables are defined, elaborated on, proven. ...
    (comp.programming)
  • Re: Why (or not) use single letter variable names?
    ... "Dan Piponi" wrote in message ... Programmers and Mathematicians have different objectives and uses for their ... the documentation, wherein the variables are defined, elaborated on, proven. ...
    (sci.math)