posted January 16, 2007 01:38 PM            

This is about all I could find on the web, its not research oriented
but it might amuse:

Computer games for pets:
http://www.moggies.co.uk/html/quake.html
http://www.cyberpounce.com/

My current interest is in computer languages. I have a tame scripting
engine, written in PB. The original version could handle dbase3 files
and was used to drive various DBs over a network. That was pre-windows.
Now I use it for testing minimalist ideas, trying to remove more than
I put in. Down to about 100 words so far without, compromising usability.
Hence my interest in philosophy and now discovering Wittgenstein.

But there are two opposing Wittgensteins. Computer Languages mostly
use the 'Tractatus Logico Philosphicus' paradigm, but what about the
more elusive 'Philosophical Investingations' which depicts the
complexties of natural language, in which logic plays a lesser role,
and expressions mean different things in different contexts.

http://www.kfs.org/~jonathan/witt/tlph.html

http://www.galilean-library.org/pi1.html
but I havent found a complete set.

>TAXONOMY

I've got Charles Petzold's "Programming Windows" buried somewhere.
Unfortunately it put me right off Windows. Life is too short. But I
take your point about taxonomy. Windows/Graphical objects fit very
nicely into the family tree idea. Databases and AI less so. Software
seems to develop in a Darwinian manner,and once prototyped, has Platonic
order imposed upon it.

>INQUISITION AND RESEARCH

Statistics are used for torturing facts, as well as students.

Did you ever use Bayesian methods for analysis, or is this a relatively
new idea in research? Bayes theorem can be closely derived from Kolmogorov's
axioms.

------------------
www.pegge.net

posted January 17, 2007 06:51 PM            

or Google "David Washburn chimp video games"
Most of Washburn's studies use computer-game-type tasks
even when this is not mentioned in his titles.
I co-authored a chapter in Washburn's (edited) book,
"Primate perspectives on behaviour & cognition" (American
Psychological Association press, 2007). The sells for $95 US
at Amazon but authors can send free .PDF reprints I would think.

>statistics and probability
My favorite program in statistics is one I wrote myself
and modelled on a chimp video-game task. Even a child
can "do statistics" with it and enjoy it -- at least
until he or she discovers that it is MATH... But re Bayes:
I'll have to look at the Bayes-Kolmogorov connection;
I am not very knowledgable about either person's work.
Throughout my career I periodically saw Bayes mentioned
in some of the major journals in my fields, but I never
felt any need to use his methods, and neither did any
other researcher I knew. The leading lights in applied
statistics, especially R.A. Fisher, http://www-history.mcs.st-andrews.ac.uk/Mathematicians/Fisher.html

definitely seemed to have a low opinion of Bayesian
statistics.

Fisher was also a geneticist & very famous for his work
on the genetical theory of natural selection, while Bayes
was an Anglican cleric; but I do not believe that explains
much. At the same time, Sokal & Rohlf's classic statistics
text, "Biometrics" does not even list Bayes in its index, at
least in my edition of it, while it has high praise for Fisher,
and chapter after chapter on the the methods which he
devised or otherwise "sold" to applied scientists.
Coincidentally, Sokal co-authored another classic that might
be of interest: "Numerical taxonomy". For a very brief account
see http://www.garfield.library.upenn.edu/classics1987/A1987F272800001.pdf

It produced great excitement & controversy at first & aimed
to do away with classical taxonomy in favor of computerized
classifications of animals.

Some of my favorite pages (thus far) in the book on "What we
believe but cannot prove", which I cited in my first post, are by
Leonard Susskind, a theoretical physicist. They are about
probability. They are reproduced on: http://blogs.warwick.ac.uk/jmiles/entry/leonard_susskind_in/

Highly recommended, especially for any reader who might remember
my old post on this forum about "Climbing Mount Impossible". I'm
not certain if this is it, but close enough...
http://www.powerbasic.com/support/forums/Forum12/HTML/001543-2.html

>computer languages

What would be the prospects of easily translating any computer
language into any other one -- or, alternatively, devising a
"universal language"? Compared to a universal human everyday
language that sounds easy. But why is it so difficult?

------------------

posted January 17, 2007 07:30 PM            

I am about 5 hours ahead of you, so I'm going off to count binomials
and dream in Gaussian curves. Good Night.

posted January 18, 2007 12:24 PM            

I remember part of a lecture by the botanical taxonomist Mr F White, at Oxford, heaping
dickensian scorn on Sokal and Sneath as utter charlatans. He belonged to a previous age
where 'receiving specimens the other day' meant getting them 10 years ago. So I've always
wondered whether Sokal and Sneath deserved Mr White's vitriol or not. I would imagine
that this approach would become essential for genotyping but not as sensitive as
traditional taxonomy for phenotyping.

>COMPUTER LANGUAGES:

Here is a Non Platonic Computer language inspired by Darwinian ideas. It does not
use classes. Objects are spawned from other objects instead and may be modified dynamically.

http://www.transmuter.org/index.html

I found it on the HOPL database which lists over 8000 computer languages, past and present.
But why isnt Powerbasic, Qbasic, Gwbasic on the list? It's obviously incomplete.

http://hopl.murdoch.edu.au/

I can think of several reasons why we end up with so many languages

Empires of influence. Sun vs Microsoft
Specialisation (domain specific languages)
Optimisation for flexibility or speed.
New concepts emerging
Pure experimentation
Proprietary: copyrights and Patents.

A universal language would end up with a huge number of intrinsics but maybe it will
happen one day.

------------------
www.pegge.net

posted January 18, 2007 05:22 PM            

A minimalist programming language would be one that had only the
prerequisites for the task at hand.

Any programming language suited to one task, will likely be
adapted or extended to complete additional tasks, rather than
develop a new and separate language for each task.

The problem then is whether one language that serves multiple tasks
is inherantly better or worse than one designed for the task at
hand.

posted January 18, 2007 11:52 PM            

With regard to a minimalist approach, I have found that replacing REPEAT DO WHILE FOR CASE
with a single block construct, and also dropping ELSEIF and ELSE makes for a much cleaner
logical design with minimal loss of efficiency. I can also make a good case for simplifying
operator precedence rules to strict left-to-right evaluation.

posted January 19, 2007 12:34 AM            

Nicely done.

posted January 19, 2007 10:37 PM            

I would have loved to hear the lecture by Mr. F. White. But
I do not agree with his characterization, any more than I'd
call all proponents of artificial intelligence charlatans.

Numerical taxonomy was oversold, I think, but it is still
worth studying, especially, say, if one finds the programs
described by Richard Dawkins in his "The blind watchmaker"
interesting and instructive -- and I do. Dawkins' programs in
effect create or evolve (metaphorically speaking) almost
innumerable forms, quite varied if not sometimes beautiful and
sometimes monstrous. Sokal & Sneath did some work that was quite
analogous: for example in one study they themselves created
many different artificial forms, but then went on to assess
how well their "blind" statistical numerical taxonomy could classify
these objects &/or deduce the systematic principles on which the
objects had been constructed. They compared their methods against
the intuitive judgments of people. I found their account intriguing
enough to consider, in the 1980's, a study of "chimps as taxonomists",
using the same objects. Of course real animals rather than human-made
replicas, would be more interesting as "objects" to be classified;
but the principles on which real animals were designed are far from
fully known.

>computer languages

My favorite story on why there are so many languages is the Tower
of Babel in the Book of Genesis, so I was pleased to see a picture
of the tower in the web site that Charles mentioned. Or, to be prosaic,
a prime function of any communication system, animal or human, is
to differentiate the good guys from the bad guys-- or to "pass" as
a good guy even when you are not.

>The Transmuter language

From the description it looks great -- especially in its ability to
modify a program or create new macros if not subroutines on the fly
or from user input. I wish PB could do that, without third-party
add-ons. Or is there some secret address that one can CALL (as in
Radio Shack's Model 100 version of BASIC, circa 1980), which can
PARSE and then EVALUATE any phrase you enter? I doubt it.

The smallest yet most generic program I've written was with
CA-Realizer BASIC, which was even better than Model 100 BASIC in
the above regard. As I recall, the program amounted to:

10 LINE INPUT A$: EVALUATE A$: GOTO 10

One could do almost anything with it, including graphics,
spreadsheets, text processing, & math. Amazing. The only
problem was that the user had to learn Realizer BASIC to do anything
at all. Nobody I knew was wild about that, and so I couldn't
even give the program away, free.

Of course this is not "minimalism" in the sense that Charles &
others use that term. Or is it? My program required the run time system
of a BASIC plus the the Windows if not also DOS operating system,
plus whatever additional utilities I needed to overcome their quirks.
Its one line was like the tip of an iceberg in a fathomless ocean
that teems not just with bugs but all manner of wildlife.

>"The problem ... is whether one language that serves multiple tasks
is inherantly better or worse than one designed for the task at
hand" -- Donald Darden

Most modern linguists would refuse to judge one natural human
language as intrinsically superior to any other one. Maybe there's
a useful lesson there, somewhere. (On the other hand, Mark Twain
rated the French as the most intelligent of races because even their
little children can speak good French, which he found impossible. One
Twain's funniest pieces was his essay on "The awful German language";
and one native German-speaking professional translator told me it
was also one of the most perceptive essays she had read on the
problems and frustrations of translation.)

posted January 20, 2007 09:02 AM            

Of course I left out swimming underwater, running 100 meters in
less than 30 seconds, climbing a tree, & doing the Tarzan call
and whatever else it takes to persuade one's Jane that it's time
to join forces and add to the operpopulation problem. As Konrad
Lorenz said, humans are specialists in curiosity & non-specialization.
I believe (but cannot prove) that artificial intelligence and
artificial languages will never equal natural ones because as
soon as they beat humans at their own games (cf chess, & the
Turing test), humans will probably change the rules, invent
a new game, or simply cheat.

posted January 20, 2007 11:38 AM            

quote:

---

011:005 And the LORD came down to see the city and the tower, which
the children of men builded.

011:006 And the LORD said, Behold, the people is one, and they have
all one language; and this they begin to do: and now nothing
will be restrained from them, which they have imagined to do.

011:007 Go to, let us go down, and there confound their language, that
they may not understand one another's speech.

011:008 So the LORD scattered them abroad from thence upon the face of
all the earth: and they left off to build the city.

011:009 Therefore is the name of it called Babel; because the LORD did
there confound the language of all the earth: and from thence
did the LORD scatter them abroad upon the face of all the
earth.

---

The Awful German Language by Mark Twain
http://www.crossmyt.com/hc/linghebr/awfgrmlg.html

Absolutely hilarious!

Inventing Computer languages:

For the most part I think new computer languages are invented, from
necessity and sometimes desperation rather than ego. My venture into
languages was an economic one to start with. I wanted to save
my client company from having to spend a fortune on Dbase3 licences.

The EVAL command, cant be used in a statically compiled language.
With my experimental scripting language, there are two options:
the Exec, which does single line interpretation from a string and
then there's the ability to write a script on the fly, then load and
execute it without a break. In fact you can write the script of a
script of a script ad infinitum. This is useful because it enables
one to customise a generic program into something very specific, that
will execute efficiently.

Natural Language.

Unless you want a robot to be your personal friend and spiritual advisor,
there is probably no need. But its often useful to feed short sentences into
Google. The search results are usuallly better than searching on keywords alone.
And there are chatterbots on the web, which although they are not particularly
intelligent, sound quite convincing: I think they could carry a conversation
in the pub quite well, once the listener has had a glass or two of something
strong.
http://www.alicebot.org/

These bots are based on AIML (Artificial Intelligence Markup Language)
You can download their brains and see exactly what is inside them.
A large repertoir of standard responses. Well maybe, thats how some people
think when thay are not thinking, so to speak.

I use a TTS voice which reads the clip board. Its not the most advanced
available but I have mine turned up to full gabble, with an English accented
voice 'Gordon' who is quite clear though obviously a heavy smoker. He is
excellent at proof reading too. The package is called Digit and offers a
trial demo.
http://www.digalo.com

I did some experimenting with Dragon speech recognition software to see
if a two way conversation might be feasable. The speech recognition ability
is very impressive but if you are already used to it, typing seems to be the easier
way to get text into a PC.

PS.
Now I remember going to Mr White after his lecture, and asking whether
numerical taxonomy would be useful in the hands of 'competent' taxonmists.
My question was met with incredulous silence...

posted January 21, 2007 06:07 PM            

http://idrewthis.org/d/20070116.html

posted January 24, 2007 01:34 PM            

Re Mr. F. White: I have an analogous "taxonomy" story about Ernst
Mayr. I once had the great good fortune to bump into him (literally)
at a conference and we got to chatting. I told him I was interested
in the taxonomy of behavior. His response to that was, There
is no such thing. He had nothing against the use of (some) behaviors
as field markers by which one might identify species; his point
was simply that Darwinian classification is of organisms, and
by descent. What I was talking about was, for him, more akin
to the cataloguing of library books.

posted February 08, 2007 06:43 AM            

Melvyn Bragg 'In Our Time' BBC Radio 4

http://www.bbc.co.uk/radio4/history/inourtime/inourtime_20070208.shtml

More in the philosophy archive

http://www.bbc.co.uk/radio4/history/inourtime/inourtime_philosophy.shtml

------------------
www.pegge.net