Re: Algorithmic Revolution?

From: <vznuri.domain.name.hidden>
Date: Tue, 19 Nov 2002 22:22:52 -0700

hi all. re the term "algorithmic revolution" here are a few
more ideas along this thread Id like to point out.

TCM wrote
>My belief is that basic mathematics is much more important than
>computer use, in terms of understanding the cosmos and the nature of
>reality.

ok, fair disclosure, I have a BS software engr, writing code
since age ~10, and it affects my worldview bigtime. or, one could
say, I really know how to pick a winning horse, haha.. seriously,
I recognized & planned my life around the "algorithmic revolution"
from a young age.

at an early point I realized that software is like
"animated mathematics".

this is a very,very deep & cosmic way of looking at algorithmics. it
captures some of the revolutionary flavor. we can suggest that
mathematics has previously attempted to grasp the concept of
change, via calculus, differential eqns etc.

but something is fundamentally new about simulation. it captures
worlds that cannot be expressed via mathematical generalities. there
are no equations we can write down that describe the outcome of,
say, a climate simulation-- its all locally defined & then globally
simulated & the outcome is "emergent". & what are the differential
equations that describe the game of life??

imho algorithmics captures the extraordinary, currently very poorly
understood property of "emergence". just as
in the game of life there are thousands of glider types, none of
which one would expect/anticipate from the simple rules.

we can argue that algorithmics is a fundamentally new way to
look at mathematics. and one could argue, all mathematics up until
now has been transformed. at this point, it seems much more correct
to classify mathematics as a subbranch of algorithmics than vice
versa. I believe much mathematics of the future will be taught
from the "algorithmic point of view" instead.

imho, the invention & harnessing of the algorithm is roughly as significant
in human intellectual development as pythagoras's original
realization about how mathematics modelled nature. its easily on that
order of magnitude as far as a milestone in human thought, possibly
surpassing it. it seems to me, fundamentally, algorithmics entails
and surpasses mathematics as a new simultaneously conceptual and physical
tool for analyzing the universe
and its variegated phenomena.

so think. we've basically got several millenia of mathematical
thought, dating all the way back to the babylonians (who played
with perfect triangles, fractions etc), and quite well
developed in greece 2000 years ago. reaching heights of sophistication
with calculus, or the abstraction in the 20th century.
Im saying to some degree, all that
is childs play compared to the new universe of algorithmics.


re: TCMs questions about some of my points.


1st, I believe that we will eventually get the math for a TOE
that matches accelerator/particle physics
so perfectly that it will be considered
redundant or wasteful to do the expensive supercollider experiments, because
the accelerators will never find anything that does not match the
comprehensive theory.

that is, after all, one of the big
reasons to look for a TOE. but I agree, until that point,
physicists are not going to give up the "big science"..

a crazy thought? perhaps. but lets look at atomic weaponry testing--
thats essentially whats happened. the US has been simulating atomic
weapons testing for many years now with powerful supercomputers. and
obviously the results are considered ***extremely*** accurate. it
can indeed be done on some level.


2nd-- alas, I wish I could cite a reference. but software is
used extremely heavily in particle physics experiments to
automatically analyze particles and classify them & find
anomalous events. its basically
AI-like software, extremely sophisticated. it can look at
very complicated particle tracks & collisions and name
all the particle tracks based on analyzing the "big picture".

this used to be done by humans & by hand, and (as I understand it)
the discovery of many
particles from the last decade or around that range could
not have been done with this highly sophisticated sorting
software that can run through millions of events very quickly.

so there is a hidden story behind massive particle accelerators.
the software infrastructure for them is all invisible and
mostly unknown to the public, but its a vast edifice at
the core of the analysis, and has gone through revolutionary
changes in a short amount of time, mirroring the algorithmic
revolution elsewhere.

how much is that software worth?? I cant really estimate, but
I wouldnt be surprised if a significant percent of supercollider
budgets was spent on developing it.

if anyone knows references on software used in particle physics
analysis, I would really like to know myself.

a nice reference on the culture behind accelerators is
"beamtimes & lifetimes" by sharon traweek.
Received on Wed Nov 20 2002 - 00:35:36 PST