Science Forums 2

[Boof-head]

I've been an IS 'man' for quite a while. I didn't realise this was the case, at least not in a definitive sense, until I actually went and studied it; though I had previously been engaged in a bit of home-science - that art of building electronic circuits, digital logic, breadboarding - I was a member of a class of geeks who join "Radio clubs" and so on.

So anyway, after some years of learning the art of programming; how a modern computer is structured; various kinds of language; databases and storage schemes; networking and communications, I got a fairly general kind of 'computer education', and I've picked up some more languages and extended my grasp of computing into the functional, OO, and list paradigms. It has become apparent that, given a digital computer built out of Boolean logic gates, an arbitrarily complex virtual system (machine) can be constructed 'on' it because the underlying logical structure is general enough and sufficiently universal, that any such a reasonable computing system might be 'built' using it.

This is congruent with the development of increasingly complex and arbitrary physical theories: start with a fundamental, irreducible universal logic (say, the logic of voltages and currents) and construct virtual 'machines', or logical devices (theories) that conform to this underlying structure - a machine algebra. Modeling a physical theory on such a computer, is using a set of fundamental algebraic functions, at the machine-level, to construct a realistic virtual set of functions, which can be used to construct other functions, and so on. The paradigm is: build a more complex logic starting with a fundamental, irreducible logic, then compose a more abstract logic from that, and on 'up'.
At any stage, the meta-logic is always reducible to a set of fundamental operations, at the machine-level, the virtual logics are in that sense, meta-languages.

Physical theories use a language which is also (meant to be) in the same domain; if you want to write a program, you use a known language - it's easier than inventing your own, you can expect that other programmers will understand it. It's like writing and playing music, where the 'tunes' are all formulas (functions) that correspond to an accepted notation.

Now, lots of people can 'bash out' a tune on a musical instrument, and they don't see a need to learn how to read (or write) the notation. Music notation is a simple kind of representative algebra, developed over centuries and simplified to the modern representation used today. Physical theories convey certain ideas, by using a similar method - notation - but the thing is, notation, musical notation say, will not convey ideas about harmony and melody, about 'why' the notation works, and moreover, why music 'works', why a given melody sounds 'right' but another sounds 'wrong', etc.

The nuances of physical theories are best supported by direct experience of the phenomena they describe, just as the nuances of music are, by listening to it - you listen to it when you play it, but anyone can hear music, just as anyone can see a laser beam on a screen, or waves in a wave-tank, etc.

So it's about sticking to accepted notation, so that 'others' will be able to read and understand (that is, play the tune). These days, some of the accepted meaning, the interpretation, is being 'abused'; we now consider certain well-understood physical processes might not be as well-understood as we think, so we have to distort meanings, we are obliged to stretch certain meanings (perhaps beyond their original intent).

For instance, we understand 'rhythm', and we understand it's connected to a linear 'flow' of regular, monotonic time. These days, the concept of 'real, physical' time is under serious review. For starters, time is not 'physical', we can't store it, or send it somewhere. We need to step beyond the fixed, invariable perception of flow of time, and perhaps consider a logic of time, generators, inverters, gates; these operators cannot be 'channels with memory' since there isn't anything to remember. But we still have to consider time (or times) as some kind of potential; then if the model holds together, although time cannot be a physical, material thing like an atom or a photon, we need some kind of irreducible structure (the machine) that allows the time logic and a time language - we should be able to build arbitrarily complex 'switching networks' in it, and design arbitrary meta-languages.

See where this is going?