Re: Next generation COBOL?

2005-11-29 11:44:55 MET

"Oliver Wong" wrote:

> ...

> Computer scientists have classified problems by how hard they are to
> solve. One such class is called "P" (for "Polynomial Time"), and problems
> in P are essentially the hardest problems that the computers we have can
> solve in a reasonable amount of time (sorting a list of numbers in in P,
> for example). There exists also a different class, usually called
> EXP-TIME, for "Exponential Time", and these are the sort of monstrous
> problems which we essentially have no hope of ever solving, even if we let
> our most powerful computers chug away at for the age of the universe. They
> are simply too complex.

One of those problems is not too far away: grammars with a
lookahead of plus infinity - that means, that a compiler needs to know the
complete program-text in order to make any sense of even the first symbol.

Some people claim that COBOL is such a language - but it is not that
bad, really! At this time, I do a subset of COBOL with pedestrian YACC
- no need for a quantum computer there!

Another example might be jokes, where the last few words do a
re-semantic on the entire text of the joke, and of course a set of some
optimization problems out of which you mentioned the traveling salesman.
The other one I did mention is the solving of an infinit number
of path-integrals in order to compute the behaviour of a simple
particle - that is computational complex beyond our wildest dreams -
and it is a simple thing for the photon itself, which *is* a quantum
computer.

Always the same message: Nature loves Quantum Computing - but not
bytes or bits or nands or nors or Intel-processors.

Herwig




.