You are not logged in.

 Discussion: Roundtable Topic: Help: generating algorithm

 Post a new topic to the Roundtable Discussion discussion

 Subject: RE: Help: generating algorithm Author: Mathman Date: Oct 24 2004
At the outset, I'm not an expert, and might easily miss the point here entirely.
So take this for what it's worth please; at least an effort to help.

I can see that a random number generator can not be perfectly random, which
implies some sort of pattern, however complex.  I can also see that a function
[pattern, algorithm] of some sort can generate a few numbers.  However, a
one-to many relationship does not determine a many-to-one.  That is, in
simple terms [that I can understand] a parabolic function can generate three
points, [and more].  However, it requires four to determine a particular
parabola.  I don't have any idea of the complexity of an algorithm to generate
pseudo-random numbers, but I doubt it can be readily determined from just four
generated values.

That is; if a pattern evolves from four numbers, and equally valid pattern quite
different could also contain the same numbers, unless there are other
restrictions.

David.