old sequences

Erich Friedman efriedma at stetson.edu
Mon Jan 24 18:57:00 CET 2000


Michael Somos writes:

  >One distinction mentioned a few times is between "core" and
  >non-core sequences. There are only a few thousand of these and it
  >may be worth it to distinquish them more clearly in the operation
  >of EIS. That way, the impact of the tens of thousands of other
  >sequences would not dilute the core.

This seems to suggest the following compromise: In any lookup of a
sequence, print any "core" sequences first.  This may make John Conway
happy (or maybe I should say "happier") to get the important sequences
first.  The prioritizing could be expanded, though this much detail is
probably not worth the extra effort.

Another possibility is to keep track of how many times a sequence is
accessed by the system, and use this as a priority ranking.  The Fibonacci
sequence will certainly rise to the top among sequences that start 1 1 2 3
5 8.  this "priority" could be of some use to those using the sequence for
research, as it might indicate the most likely fit to a given sequence.

Michael Somos also writes:

  >A resived print version of EIS may include only the designated core.

I would not favor this, as some are attracted to recreational sequences
that wouldn't make it into the core.  Derived sequences like Fib(n) mod 9
could certainly be left out of any further printing.

erich friedman
erich.friedman at stetson.edu

More information about the SeqFan mailing list