OEIS Keywords

[Dean Hickerson]

Mostly to Neil Sloane:

> unpr[oved] is actually a good suggestion
>
> for sequences where it has not been proved 
> that all (sometimes any) of the terms in the sequence are correct.  

Assuming that this keyword is added, then what's your preference for
sequences with uncertain terms?  When should we include terms that we're
only guessing about and use the "unpr" keyword, and when should we end a
sequence before the first uncertain term?  How confident should we be about
our guesses in order to justify including guessed terms in the sequence?

For example, I recently edited A119479, "Length of longest string of
consecutive integers having exactly n divisors".  The sequence currently
stops at a(11), because the value of a(12) isn't known.  But I'm almost
certain that a(12)=15, in which case the sequence could be continued up
to a(17).  (Probably even more terms could be guessed without much trouble.)

In this case, the value of a(12) follows from a widely-believed conjecture
about prime numbers, so I'm pretty confident of its value.  But I've edited
some sequences whose authors made guesses based on observations about what
happens for small numbers, but which were easily disproved using known
asymptotic formulas.

Dean Hickerson
dean at math.ucdavis.edu