Question about missing terms

[Marc LeBrun]

 >=Mark Hudson
 > I apologise if this question has been asked/answered many times in this 
list, but what is the correct protocol for submitting a sequence when one 
knows many terms, say from a(1) to a(N), but does not know the values for 
some m: 1<=m<=N?

The best thing is to use some "flag" value, such as 0 or -1, outside the 
range of a(n).

This is also often done even for complete sequences where some of the terms 
are non-integral for other reasons (eg fractional, infinite, undefined, etc).

Sometimes to do this you may have to resort to submitting a transformed 
sequence f(a(n)) that somehow "encodes" a(n)-or-flag.

An entire auxiliary "flag" sequence could also work, but then when more 
terms become known someone would have to update two sequences instead of 
just one.

Or you could just add a comment like "a(n) not currently known for n=6, 9 
and n>50" or whatever.


 > Could I submit a sequence with "?" marking the missing terms?

Don't, because it will screw up facilities such as the transforms (ie 
superseeker) that expect all the values to be numeric.