[seqfan] Re: Contributing sequences with holes and infinities

[Frank Adams-Watters]

In extreme cases, where not enough initial terms are known, one may sometimes enter a 
sequence with a comment like "All values for powers of two are conjectured" (or whatever 
the condition is). This should be an absolute last resort, however; we want to avoid having 
conjectured values in the data section if at all possible. In particular, if there is anyhow some 
question whether the sequence merits inclusion, and it would also have to have conjectured 
values like this, it is probably better to simply reject it.

Franklin T. Adams-Watters


-----Original Message-----
From: Christian Stump <christian.stump at gmail.com>

Hi,
> Regarding 2: As Felix said, you should use a placeholder, typically -1 in a
> case of "the number of". See e.g. A215418.

Okay, this seems to be a reasonable solution (as otherwise, the
sequence is just not an integer sequence).
> Regarding 1: the procedure would be to have the Data include the entries
> before the hole, i.e. a(0) to a(3) in your example, and state a(8) and a(24)
> in a Comment.
This seems wired, though -- what if I know the sequence everywhere
except for powers of 2 then? Is this saying that well-defined
sequences for which the terms are both know and unknown for infinitely
many entries not suited?There is no problem in searching for partial sequences, and to me it
seems better to be able to recognize a partially known sequence than
not to find it.

> In this case (where 0 cannot occur) I'd put 0 for infinity.
> Sometimes -1 is a (somewhat fugly) way out if known terms
> are non-negative.> It's often a judgement call: if looking at it makes you unwell.
> then it may be the better option to not submit.
I think it is best to have a well-chosen general convention rather
than an ad hoc solution. This is certainly better to maintain integrity
of the sequences.
Cheers, Christian

--Seqfan Mailing list - http://list.seqfan.eu/