request for advice

[Antti Karttunen]

On 3/11/06, Hugo Pfoertner <all at abouthugo.de> wrote:

>
> Please don't be too rigorous ;-)
>
> I made another random try:
>
> http://www.research.att.com/~njas/sequences/A041800
>
> and then continue with "Adjacent sequences" in either direction. All to
> die? You probably know "Quod licet Iovi ...."



Here we fortunately have the keyword "cofr", which people can
use as keyword:-cofr if they want to exclude continued fractions.

Partial SOLUTION to this problem is to have a lot MORE keywords.

For example, since the new
search interface has allowed entering terms without commas
between them, I have noticed that when I give a few small terms,
I often get back many of the permutations of natural numbers
that I and others have submitted. Indeed, I can guess that
it might p*ss up a few persons who seek for sequences that
really count something. However, I still think that many
permutation sequences are valuable in their OWN CONTEXT
(including the many signature-permutations of Catalan automorphisms, i.e.
"gatomorphisms" that I have submitted).
Here one would just a need to add a keyword "perm" to them,
so that they can be filtered off from a search when not wanted.

(If Neil decides to add the "perm" keyword, I can try to find all the
permutations in the OEIS (including conjectured), for which to
add it.)

Other keywords: something like "reg2", "aut2", "bina" or "code" for
the sequences whose definition is based on the base-2
expansion of the terms (2-regular and 2-automatic sequences).
Note that e.g. Marc LeBrun includes the keyword "base" also for the decimal
representations of such sequences, although I use "base" only for the
illustrating binary expansion of the same sequence, and leave "base" out of
the (decimal representation variant) of the sequence itself. (Which seems to
be also the practice Neil has followed...)

This is because many such sequences can occur in somewhat unexpected number
theoretical contexts.
See e.g.: http://www.research.att.com/~njas/sequences/A003714
which I recently edited and removed the "base" keyword.

Keyword "prime" for sequences containing only primes.

Keyword "enum" or "count" for sequences that have been
proved to really COUNT some "naturally defined" combinatorial
objects. These are the pearls of the OEIS, only ones that many serious users
search for.

And then apply these keywords as completely as you can,
so that they are of any use.

Furthermore, the OEIS could include a small feedback-form for every sequence
shown, where one could choose options like
"I think this sequence ..."
1) should be deleted
2) doesn't match with its definition
3) needs a keyword "base"
4) needs a keyword "perm" (is a permutation but has not been marked as such)
... needs a keyword "....", etc.

15) is a duplicate
16) would reserve a keyword "nice".

... etc.

and there would be an automatic feedback-analyzing script,
that would alert Neil whenever A) a trusted correspondent/associate editor
with enough credibility has commented on some sequence,
B) enough many of us with less credibility have commented
on the same sequence.

Terveisin,

Antti Karttunen,
writing from Prague.

PS. Some of the more fluffy sequences I have submitted
should probably be deleted as well.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.seqfan.eu/pipermail/seqfan/attachments/20060312/53c29afe/attachment-0001.htm>