[seqfan] Re: Announcements: faster search, new Sys Admin

Russ Cox rsc at swtch.com
Sun Jun 23 23:19:35 CEST 2019


On Sun, Jun 23, 2019 at 1:08 PM Neil Sloane <njasloane at gmail.com> wrote:

> Russ Cox has speeded up the search for when the query only contains
> numerical data. Now even sequences of 0's and 1's get a lightning-fast
> answer. Superseeker is also much faster.


As part of this, I have been looking at actual searches, and I've noticed
that there is a steady stream of people searching for space-separated
number lists like:

    1 2 3 6 11 23 47 106 235
    "1 2 3 6 11 23 47 106 235"

My guess is that these searches intend to search for that exact sequence
(this being the Online Encyclopedia of Integer SEQUENCES and all),
and that despite the hints and warnings about using comma syntax
instead, many people don't realize what is going on.

As an experiment, I've changed the rules on the server to turn three
or more space-separated numbers into a sequence search, as if they'd
been comma-separated. So now these are all the same search:

    1,2,3,6,11,23,47,106,235
    1 2 3 6 11 23 47 106 235
    "1 2 3 6 11 23 47 106 235"

To search for those numbers appearing all together in a sequence entry
but not necessarily in order, they can be quoted individually, as in
["1" "2" "3" "6" "11" "23" "47" "106" "235"]. And of course if the order
is known but not where index gaps are, there are always subsequence
searches like:

    subseq:2,6,47,106
    2,3,6,__,235

The hints page https://oeis.org/hints.html is updated with the new
state of the world.

Please let me know if you think the new behavior is problematic or
can be improved.

Best,
Russ



More information about the SeqFan mailing list