[seqfan] Re: Inconsistencies in OEIS

Charles Greathouse charles.greathouse at case.edu
Thu Sep 17 17:29:25 CEST 2015


> If there are any more combinations of keywords that should trigger warnings
I'd like to know. I currently have:
>
> - "full" without "fini" --> warning (never occurs).
> - "tabl" and "tabf" together --> warning (occurred 1 time, fixed now).

Let's see. Keywords "recycled", "allocated", and "allocating" should never
appear with other keywords. Really keyword:dead should also appear alone,
but in practice there are cases where it does not (I count 54).

kw:nice should never appear with kw:less.

kw:easy should never appear with kw:hard.

kw:nonn should never appear with kw:sign.

The keywords done, dupe, huge, and part have been deprecated and should not
be used in any sequences.

Charles Greathouse
Analyst/Programmer
Case Western Reserve University

On Thu, Sep 17, 2015 at 2:01 AM, Sidney Cadot <sidney at jigsaw.nl> wrote:

> Hi,
>
> > I fixed 21 sequences that had such an error. But you might check that
> > I caught them all.
>
> It seems so - and from what I can tell you also took care of all
> entries with superfluous commas, so that should care of issues (P11)
> and (P13) on the list.
>
> It will take ~ 1.5 weeks for my local copy of the database to refresh
> the entire database, which will allow the parser to check the fixes
> automatically; this is because I get all data via HTTP and I throttle
> access to under a thousand requests per hour. In fact while starting
> this project I went beyond that, which may show up in the server usage
> statistics and I hope this hasn't cause performance issues -- if so I
> apologize.
>
> If there is a more direct way to access the data that doesn't burden
> the server as much, I'd be happy to implement it. Also, if there is
> interest to automate the process of detecting issues in the database,
> I will be happy to hand over my scripts once they have stabilized.
>
> > There was one that had both tabl and tabf keywords.
> > I don't know if you considered that an error - it is
>
> I added that condition as a check and re-ran, A212013 seems to have
> been the only entry that had this particular combination.
>
> If there are any more combinations of keywords that should trigger
> warnings I'd like to know. I currently have:
>
> - "full" without "fini" --> warning (never occurs).
> - "tabl" and "tabf" together --> warning (occurred 1 time, fixed now).
>
> Kind regards
>  Sidney
>
> _______________________________________________
>
> Seqfan Mailing list - http://list.seqfan.eu/
>


More information about the SeqFan mailing list