[seqfan] Conventions on timestamps?

Peter Luschny peter.luschny at googlemail.com
Sun Feb 1 22:11:37 CET 2009

Peter> Thus it seems that it is not possible to draw conclusions
Peter> from the timestamps on OEIS.

Rainer> I don't see it that way. It is most welcome, I think, if
Rainer> sequences submitted at time T1 will become enhanced by the
Rainer> time by comments and references dating from time T2 > T1.

Absolutely! So you see it my way.

Rainer> But maybe I didn't get your idea correctly.

Correct! I made no suggestion. I made an observation.
It puzzled me to find on a page dated 2004 a reference
not older then a month. That's all. I just want to
understand the conventions used on OEIS.

It is a problem of the interface often discussed by software-developers:

http://de.wikipedia.org/wiki/Principle_Of_Least_Surprise (for Rainer ;-)

"The principle of least astonishment surprise states that the behaviour
of an interface should be that which will least surprise the human user."

Richard> References in old (almost crusted) OEIS sequences are added
Richard> to newer papers which shed new light on their algorithms. I consider
Richard> this very legal and absolutely convincing. It indicates that even
Richard> in old sciences like mathematics progress is in some way inevitable.

Mais, bien sure! No doubt! I did not say anything contradicting.

Peter> So I was surprised to see in A098777, an entry with
Peter> timestamp [Oct 04, 2004], reference given to a paper
Peter> which bears the timestamp [January 12, 2009].

Richard> The conclusion is, in this case, that the sequence was inaugurated
Richard> in 2004, and that information dated from 2009 may be found in the
Richard> reference.

There was no indication that the description of A098777 was ever
updated since 2004! The paper had the timestamp 2009, not the
reference to that paper.

Maximilian> Maybe it is not crucial if a simple reference does
Maximilian> not bear the date when it had been added ; there's not really any
Maximilian> intellectual property associated to /adding/ the reference.

Sure. This is not a problem. However, it is an inconsistency of the
interface since other newly added references do have a time stamp.

Peter> Thus it seems that it is not possible to draw conclusions
Peter> from the timestamps on OEIS.

Yes. Contrary to my naïve expectations I can not longer assume that
a reference points to an 'old' paper when the timestamp is from 2004.

The paper mentioned is authored by Roland Bacher and Philippe Flajolet
and says: The sequence "appear as A098777 in the On-line Encyclopedia
of Integer Sequences, which is brilliantly maintained by Sloane
and a gang of dedicated volunteers."

I fully agree with Bacher and Flajolet!

Peter Luschny

More information about the SeqFan mailing list