Sequence sorting

[franktaw at netscape.net]

Currently, when searching the OEIS, after taking into account the 
quality of the match of the search string, the only criterion used to 
determine which sequences are "more important" is the number of other 
sequences referencing it.

I would like to suggest another criterion, one which, in many cases, 
provides a complementary measure: the length of the entry itself.  This 
could be measured in lines, characters, or whatever is convenient - it 
doesn't make much difference.  One might weight different kinds of 
lines differently, but I doubt that this would add enough value to be 
worth while.

For example, A002275 (repunits) has 674 references, while A000110 (Bell 
numbers) has only 217.  But the entry for A000110 is much longer; and, 
in my opinion, it is clearly a much more significant sequence.

Certainly there are anomalies with a particular sequence having a 
longer than normal entry; but there are similar and perhaps more severe 
anomalies with the number of references.  For example, the numerators 
and denominators of sigma_{-k}(n) all reference "A017665-A017712", so 
A017712, the denominators of sigma_{-24}(n), has a large number of 
references.

Franklin T. Adams-Watters