[seqfan] Re: how is A177018 defined?

Marc LeBrun mlb at well.com
Fri Jun 18 18:15:14 CEST 2010


The description does seem needlessly opaque.  Some comments/suggestions:

  *  A description of the form "Numbers n such that <predicate on n>."
implies that the sequence is a set (aka "list").  As such it shouldn't
contain duplicated elements.  (Also be increasing, have offset=1, etc).
For instance, a good example is A067076.

  *  In descriptions A-numbers are basically opaque "pointers" and should be
avoided--unless they *significantly* improve clarity--otherwise use text,
preferably something very similar to the description found at that A-number.

Plausible exceptions to this might be when the referenced sequence has a
complicated description, when the referring sequence is very strongly
related or derivative in some way (eg "Record values of...") or when the
resulting text would be too confusing (eg nested conditionals resulting in
complicated dependent clauses, ambiguous pronoun references, etc).

(In fact, in seqfan eMail, I wish people would include at least some brief
explanatory text when introducing a new A-number in a thread.  Following
links is a distraction that I suspect often results in ignored queries.)

  *  Conversely, it's reasonable to refer to A-numbers in formulas etc.  (Of
course every such A-number should then appear in the "see also" list too).

  *  Also, an A-number refers to an entire sequence.  Since the OEIS
contains both sequences that are sets and sequences that are maps notations
such as A067076+A001477 might potentially mean many things, including:
  the union of their members
  the sequence formed by summing corresponding sequence elements
  the set of distinct sums of corresponding elements
  the set of distinct pairwise sums
  the table of every pairwise sum

  *  In particular, the reference to A001477 seems especially obfuscatory,
and the presence of n in the same expression renders it unintelligible.

  *  If I understand it, the code seems to be generating a pair of
non-decreasing sequences, which I'll notate a(n) and c(n), such that
  p(n) := a(n) + c(n) + A067076[c(n)]
is always an odd prime.

The algorithm travels from (0,0) in the (a,c) grid: when p(n) is an odd
prime it outputs the a coordinate and steps in the c direction, otherwise it
steps in the a direction.

While I don't get the motivation for this, I observe that the condition
implicitly references the sequence

  q(n) := p(n) - a(n) = n + A067076[n]

(Yow, is q = A080370?!  Sorry, I'm working by hand.  Anyway, if it isn't
already we should put it in the OEIS)

Given q(n) I suspect it might be possible to compose a reasonably
comprehensible description of a(n), but I'm out of steam for now...
 


>="Maximilian Hasler" <maximilian.hasler at gmail.com>

> The sequence is created by the following code, which I tried to make
> as clear as possible:
> 
> n=0; c=0;for(i=1,99, t = n+c+A067076(c); if( t%2 == 1 && isprime(t),
> print1(n", "); c++,/*else*/ n++ ))
> 
> 3, 3, 3, 4, 4, 5, 5, 6, 8, 8, 10, 11, 11, 12, 14, 16, 16, 18, 19, 19,
> 21, 22, 24, 27, 28, 28, 29, 29, 30, 36, 37, 39, 39, 43, 43, 45, 47,
> 48, 50, 52, 52, 56, 56,
> 
> Maximilian
> 
> On Fri, Jun 18, 2010 at 8:50 AM, Richard Mathar
> <mathar at strw.leidenuniv.nl> wrote:
>> 
>> I have problems understanding the (apparently simple)
>> http://oeis.org/classic/A177018
>> "Numbers n such that n+A067076+A001477 is odd prime"
>> Why are values occurring more than once?
>> Which indices (or index combinations) are missing in the definition?
>> 
>> If we use "Numbers n such that n+A067076(n)+A001477(n) is prime",
>> which is of course the same as
>> "Numbers n such that 2n+A067076(n) is prime"
>> we get the very different
>> 1,2,6,9,12,23,28,29,32,36,39,43,49,52,59,64,69,73,87,99,..
>> 
>> Lead by the examples I considered "Smallest k such that
>> k+A067076(n+1)+A001477(n) is an odd prime"
>> but this doesn't match either.
>> 
>> RJM
>> 
>> 
>> _______________________________________________
>> 
>> Seqfan Mailing list - http://list.seqfan.eu/
>> 
> 
> 
> _______________________________________________
> 
> Seqfan Mailing list - http://list.seqfan.eu/






More information about the SeqFan mailing list