easy, hard

[Brian L. Galebach]

The help file indicates that it is hard "if the next term is not known".
However, that doesn't make sense if we know a billion terms.  I think that a
hard sequence should be one for which it becomes very difficult to calculate
values after just a few terms, (unless the values of the terms are easily
understood such as A014221 where a(n+1) = 2^a(n) ).

Brian Galebach
briang at ProbabilitySports.com

-----Original Message-----
From: djr at doctor.nl2k.ab.ca [mailto:djr at doctor.nl2k.ab.ca]On Behalf Of
Don Reble
Sent: Tuesday, April 02, 2002 5:41 AM
To: Number Sequence Mailing List
Subject: Re: easy, hard


> Is there an understanding among OEIS users about how easy is 'easy'
> and how hard is 'hard'? ... I'm ambivalent about whether that should
> include programming time as well as running time.

I doubt that there is such an understanding. Furthermore, one
can cheat to get either answer. To illustrate:

I just computed the next term of A048824. Twice. My first
program was a bit crude, and needed over an hour to get there.
After the mandatory head-thwack, I improved the algorithm:
this time it took maybe ten minutes. But now it occurs to me,
that there's a much better algorithm, both quicker to code, and
faster to run. (Here's a BASIC version:)

    10  print "A048824:"
    20  print 3, 7, 23, 347, 1439, 10780559

If only I had known that algorithm.

Anyway, one can code slowly and stupidly, to make a problem
look hard; or quickly and psychically, to make it easy.
Perhaps there's a better way to ask the question?
--
Don Reble       djr at nk.ca