Thanks Dan. Yes, I like the idea of using OEIS for getting suggestions on
the sequences.

Regards,
Yuva

On Fri, Jan 15, 2010 at 12:09 PM, Dan Bron <[email protected]> wrote:

> Yuvaraj Raghuvir wrote:
> >  1) For the J pattern I have found, is there a way to use ^: or $: forms?
> >  2) Based on the math behind the Java Code, is there a different way to
> >  achieve this result? I am not able to formulate the problem so I did it
> the
> >  debugging way.
>
> If this is the kind of output you want:
> >     t4 i.2
> >  0 1 1 2 1 2 2 3 1 2 2 3 2 3 3 4
>
> and you wanted to use  ^:  then you could say something like:
>
>           t =: >: (, >:)@:]^:[ 0:
>
> or, more directly:
>
>           t =: +/"1...@#:@:i.@:(2 ^ >:)
>
> for example:
>
>           t 2
>         0 1 1 2 1 2 2 3
>            t 3
>         0 1 1 2 1 2 2 3 1 2 2 3 2 3 3 4
>            t 4
>        0 1 1 2 1 2 2 3 1 2 2 3 2 3 3 4 1 2 2 3 2 3 3 4 2 3 3 4 3 4 4 5
>
> But I'm not sure this is what you're asking for, because I haven't read the
> Java or J code yet - I just took the last line of your
> message as the desired output.   And I'm  no mathematician, but when I
> encounter an interesting integer sequence I turn to OEIS for
> insight.
>
> For example, in this case, I copied your longest sequence (quoted above),
> inserted a comma between each number, and pasted it into
> the search box at the top of
> http://www.research.att.com/~njas/sequences/  and was directed to
>  A000120:
> http://www.research.att.com/~njas/sequences/A000120
>
> That sequence is entitled "number of 1's in binary expansion of n", which I
> mechanically translated to J to produce the 2nd version,
> above.  A little further down the page, we read:
>
>        To construct the sequence, start with 0 and use the rule: If k>=0
> and a(0),a(1),...,a(2^k-1)
>        are the 2^k first terms, then the next 2^k terms are
> a(0)+1,a(1)+1,...,a(2^k-1)+1.
>        - Benoit Cloitre, Jan 30 2003
>
> which is how I derived the first (^:  based) version.  So you see I didn't
> have to read the code or understand its math (and
> refactor) to reproduce its results.    A  $:  based version would be pretty
> easy, and is left as an exercise for those want to learn
> to use OEIS.
>
> -Dan
>
> PS:  Incidentally, it took me much longer to write this message than it did
> to write the J code after reading the OEIS entry (the
> time for that was de minimus).
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to