Also, you should note that if you "start off" with extended integers
things work ok - e.g.
*/ ~. 4 p: i. 53x
32589158477190044730
At 11:05 +1100 2010/01/20, Gerry J wrote:
>Thanks very much to those who responded.
>It seems that most of the difficulty i had was traceable to x: not
>forcing the equivalent of the sqrt function before floor was applied.
>I was mainly interested in getting the correct floor but the other
>information and directions are most useful.
>I clearly have a lot of catching up to do.
>As regards the second area "the product of primes" i found that it was
>necessary to use x: at the very beginning, so that problem is overcome.
>In case my misunderstanding is helpful to others...
>
>NB. it appears that "ordinary" integers need to be processed with x: in
>some cases
>NB. or else a later calculation may not be in ep mode
>NB. "ep" == extended precision
>
>
>NB. This case doesn't tag the output as exact integers for later
>calculation, ie extended precision isn't forced in respect of earlier
>variables simply by using x: in later calculations
> pns=: ~.4 p:i.53
> pns
>2 3 5 7 11 13 17 19 23 29 31 37 41 43 47 53
>NB. i had expected if x: was used in the following statement that ep
>would be forced for all inputs
>NB. maybe a "nice to have"?
> pp=:x:*/pns
> pp
>32589158477190045696 NB. incorrect because input has 2*5 the
>result has to end in 0
> x:pns|pp
>0 0 1 0 9 4 14 16 0 9 5 4 23 20 26 12
> NB. oops the ordinary integers in */pns though just plain integers
>don't produce a correct result for pp just because x: is used in the
>statement
> NB. Now tag the first calculation with x:
> pns=:x: ~.4 p:i.53
> pns
>2 3 5 7 11 13 17 19 23 29 31 37 41 43 47 53
> NB. looks the same, but treated differently
> pp=:x:*/pns
> pp
>32589158477190044730 NB. correct
> NB. and gives different and correct result below
> x:pns|pp
>0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 NB. what i needed, but also
> pns|pp NB. didn't need the x:
>there; ... it is just pns that needed to be set up "right"
>0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
>
>
>Thanks again for the earlier advice.
>
>--
>Regards,
>Gerry Jensen
>02 9713 6004
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
