Thinking this through just a little bit more: NB. brute force adverb: pe1Ab=: 1 :'m&([:+/@I. 0 +./ .= (|/ i.))'
NB. moebius adverb: mu=: (1 _1 0 */@:{~ 2 <. _ q: ])"0 mf=: -.&0@(- * mu)@(, ~.@,@:(*/)~) pe1Am=: 1 :'(mf m) +/@([ * 2 ! 1 + (<.@% |)~) <:' 3 5 pe1Ab 1000 233168 3 5 pe1Am 1000 233168 3 5 pe1Ab 3 5&([: +/@I. 0 +./ .= (|/ i.)) 3 5 pe1Am 3 5 _15 +/@([ * 2 ! 1 + (<.@% |)~) <: 3 5 pe1Ab 1e6 233333166668 3 5 pe1Am 1e6 233333166668 timespacex '3 5 pe1Ab 1e6' 0.028942 2.51717e7 timespacex '3 5 pe1Am 1e6' 6.4e_5 8576 3 5 pe1Am 10000000000000000x 23333333333333331666666666666668 Thanks, -- Raul On Thu, May 5, 2016 at 2:48 PM, Raul Miller <rauldmil...@gmail.com> wrote: > If this matters: > > mu=: 1 _1 0 */@:{~ 2 <. _ q: ] NB. moebius > > But it's maybe simpler to describe this as removing the "double > counting" which would result from counting multiples of 15 twice (once > for multiples of 3 and the other as multiples of 5). > > -- > Raul > > On Thu, May 5, 2016 at 10:03 AM, Mike Day <mike_liz....@tiscali.co.uk> wrote: >> Yes, quite right! You're my conscience as ever, Raul. >> >> Sorry - I misremembered and didn't check. >> Moebius mu(n) is 1 for 3 and 5, _1 for 3*5=15, and 0 for powers higher >> than one of 3 and 5 and their products, so (eg) 0 for 25, 27 etc. >> Also, 99 is the upper inclusive bound. >> >> So: >> +/99 (]*2!>:@(<.@%|)) 3 5 _15 NB. correctly hand-coded moebius! >> 2318 >> >> cf >> +/I.+./3 5 (0=|)"0/i.100 NB. Direct method >> 2318 >> >> Right idea. Previously almost completely wrong example!!! >> I hope it hasn't put Geoff off. >> >> Mike >> >> >> On 05/05/2016 14:31, Raul Miller wrote: >>> >>> On Thu, May 5, 2016 at 4:31 AM, Mike Day <mike_liz....@tiscali.co.uk> >>> wrote: >>>> >>>> +/1683 1050 _594 315 _250 162 _135 >>>> 2231 >>> >>> Er... is that the right answer, somehow? >>> >>> Thanks, >>> >> >> >> >> --- >> This email has been checked for viruses by Avast antivirus software. >> https://www.avast.com/antivirus >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm