Here's a really perverse solution ... bpv=: 3 : '}:8}.5!:5<''y''' bpv z alpha bravo charlie
I've played with Roger's suggestion: }: ; z,&.> ' ' to get: rrb=: [: }.@; SP ,&.> ] (where: SP=: ' ') I've also done some repeated timings of solutions to-date, with a longer z of 1000 atoms. Here's a sample... 0.000110016 (}.@((<SP) ;@,. ]))z NB. current b2o (6Sep12) 0.000179138 (;:^:_1)z NB. ;:inv 0.000226898 ([: }: [: ; SP ,&.>~ ])z NB. Roger 0.000220947 ([: }:@; SP ,&.>~ ])z NB. Roger, combining }: and ; 0.000233917 ([: }.@; SP ,&.> ])z NB. Roger, elim ~ 0.000143127 ([: }:@; (<SP) ,.~ ,.)z NB. Ian1, combining }: and ; 0.000147095 ([: }: [: ; (<SP) ,.~ ,.)z NB. Ian1, avoiding ,each 0.000223083 ([: }. [: ; SP ,&.> ])z NB. using ,each 0.000231018 ([: }.@:; SP ,&.> ])z NB. c/f Roger, elim ~ 0.000224152 ([: }. [: ; SP ,&.> ])z NB. c/f Roger, sep }.@:; 0.000196838 bpv z NB. perverse soln bpv comes out surprisingly well. As does: ;:inv On Thu, Sep 6, 2012 at 5:46 PM, Roger Hui <[email protected]> wrote: > Nice. > > > > On Thu, Sep 6, 2012 at 9:44 AM, Aai <[email protected]> wrote: > >> >> ;: inv z >> alpha bravo charlie ... ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
