On Feb 10, 2014, at 5:55 PM, Attila Vrabecz wrote: de-parenthesize k)" *"4=x+/:x:4&i,1_|i:|!6 getting rid of one reverse k)" *"6=x+/:x:1|i,1_|i:!6 moving the 1| to the end (no need for x and i separately) k)" *"6=x+/:x,:1_|x:1|!6 i like the symmetry of 6 and 6 ;) still 4 bytes longer than J but as Stevan said given the number of primitives in K vs J is quite excellent On 10 Feb 2014, at 20:18, stevan apter wrote: you know, that's not bad, considering how much smaller k is than j. too bad the pure version adds the extra characters for the lambdas, but it really isn't too shabby. On Feb 10, 2014, at 2:49 PM, Attila Vrabecz wrote: / /: is enough k)" *"4={x+/:x}4&(|1_i),i:!6 / dropping the lambda k)" *"4=x+/:x:4&(|1_i),i:!6 On 10 Feb 2014, at 18:58, stevan apter wrote: k)" *"4={x+\:/:x}4&(|1_i),i:!6 a trivial improvement: 2 bytes shorter, but still 10 characters longer than lochbaum's. On Feb 10, 2014, at 1:25 PM, stevan apter wrote: http://scottlocklin.wordpress.com/2013/10/31/shannon-information-compression-psychic/ from one of my favorite bloggers - scott has been learning j. i've just mimicked the j solution he describes. i don't think we can quite match j's beautiful terseness, but i wonder if there's a better k solution than this. q)k)"_*"4={x+\:/:x}{4&(|1_x),x}@!6 "_____*_____" "_____*_____" "____*_*____" "___*___*___" "__*_____*__" "**_______**" "__*_____*__" "___*___*___" "____*_*____" "_____*_____" "_____*_____"