The reason I like Cliff's code is it lets me draw lines through the color cube between arbitrary corners. In my paper, I reduce the many possibilities this raises by deciding that a "Z-to-Z" mapping gives nice palettes; that is, connect corners on one face in a "Z" pattern, then continue on the opposite face with another "Z" pattern.
On Sat, Aug 10, 2013 at 12:21 AM, Raul Miller <[email protected]> wrote: > (personal response sent off list) > > On Fri, Aug 9, 2013 at 5:18 PM, Devon McCormick <[email protected]> > wrote: > > The basic code I stole from Cliff are these two things: > > > > NB.* pwlin: piecewise linear function > > pwlin=: 4 : 0"_ 0 > > 'p c'=. x > > i=. i.&0 p<y > > ((1-r),r=.(y-~i{p)%-/(i-0 1){p)+/ .* (i- 0 1){c > > ) > > > > hue=: <.@(0.5&+)@(((6%~i.7);255*|."1#:7|3^i.7)&pwlin)"0 > > NB.EG hue 10%~i.10 > > That looks more general than I was comfortable using. > > Generality can be a problem when it obscures the purpose of the code. > So I (perhaps arbitrarily) avoid it when it "feels wrong to me". Of > course, perhaps ironically, that's probably not an explanation which > would be generally useful to other people. > > Here's a rule of thumb though: if making the code general more than > doubles the code length? It's high time to think about if it really > solves anything for you. > > Thanks, > > -- > Raul > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > -- Devon McCormick, CFA ^me^ at acm. org is my preferred e-mail ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
