cool! too bad I was not that well yesterday, but I enjoyed our geometry deciphering session on the whiteboard! :)
On 2013-04-19, at 06:55, stephane ducasse <stephane.duca...@free.fr> wrote: > :) > This is cool that you push athens. > Because this is really important for Pharo. > > Stef > >> Not really impressive (that's why it called new wave) >> >> Here is results of 2 days of preparing/implementing and polishing a >> quadric bezier curve approximation >> algorithm by series of lines. >> >> The red pill.. err fill is polygon consisting purely from lines, >> the white outline is same quadric bezier, but drawn directly by cairo >> with stroke width = 5. >> >> Cubic bezier , then elliptical arc and then generic parametric >> functions ahead! :) >> >> There is two reasons why i implementing highly inefficient (because of >> boxed floats) >> path flattener: >> - some path segments are not supported by cairo, therefore i am >> forced to approximate it with something else (and something else is >> lines ;) >> - having path geometry turned into silly polygon, there is very easy >> algorithm to tell if given point inside or outside of polygon >> >> so, potentially, this will give use following: >> - being able to use any kind of curves as path segments (well, it >> could be a bit slow, but it is up to you, right?) e.g. something like: >> builder >> lineTo: 50@50; >> parametricCurve: [:t | t sin * t squared ] tMin: 0 tMax: 10 to: 20@30 >> >> - being able to detect whether mouse hovers over shape or not :) >> >> >> -- >> Best regards, >> Igor Stasenko. >> <Screen Shot 2013-04-18 at 10.43.07 PM.png> > >
