Closed jgamble closed 6 years ago
Okay, implemented in SVG so far in branch splitradius (pay no attention to the radiuszero branch, which may be deleted by the time you read this).
Next step is to implement in postscript. Since radius wasn't implemented in ps in the first place, I'm starting from zero.
I've been hitting my head on this off and on, and finally did a comparison between the old (current) EPS output and my new output. And found that with default values, there is no difference in output, and that the postscript viewers I installed are choking on the files. Which means that there's a bug that goes back to the version 2 days (or even before), or postscript parsers have gotten stricter in the mean time.
This means that the postcript code is holding up the SVG code, which is not good. So I'm merging the code and will open up a postscript-only issue.
Right now a single radius value is used for the endpoints of both comparator and input lines. But the Knuth diagrams in The Art of Computer Programming have, for example, circles on the comparator lines, but not the input lines.
There need to be two radius keys ('compradius' and 'inputradus') to let users mimic the style of TAoCP.