Re: [Groff] polygon drawing compatibility

[Bernd Warken]

On Tue, Nov 20, 2001 at 10:18:17AM -0500, T. Kurt Bond wrote:
> 
> By "classical positioning" you mean "at the end", right?  What should
> be "the end" for polygons: (1) at the last point of the drawing, like
> splines and lines?  Or (2) at the right side of the drawing, with the
> verical position back at the baseline, like circles and ellipses?  Or
> (3) something else?
> 
I think I mailed that often enough.  1) The "end" is the end of drawing.
2) With circles and ellipses (and only these!), the "end" was defined as
the right-most point (as if two symmetric half-figures were drawn).
3) Classically, there isn't anything else, only (1) and (2) are left.

> If we follow the example of splines (which also have multiple delta
> pairs) we'd choose (1).  But polygons are always closed (even if we
> leave off the delta pair that moves back to the start it will
> automatically be closed), so if we follow the example of the other closed
> shapes, circles and ellipses, we'd choose (2).
> 
A closed polygon is neither a circle nor an ellipse, so (2) is not 
applicable.  Moreover, in former mails, we agreed that there isn't a
defined right-most point for an arbitrary poligon (e.g. the rectangle).
So only (1) is left, i.e. the "end" is the same as the starting point.
And that's what also the ditroff(1) had implemented.  So if we make
our polygons not changing position we restore compatiblity with
classical troff _and_ with commercial ditroff.  If we adopt the stupid
point-before-end positioning we might have a good chance to get
Microsoft as a sponsor for groff :-)

Is it clear now?  If not read my former mail another 100 times!

Your examples might be helpful for testing non-groff roff systems.
Who has access to such systems?

Bernd Warken