- From: Chris Hennick <christopherhe@trentu.ca>
- Date: Sun, 30 Mar 2014 00:48:45 -0400
- To: www-svg@w3.org
- Message-ID: <CAPvCkbgM-zL12rzwyhAEPBRSi0wtt8bwKv1t0xLYDLZkUAxc0w@mail.gmail.com>

Under the current SVG2 working draft, the curve produced by a "smooth curveto" command *isn't* smooth if the path happens to mix quadratic and cubic spline segments, or mixes either kind of spline segment with elliptical arcs. This is a disappointing limitation, since it means one can't combine the shorthands with another technique for shrinking file size -- simplifying cubic segments to quadratic ones when the control points happen to be equal -- or with the use of real elliptical arcs (which Bezier curves can only roughly approximate). Following a straight-line segment, this limitation is understandable, since there's no obvious choice of length for the control handle; but the rest of the time, it makes no sense to me. Is there a compelling reason for it? If not, I propose the following amendments to the commands T,t,S,s: - When either follows a quadratic-spline segment, it mirrors that segment's control point. - When either follows a cubic-spline segment, it mirrors that segment's second control point. - When either follows an elliptical arc of 180 degrees or shorter, it behaves as though that arc were the cubic-spline segment that best approximates it. - When either follows an elliptical arc longer than 180 degrees, it behaves as though that arc were the cubic-spline segment that best approximates the last 180 degrees of it. On a related note, is there any reason we can't have smooth-closepath commands? (They could work even if the first path segment *had* been a straight line.) Sincerely, Chris Hennick Trent University Peterborough, ON, Canada http://softwetware.blogspot.com/

Received on Monday, 31 March 2014 07:54:16 UTC