Re: Thoughts about variable-width strokes in SVG2

On Tue, May 7, 2013 at 1:36 AM, Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>wrote:

> I think, you can provide all options,
> if you align with the corresponding functionality of animateMotion.
> If one has a corresponding set of attributes and lists, one can
> get all the functionalities.
> This allows as well to use cubic Bezier curves and a smooth
> change of the width.
> The Catmull-Rom method seems to have the problem, that
> the derivatives typically do not approximate intended paths
> very good, therefore one needs a lot of points (large document
> size) to approximate with good quality.
> This is ok for fast approaches, but advanced authors should
> have the option to provide control points as well.
>

That sounds reasonable. Going with bezier might give more control.
However, at the same time beziers have certain capabilities (for instance
they can fold over themselves) which don't make sense on a stroke.


>
> Asymmetrical stroke-width can be relevant as well,


We decided to move that to a later date.


> negative
> width is not completely useless.  If an author
> sets it to a negative value - is it a problem for implementation?
>

I *think* we should spec that negative numbers are interpreted as 0.


> If not, then one can interprete it simply as a width to the other
> side, this would allow a simple method to get a zero width
> somewhere within the segment without the need to specifiy
> this as a point of the path or knowing exactly the path fragment
> length.
>
> The asymmetry itself is not really a problem,  because one
> can or should align it according to the pace of path data,
> respectively the directionality of the path data as used
> for some other features as well, therefore it is already defined,
> what the direction is and therefore it is known, what it means
> to be on the left or the right of a path.
>

Yes. We can add this later.