SVG Working Group Meeting Report

The SVG Working Group had a three day Face-to-Face meeting in Hamburg this week. The last day was devoted to working with the CSS Working Group on joint items. I attended most of it via telephone. A lot of the discussions had to do with stategies for getting SVG2 out or with technical stuff. But there were a number of things that might be of interest to Inkscape users and others.
  • Paint Order: It was resolved that the SVG will allow the order in which fill, stroke, and markers for a single graphics element are painted to be specified. At the moment the stroke is always painted on top of the fill, and markers on top of everything. Being able to paint the fill on top of the stroke is quite important for text.
    The letter 'B' with normal fill order, fill on top, and markers underneath.

    A letter with various paint orders.

  • Marker Clipping: At the moment, in order to avoid the path from showing under the tip of an arrowhead, the arrowhead must extend past the end of the path. In Inkscape, this prevents things like snapping the arrow tip to a line since the path end is not at the same place as the arrow tip. One of the SVG group members is going to work out a way to specify a clipping region for a marker that would apply to the path below thus eliminating this problem.
    Marker clipping applied to an arrow and to a circle.

    Demonstration of the use of marker clipping. The red dotted line indicates the clipped area. Clipping only applies to the path the marker is attached to.

  • Marker positioning: The current spec only allows markers placed at the ends of a path or at nodes. The WG agreed that more complex marker placement would be supported, For example one will be able to alternate between two markers and place them every 20px along a path.
    A path with two alternating markers placed at equal intervals.

    A path with markers. The markers are placed by distance along path.

  • Masking: Masks will gain an attribute to specify whether to use alpha or luminance in the masking calculation.
  • Screening Filter Primitive: The group was supportive of my proposal to add a screening filter primitive. It would not be added to the SVG2 specification directly but to the CSS/SVG joint filters specification that has been split out from the SVG spec.
  • Gradients Along/Across Paths (and Variable Stroke Opacity): While group members were interested in this, it turns out to be quite difficult to implement so the group pushed this off to the future. (The Adobe rep said that it was a real pain to export it to other formats.) If Inkscape were to support this somehow (mesh fallback?), it would help push it into the spec.
  • New Stroke Join: Coming from Johan Engelen’s Power Stroke work inside Inkscape, I propsed that the SVG2 specification add a new stroke-join option. Currently you have the choice of beveled, rounded, or mitered. When a path contains two curved segments joined at a sharp point, the mitered option doesn’t look so good. Johan’s Power Stroke allows an “extrapolated” join where the curvature of the paths is taken into account. The SVG WG liked this idea and approved it subject to defining the math needed precisely. One factor in its adoption will be than none of the graphics libraries used by the browsers includes such a join.
    A demonstration of the different existing SVG join styles and the proposed new style.

    The existing join styles with the proposed extrapolated join style.

One thought on “SVG Working Group Meeting Report”

  1. Great to see joints being discussed for SVG2.0 and great to have witnessed Johan’s Engelen discussion at the lgm 2012.

    Any chance the horrible path’s d-attribute (aka graphical assembler) will get an update as well? my vote would go to a more scripted approach as iPath outlined in ‘’.

    But I’m not totally unbiased.

Comments are closed.