Re: Proposal: <star> element

On Mon, Apr 14, 2014 at 2:47 PM, Tab Atkins Jr. <jackalmage@gmail.com>wrote:

> On Mon, Apr 14, 2014 at 11:41 AM, Rik Cabanier <cabanier@gmail.com> wrote:
> > On Mon, Apr 14, 2014 at 10:59 AM, Tab Atkins Jr. <jackalmage@gmail.com>
> > wrote:
> >> Private discussion with Philip has convinced me that adding a <star>
> >> element (or a <polar> element, or any other particular instance of
> >> something that's star-like and possibly does more) is probably not
> >> worth it.  Stars happen, but they're not really common.
> >
> > I thought so too, but searching for "SVG images" brings up a lot of
> artwork
> > that uses stars (and triangles which are also stars)
> > In addition, applications such as Illustrator and Inkscape offer them as
> > primitives alongside circles and rectangles and it's easy to find js
> > libraries that support them (ie http://paperjs.org/reference/path/).
>
> I could be convinced by arguments like this, though the definition of
> what a "star" is is obviously a bit loose (as evidenced by the
> multiple substantially different attempts to provide them).
>
> >> Plus, the
> >> bearing command, which I think *does* justify itself, makes generating
> >> stars fairly easy.
> >
> >
> > I don't see much value in the bearing command. Do you have any examples
> of
> > drawing applications or popular js libraries that offer this?
> > It is also more invasive as it impacts the already complex path command
> > string. See http://www.w3.org/TR/SVG2/paths.html#PathDataLinetoCommandsfor
> > all the commands that are impacted.
>
> The bearing command is useless for a drawing application.  It makes
> hand-authoring far easier, particularly when performing some actions
> that are symmetrical about some axis.  Rather than doing trig to
> create a pentagon, you can just do some simple middle-school geometry:
>
> <path d="h10 b108 h10 b108 h10 b108 h10 z">
>
> ~TJ
>

I also am concerned about the bearing command due to the complexity it adds
to path handling. As the most basic drawing primitive, we want path to be
fast.

Looking to the future, I believe that most SVG will be generated by some
other program, be it a javascript library, or a plotting package, or a
vector drawing tool, or a custom visualization program. My sense is that we
are already in that world, but nobody seems to have data to confirm it one
way or the other.

The prevalence of stars in existing artwork is not enough. We also need to
know whether or not someone hand-authored it along with the amount of path
data needed to draw it. That's what would inform our benefit analysis.

Stephen.