Re: Thoughts about variable-width strokes in SVG2

On Tue, May 7, 2013 at 12:07 AM, Brian Birtles <bbirtles@mozilla.com> wrote:

> (2013/05/07 15:48), Rik Cabanier wrote:
> >     Because as you push harder on the tablet or you rotate the stylus to
> >     use only the narrow end you expect the stroke width to change.
> >
> >
> > OK.
> > Will that pressure point coincide with an actual point? What if you just
> > draw a straight line?
>
> Yes, it has to coincide with an actual point due to the way pointer events
> and touch events are defined.
>
> I don't understand the question about straight lines.


If you have a straight line, there will only be 2 points but there might be
multiple widths.


>
>
> > I think that's more complicated than it needs to be.
>
>> You most likely want to reuse a variable width stroke definition across
>> multiple paths. Tying it to points makes that impossible.
>>
>
> This is why we need real use cases. What is the use case where you
> actually re-use variable-width stroke definitions?
>

http://bit.ly/ZDlWyC for lots of art that's using variable width strokes.
This article: http://adobe.ly/18R887J shows reuse of variable width strokes
on eyelashes and lips.


>
> Here's one: You have a diagram where boxes that represent processes are
> represented using a stroke that is thickest at the first and third corners.
> I've drawn an example picture at.[1]
>
> You resize the width of the box depending on the length of the text
> content.
>
> If the stroke widths are tied to percentage offsets of the path length
> then as the aspect ratio of the box changes you will get odd results where
> you get a bulge before or after the corners of the box. You have to
> generate a different 'stroke-widths' property for every such box and if you
> change the text content dynamically you have to recalculate it.
>
> If you tie the stroke widths to points in the path you can re-use the same
> variable width stroke definition on all such boxes regardless of their
> aspect ratio. If the text content changes, you don't need to adjust the
> styles at all.
>
> So it's not the case that tying to points prevents reuse. That said, I'm
> pretty sure there will be use cases where percentage-based offsets make
> more sense. Without real apps though, we're just guessing.
>
> We agree about only supporting symmetrical and positive stroke widths for
> now. That's good. Also, we agree that if we have to choose between absolute
> stroke widths or magnification factors, magnification factors are more
> useful. That's good too.
>
> I think the next step is to look for a middle-ground approach that allows
> addressing the drawing app use case and also situations where you want to
> tie a stroke width to a particular point in the path.
>

I think that use-case is pretty obscure and can be achieved with simple
math.


>
> [1] http://people.mozilla.org/~**bbirtles/variable-width-**
> stroke/variable-width-stroke-**eg.png<http://people.mozilla.org/~bbirtles/variable-width-stroke/variable-width-stroke-eg.png>
>