W3C home > Mailing lists > Public > public-svg-wg@w3.org > July to September 2010

Point-on-Path Curves (was: Agenda July 13 2010 SVG WG telcon)

From: Doug Schepers <schepers@w3.org>
Date: Tue, 13 Jul 2010 23:03:36 -0400
Message-ID: <4C3D2908.4070802@w3.org>
To: Chris Lilley <chris@w3.org>
CC: Erik Dahlstrom <ed@opera.com>, public-svg-wg@w3.org
Hi, Chris-

Chris Lilley wrote (on 7/13/10 1:29 PM):
>
> ED>  - Catmull-Rom curves ED>     http://schepers.cc/?p=243
>
> awesome.

:D


> Is that the same as the thing Robert Miner was talking about?

I don't recall, actually.  It may have been, but I think that was Spiro 
curves.


> Smooth curve through a bunch of points is a needed feature. It can be
> complicated if there are constraints on where the curve goes between
> points. Sometimes curves like that get really wild yo-yo effects
> between points.

The tension on my Catmull-Rom is pretty tight.  Spiro makes astounding 
curves, but they are a bit unpredictable.


> By the way, besides producing drawn geometry, there is also an
> application of 'smooth curve through some points' to gradients.
> Consider each stop colour as a point in 3D space (the colour space).
> What we currently do in gradients is draw a straight line segment
> between each of the points. This leads to visual discontinuities as
> the slope of the line changes abruptly from point to point. A smooth
> curve avoids this. But has to be constrained to not move outside the
> colourspace, otherwise some colours in the resulting gradient are
> not, well, colours.

Great point.  I alluded to some of the other non-drawing cases, like 
text-paths, glyphs, motion paths, and keysplines, but the colorspace 
aspect is really important.

I don't know if this also covers the idea of a gradient that "follows" a 
path... so, just like a horizontal linear gradient would interpolate 
along the length of a totally straight horizontal line, a curved 
gradient would interpolate along the length of a line with arbitrary 
shape.  Assuming that we do want to do diffusion curves, having a 
curve-based gradient would really help with some of the more subtle 
effects.  I have the feeling that they are related but separate points, 
am I right?


> I know, this was a long 'regrets'. Consider me conflicted. If I get
> guilted enough I might even show up tonight.

We had a long discussion about curves, and it would have been good to 
have you there, but this was only the first discussion of many.

Regards-
-Doug Schepers
W3C Team Contact, SVG and WebApps WGs
Received on Wednesday, 14 July 2010 03:03:38 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:20:12 UTC