W3C home > Mailing lists > Public > www-svg@w3.org > August 2012

Re: [svg2] radialGradient @fr constraints

From: Dirk Schulze <dschulze@adobe.com>
Date: Wed, 29 Aug 2012 13:25:49 -0700
To: David Dailey <ddailey@zoominternet.net>
CC: Erik Dahlstrom <ed@opera.com>, "www-svg@w3.org" <www-svg@w3.org>
Message-ID: <92980D20-B846-496F-A4B6-0C3E77256E71@adobe.com>
Just as a note, Firefox and WebKit never let the focus be on the circle itself. Both have a small offset. Your can see that on huge radial gradients. There would be to much noise without this offset.

Note that the HTML Canvas spec did not specify what happens for focus points outside of the radius for a long time. Even if the behavior is specified now, it looks totally different on every browser.

Browsers are limited to the underlaying graphics library. Even if some cases can be special cased, it won't work on all cases. I would suggest to continue disallowing the focus point to be outside of the radius circle and extend this definition for focal radius as well to have compatible implementations. We could go on with the definition of HTML Canvas, but would risk different behavior on different implementations.


On Aug 29, 2012, at 12:04 PM, David Dailey <ddailey@zoominternet.net> wrote:

> Yes, I think that is the issue. I've simplified my example to get rid of the animation and to include only one gradient:
> http://cs.sru.edu/~ddailey/svg/gradient11c.svg
> The different browsers' behavior is shown at http://cs.sru.edu/~ddailey/svg/gradientscreen.jpg
> ASV and Chrome agree with me. (I hold the opinion since it is the coolest effect, and isn't that what reflected radial gradients are largely all about anyhow?)
> Using a pair of these gradients overlaid with opacity provides a very nice animated effect when combined with feDisplacement to give water effects, with the activity beyond the horizon providing a somewhat realistic impression of fractal reflection.  
> Cheers
> David
> -----Original Message-----
> From: Erik Dahlstrom [mailto:ed@opera.com] 
> Sent: Wednesday, August 29, 2012 7:51 AM
> To: www-svg@w3.org
> Subject: Re: [svg2] radialGradient @fr constraints
> Hi David,
> You mean the condition displayed in figure 8 here, https://svgwg.org/svg2-draft/pservers.html#RadialGradientNotes right?
> The discussion and rationale for that addition can be found here:
> http://www.w3.org/2011/07/29-svg-minutes.html#item05
> [[ Summary of the issue: when the focal point is on the circle edge, with repeat, then the distance between the first and last stop for the repeating colors is 0 and the (SVG1.1) spec. does not define what should happen. ]]
> On Wed, 29 Aug 2012 12:32:01 +0200, David Dailey <ddailey@zoominternet.net> wrote:
>> Fwiw, the example at
>> http://srufaculty.sru.edu/david.dailey/svg/newstuff/gradient11c.svg
>> (that is now many years old and appears in more than one book) is 
>> rendered in ways that I consider proper in only Chrome and ASV. Opera 
>> makes it too grainy; Safari doesn't reflect gradients; FF refuses to 
>> differentiate between fx and cx (and even crashed for me thrice this 
>> morning while running it for a while!); IE9 doesn't reflect the 
>> gradients beyond the horizon (and they seem to have dropped animation 
>> from the graphical web).
>> HTML had the opportunity, in 2007, to have canvas follow SVG's syntax 
>> for radial gradients. Some people even advised as much. Reversing the 
>> role of leadership here would seem silly. Apple, despite, its recent 
>> victories in the wars*, should not have so unilateral a role in 
>> defining graphics for the web.
>> Regards
>> David
>> *There was a reason the policy of mutually assured destruction worked.
>> -----Original Message-----
>> From: Erik Dahlstrom [mailto:ed@opera.com]
>> Sent: Wednesday, August 29, 2012 4:33 AM
>> To: www-svg@w3.org
>> Subject: [svg2] radialGradient @fr constraints
>> Some issues regarding the added 'fr' attribute on radialGradient:
>> a) Should 'fr' be allowed to be negative? (this is disallowed in 
>> <canvas>). What should happen if it is?
>> b) Should we still keep the constraint[1] to move the focal point 
>> inside the other circle? <canvas> doesn't do this. What the spec 
>> currently defines means some kinds of conical gradients aren't 
>> possible to do with <radialGradient>.
>> c) Related to b): the case where the focal point is outside the other 
>> circle, but the focal radius makes the two circles intersect, how 
>> should that be handled?
>> Proposal:
>> a) disallow negative values for 'fr', and let these cases fallback to 
>> the lacuna value '0%'.
>> b) remove the constraint and handle it the same as in <canvas>, noting 
>> that this may break some existing content. If we do this way it 
>> doesn't matter how the two circles are positioned relative to one 
>> another, so it addresses c) as well.
>> [1] https://svgwg.org/svg2-draft/pservers.html#RadialGradientNotes
>> --
>> Erik Dahlstrom, Core Technology Developer, Opera Software Co-Chair, 
>> W3C SVG Working Group Personal blog: http://my.opera.com/macdev_ed
> --
> Erik Dahlstrom, Core Technology Developer, Opera Software Co-Chair, W3C SVG Working Group Personal blog: http://my.opera.com/macdev_ed
Received on Wednesday, 29 August 2012 20:26:26 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 8 March 2017 09:47:29 UTC