Re: Enabling pixel grabbing in SVG just like Canvas

Hi Rob,
Sounds intriguing to me. Maybe:

<svg id="aaa" canvas="true">...</svg>

In the DOM, you could get the canvas object via

var canvas = document.getElementById("aaa").canvas;

This markup approach would allow you to make canvases out of subtrees of
SVG objects and allow reference via <use>, which would use the saved raster
rather than render  the content from the DOM:

<svg>
  <g id="bbb" canvas="true">
     <!-- Complicated graphics -->
  </g>!
   <!-- reference the saved canvas --->
   <use href="#bbb"/>
</svg>

Something in this space could provide huge performance gains for certain
types of complicated graphics.

Jon




|------------>
| From:      |
|------------>
  >-----------------------------------------------------------------------------------------------------------------------------------------|
  |"Robert O'Callahan" <robert@ocallahan.org>                                                                                               |
  >-----------------------------------------------------------------------------------------------------------------------------------------|
|------------>
| To:        |
|------------>
  >-----------------------------------------------------------------------------------------------------------------------------------------|
  |Jon Ferraiolo/Menlo Park/IBM@IBMUS                                                                                                       |
  >-----------------------------------------------------------------------------------------------------------------------------------------|
|------------>
| Cc:        |
|------------>
  >-----------------------------------------------------------------------------------------------------------------------------------------|
  |www-svg@w3.org                                                                                                                           |
  >-----------------------------------------------------------------------------------------------------------------------------------------|
|------------>
| Date:      |
|------------>
  >-----------------------------------------------------------------------------------------------------------------------------------------|
  |10/03/2009 06:36 PM                                                                                                                      |
  >-----------------------------------------------------------------------------------------------------------------------------------------|
|------------>
| Subject:   |
|------------>
  >-----------------------------------------------------------------------------------------------------------------------------------------|
  |Re: Enabling pixel grabbing in SVG just like Canvas                                                                                      |
  >-----------------------------------------------------------------------------------------------------------------------------------------|
|------------>
| Sent by:   |
|------------>
  >-----------------------------------------------------------------------------------------------------------------------------------------|
  |rocallahan@gmail.com                                                                                                                     |
  >-----------------------------------------------------------------------------------------------------------------------------------------|





On Sun, Oct 4, 2009 at 2:17 PM, Jon Ferraiolo <jferrai@us.ibm.com> wrote:
  At SVG Open, there was discussion today about adding Canvas's pixel
  grabbing APIs to SVG, with claims that SVG implementations don't
  necessarily retain a rendering surface. It now occurs to me that there is
  a CSS property ('enable-background') in the SVG 1.1 spec just for this
  purpose:

  * http://www.w3.org/TR/SVG/filters.html#AccessingBackgroundImage

  Pixel grabbing APIs in SVG could require that 'enable-background' is
  placed on the outermost <svg> element as a hint to the implementation
  that it needs to retrievable retain pixel values for image.

  Note that implementations pretty much have to implement saved rendering
  surfaces to do group opacity even if they haven't implemented filter
  effects yet.



Why don't we just make it more convenient to render SVG to <canvas> and
then use canvas's getImageData?

Rob
--
"He was pierced for our transgressions, he was crushed for our iniquities;
the punishment that brought us peace was upon him, and by his wounds we are
healed. We all, like sheep, have gone astray, each of us has turned to his
own way; and the LORD has laid on him the iniquity of us all." [Isaiah
53:5-6]