Re: [SVG12] comment

Andrew,
I believe that you are requesting that SVG support a feature which is 
sometimes called "gradient mesh". Illustrator has a tool on its tool 
palette called "Mesh tool" which is used to create such graphics. The PDF 
language spec uses the term "Coons patch mesh" and supports that plus a 
half dozen more vector-based advanced shading techniques.

These features are powerful, create cool effects, and generally cannot be 
approximated by other graphical techniques, other than by rasterizing. A 
large subset of these features are included in Illustrator and (I believe) 
CorelDraw. As I mentioned, they are supported in PDF, including the ISO 
standard subset of PDF for document archiving, PDF/A.

While powerful, these features aren't used that much. Because of this, the 
SVG Working Group hasn't received very many requests for advanced shading 
techniques so far and thus has given priority to other SVG issues. For 
example, there have been countless requests for enhancements to SVG in the 
general area of supporting user interface development and multimedia, but 
just a few people who have even mentioned gradient meshes.

Despite the small amount of industry interest so far, the SVG Working Group 
does have advanced shading on its list of features to consider for the 
future. It seems likely that Coons patches at least will get into SVG at 
some point, at least in SVG Full, but it is unclear when exactly that will 
happen. One issue facing the SVG Working Group is whether there are viewers 
which will implement a given feature. The W3C process requires two 
interoperable implementations of each new feature. A feature cannot get 
into the standard until two organizations find a way to implement the 
feature (which for SVG usually means within an SVG viewer). The viewer 
companies  often have their hands full and cannot implement all requested 
features.

Jon Ferraiolo
Adobe
Member SVG Working Group

At 03:30 AM 7/6/2005, andrew matseevsky wrote:
>Dear Sirs, I think, that SVG specification should be expanded a little. 
>Now it doesn't seriously supports gradient filling (I know about linear 
>gradients). Look at two examples, where and how gradient filling could be 
>used (I'm not sure, that attachments won't be deleted). Skoda.png shows, 
>how one can paint something using gradient fills. BorisRVR2.jpg shows, 
>what result could be achieved in vectorization (I made 
>raster-to-vector-to-raster conversation). Corresponding SVG would have the 
>size approximatedly 350K. When you seriously want to convert some 
>photorealistic image from raster to vector format, some variant of 
>gradient filling is a must. Result is enough impressive and I have no need 
>to comment it. I think, some additional info could be inserted into SVG 
>specification in such way:
>Old variant with constantly colored area
><path d="M790 380 l0 41 c-5.742 192.3 -179.9 359.3 -372 362 l-21 0 c-197.6 
>4.443 -380.8 -168.5 -384 -367 l0 -29 c2.254 -194.1 174.1 -361.7 365 -370 " 
>fill="#255" stroke= "red" stroke-width="5" />
>New variant with gradient filling
><path d="M790 380 l0 41 c-5.742 192.3 -179.9 359.3 -372 362 l-21 0 c-197.6 
>4.443 -380.8 -168.5 -384 -367 l0 -29 c2.254 -194.1 174.1 -361.7 365 -370 " 
>filling="v112 343 #ffccaa v343 298 #ff00dd" stroke= "red" stroke-width="5" />
>Here number of vertices is unlimited, large area usually requires more 
>than the small one. Each vertex contents two coordinates, relative or 
>absolute, and RGB color. Viewer interpolates this info to the whole area, 
>bounded with path. There are some another formats, what could be expanded 
>in similar manner. Adobe Illustrator, for example. This is the way (in 
>fact, one of possible) to make SVG graphic much more flexible and impressive.
>Andrew Matseevsky, Kamchatka.
>
>
>

Received on Friday, 15 July 2005 04:00:43 UTC