Re: SVG 2.0 (was: Minutes, 12 November 2009 SVG WG telcon)

On Fri, Nov 13, 2009 at 4:26 PM, Doug Schepers wrote:

>> Thank you, Doug. So e.g. the planned 2.5D transformations spec will
>> move to 2.0 as well?
> Correct.

Interesting :)

>> Is there general agreement at this stage what primary focus of 2.0
>> would be, if such could be specified?
> The primary focus, as far as I see it, is refining the SVG DOM to be more
> easily implemented and more performant, and adding features that make SVG
> easier to use (syntactic sugar, parameters, canned filter effects, that sort
> of thing), as well as tightening up the language in the spec and making more
> comprehensive tests to get better interoperability.

I see, thank you.

>> I see that most participators
>> you mention generally fall into "web" category. How much attention
>> could be given to "design" category which is rather omnivore :) and
>> deals with both web and print.
> There are some things that cater more specifically to print than to Web,
> such as the color management stuff.  Many of the improvements needed relate
> to performance and interactivity, so that's obviously stuff that needs to be
> worked out with the Web browser implementers.
> I'm personally less knowledgeable about what other considerations are needed
> for print... can you elaborate?

Well... not necessarily print... :) But we would do with multiple pages.

>> And what would be expected from "design" collaborators?
> You tell us!  We are open to feedback from designers, and have even sought
> it out, but so far haven't heard enough from them.  If you are interested in
> helping follow though on this, let us know what you have in mind.

Well, in Inkscape we have a number of issues that could be solved on
your end, but I have a suspicion that we are actually in a chicken-egg
situation :) I'll just provide some examples of things we need.

Document model:

1. Users want us to move axis origin from Cartesian to top-left. Can
we do this and still make backwards compatible documents?
2. Users want us to make origin shiftable.

Fill and stroke:

1. Multiple strokes (linked offset solves the issue only partly)
2. More types of fill. Earlier this year we were sent a patch that add
conical gradient fill. We had to reject it, because a) it added the
feature inconsistently re. SVG spec and b) didn't provide fallback
rendering. See for

Gradient mesh

It was brought up a lot of times in this very list, but without much
outcome, which is quite a pity. As someone who is involved into
redoing AI tutorials for Inkscape I see way too many interesting
things that are low-hanging fruit to AI users thanks to gradient mesh,
while demanding too much jumping around in Inkscape. Sad, sad.

There are some hot things like diffusion curves too. The problem is:
there are lots of interesting things around, and standardization is
not something that happens overnight.


1. Users want us to make it possible to apply several SVG Filters.
Currently each new preset applied results in appending its primitives
to an existing applied filter.
2. Users want us to make it possible to bypass selected primitives in
a filter. Not that this is a problem on your side, but how exactly the
primitive after the bypass point should behave might need thorough
3. I swear there a lots of filters we could put to a good use. Radial
blur in AI seems to be quite popular, for instance.
4. SVG 1.1 doesn't provide lots of useful blending modes. IIRC
proposed SVG 1.2 Color solved that, but then again here comes
chicken-egg situation.

I think I could provide more of this. The above is just off top of my head.


Received on Thursday, 19 November 2009 14:04:06 UTC