Getting to SVG 2.0: A report from the SVG Working Group Face-to-Face (May 24th – June 1st 2010)


After months of conference calls, the most recent face to face meeting of the SVG Working Group helped us make progress on this web standard.  The biggest ‘news’ from this meeting was the group’s decision to take the next step to finish the current SVG proposal – “SVG 1.1 2nd Edition” – and send it as a Proposed Recommendation to the W3C Advisory Committee for final endorsement. We also started more serious conversations about SVG 2.0. As the Web has advanced in a lot of areas in and around SVG, one of the most important things we can do as a group is ensure sure different parts of HTML5 work well together. There are good real life examples of this in SVG Fonts and WOFF that demonstrated the challenge the group discussed and how we move forward to make the web more interoperable in the right way; more on that below.

The Meeting

The SVG Working Group has a quarterly face to face meeting. As this one was my first, I was more than happy to arrange for a place for us to meet at the Microsoft Executive Briefing Center in Brussels:

Buildings and waterway in Bruges

This photograph is of Bruges, a short trip to the west of Brussels. The actual meeting looked a bit different

Finishing SVG 1.1

During our first few days, the working group decided that we needed to close out SVG 1.1 2nd Edition Specification. The entire group hunkered down and finished all open items and made our last edits to the spec, prepping it for the next publishing stage, Proposed Recommendation.  We are now analyzing the quality and viability of the test suite, which is the final step.

Moving Forward with SVG 2.0

After moving past the SVG 1.1 2nd edition, the working group started to push forward on SVG 2.0.  We all agreed that we needed to approach the design of SVG 2.0 differently in order to achieve our goals.  The main principles involve scenarios and use cases, test driven feature development, and (most importantly) community feedback.

Over the next few weeks, the working group will draft scenarios and supporting features and socialize them with the community. We want to scope the work in SVG 2.0 appropriately to what will help developers the most.

Hot Topics: Modularity and SVG Fonts

There were a handful of hot topics, many of which we violently agreed on. For example, we agreed strongly about creating “optional” modules in SVG 2.0 in order to make it easier for everyone involved in web graphics to focus on the core important parts.

Modules in SVG 2.0 are more consistent with the CSS working group’s approach. They should make faster progress easier.

As Doug Schepers, the W3C Team Contact for the SVG Working Group, put it plainly we need to : “Help progress the Web Platform.”  Optional modules will replace “profiles.” (Not many people understood the relationship between tiny, basic, and full within SVG 1.1.)

SVG Fonts came up as a candidate for optional status. Chris Lilley, the technical expert on the SVG Working Group officially made the proposal. The WOFF Working Group just standardized fonts, and site developers need more consistency across different web technologies. I agreed with the Firefox representative here. No one saw high demand for the scenarios. We agreed that if there’s a strong demand, we can revisit this decision.

The working group has not reached a conclusion about SVG Fonts. This discussion still goes on today. Another Mozilla person, Robert O’Callahan, recently posted a similar position about SVG fonts.

Consistency Breeds Interoperability

SVG Fonts are one of several technologies that are inconsistent between SVG and other parts of HTML5. We have a CSS/SVG Task Force to normalize features that had previously been isolated to SVG, and to expand them consistently for both HTML and SVG.

This is of course key to our primary mission of progress on the web. It’s important to have a single consistent development model for the web, one that reflects the well understood patterns, frameworks and tools used today, and provides capabilities for emerging needs of developers.

There is even more inconsistency between SVG and CSS around animations.  We all agreed that developers benefit from one consistent model here.  CSS Styling has made a lot of great progress down one path; SMIL is quite different. Reconciling these differences and commonalities between SVG and CSS features (transitions, animations, and transforms) is also part of our responsibility as a standards body working group.

Prior to going to Brussels, Doug Schepers and I exchanged an informal interview around these subjects and Internet Explorer working with the SVG Working Group.  Doug’s reaction to Internet Explorer 9’s support for SVG and his perspective on the status of SVG implementations across browsers is clear.

New Ideas for Graphics

The group also saw some new ideas for graphics; one which caught everyone’s interest was diffusion curves. I was quite surprised by the quality of graphic achieved through a diffusion algorithm.  As an example, a manually rendered set of paths below demonstrates the small amount of data required to render a lifelike image.

a red peper first rendered in diffusion curves and then filled in

Or the reverse where a picture was taken, and a tool reengineered the image into diffusion curves.

blue ash try image and then diffusion curves

Graphics on the web will continue to be exciting for some time to come!

Thanks –

Patrick Dengler
Senior Program Manager

Comments (16)

  1. blah says:

    All you need to do now is support more than 5% of SVG 1.

  2. 8675309 says:

    @ least the blog has had a nice facelift

  3. Ms2ger says:

    Please, please, please, SVG isn't a part of HTML5. Yes, HTML5 explicitly supports including SVG into HTML, but that doesn't make it part of HTML5, just like CSS isn't part of HTML5. I'd hope you knew better.

    (Oh, and javascript URLs still suck. Can't you just use real buttons?)

  4. tobi says:

    Those curves look like an interesting image compression technique.

  5. Josh says:

    SVG 2 is going to be really exciting. What I'd like to see in it is filters being changed from SVG elements to CSS properties.

    @Ms2ger

    You do realise that HTML5 wasn't mentioned once in this blog post, don't you? HTML5 isn't the only thing they need to support.

  6. Michael Kozakewich says:

    The lady bug gzips to 13 KB, while the pepper gzips to 8 KB. That's pretty impressive.

  7. Klimax says:

    1)HTML5 was mentioned – it seems that SVG is considered as partially under HTML5,respectivelly there is need to have it nicely cooperate with rest of doc.

    2)All these talks seem to point what will be supported. Meaning whole SVG 1.1 and it looks like it has priority over canvas.

    (Since canvas will probably require a lot of optimisation in js and in rendering,I am not suprised.)

  8. Fleet Command says:

    Let's assess the significance of this meeting for people who are actually reading this blog post.

    Speaking of SVG 2, why do I not fail to notice that Internet Explorer currently does not support SVG at all? In time, I also do not fail to notice that Internet Explorer 9 has promised to support SVG. As for the customers who benefit from this support, Internet Explorer 9 has also mentioned that it will not be available for Windows XP or any other earlier operating system.

    Result: It doesn't matter for a guy like me whether you've discussed SVG 2 or not; all it matters is that I'll get none of that through Internet Explorer.

  9. Gyrobo says:

    Diffusion gradients look awesome! Is there an effort to get them into the Image Values Module [1] or would that be yet another case of overlap?

    [1] dev.w3.org/…/css3-images

  10. RTFA says:

    @blah

    if you read the linked article http://www.w3.org/…/interview_ie you'll see that IE9 will ship with 20 of 23 of SVG 1.1 modules implemented.

  11. sodaxp says:

    It's great that IE9 supports Windows Vista and later operating systems; IMO Windows XP (born in 2001!) seems too old to support the new features of IE9 ;) Also, from Windows Vista, security in Windows has become more robust. It's time to move on, guys! :D

    Keep up the good work, Microsoft! I hope we can see mucho more on HTML5, SVG and CSS ;)

    Best regards from Peru!

  12. Jon says:

    Are diffusion curves much different from gradient meshes? I thought they were the way to achieve photo-realism in vector images.

  13. Neil Dunensach says:

    @Sodaxp – there's just far, far too many XP installations out there for MS to totally abandon, I feel.  By the time IE9 comes out there will be less, but still a significant number.  I think an IE-9 without the GPU rendering stuff should be an option – that's the main part that spanks XP.  I'd rather see Vista dropped before XP!

    Good to see MS getting up to speed on the SVG stuff though, great news.

  14. Wurst says:

    What about SVG 1.2?

    Also, what about making SVG comparable to 'classic' image formats by making it possible to display it in a script-less sandbox so that SVG could be also used in user-generated content (as an example: making it possible for me to have a SVG image as an avatar).

  15. Jon says:

    It's amusing that web developers constantly *** about people who still use IE6, but then scream and shout when Microsoft says they wont support their 10 year old operating system.

  16. nrdgrl says:

    SVG is great and we are glad to see a renewed interest in it… My company – http://www.visual-integrity.com is specialized in converting vector-to-vector. We've had a module that generates SVG since the beginning so know it well… most people use our software to go from PDF to SVG. We also can go from PostScript and WMF/EMF to SVG via API/DLL or command-line. As it gains momentum as part of IE9, do you think people will be needing to convert legacy content to SVG or that most of it will be generated on the fly?