SVG Working Group Teleconference -- 23 Sep 2016

<scribe> Meeting: TPAC 2016

<shepazu> Draft charter: https://w3c.github.io/charter-drafts/svg-2016.html

Charter

<scribe> scribe: nikos

<scribe> scribenick: nikos

nikos: Charter period ends in November. Doug has prepared a draft charter that focuses largely on completing SVG 2 with a test suite. So we've cut back the number of deliverables.

shepazu: As you know w3c is going through a reorganisation. Along with that we are evaluating all of our WGs to see where we continue to put resources
... svg has had low participation from implementors so we need to continue to examine whether to continue work on svg
... we published CR last week
... After discussing things with the svg wg and with plh, I came up with this charter
... it does two things
... 1. tightens the specs that we are working on
... 2. syncs joint specs with fxtf charter
... the main thing that we did is put svg 2 as the core and dropped most of the other modules
... we do want to continue the module approach in future if we have manpower
... there is the a11y tf
... browsers have been great about implementing features to make svg more accessible
... there is the prospect of future work in that area, but depends on implementor interest
... then we get into the joint specs. We consider those to mostly be in the domain of the CSS WG at the moment - that's where the energy is
... some of those are CR and it would be good to move them along
... could help marshal man power and examine how the CR specs relate to implementations
... the last set of specs are svg parameters and svg connectors
... svg parameters is probably going to be folded into a spec Tab is working on
... one of which is css variables, the other is part of wicg
... connectors is an accessibility feature. We'll probably fold the semantic aspect of this into one of the accessibility specs
... make a connector role rather than an element
... the last three specs are the paths, markers, and strokes - these are the beginnings of the modularisation which will continue after svg 2
... paths and strokes could be part of the FXTF

nikos: They're also in the charter so that we can publish updated WDs with some tidy up and status information so as to not confuse people

shepazu: they're not high priority
... this charter extends for one year - what we'll likely do is make sure we have all the tests and try to gather developer feedback
... we would probably not continue the group after we complete the test suite unless there is a big increase in interest
... maybe move it to a cg
... the last thing is the authoring guide, which includes accessibility recommendations
... it could also explain about the features of svg 2 to help solicit feedback

TabAtkins: After SVG 2 is done and you set up a cg or something - would you consider moving this to wicg or web platform or css or something?

shepazu: everything is on the table
... personally I'd like to see it continue
... but don't care if it's as the svg wg - just want it to continue
... there's also talk of a graphics working group
... will be some stuff going on in the vr space
... + canvas, etc

Rossen: are all the specs that are here really necessary? The css ones are already on our charter

<TabAtkins> https://tabatkins.github.io/specs/svg-params/

Rossen: do you expect much contribution from the svg wg people?
... Tav and Amelia have been very productive
... I suspect there are some joint deliverables they would like to help with
... for all practical purposes they are CSS specs now
... think that cutting back the deliverables more would help focus on svg testing, verification, and PR readiness
... that would be really good to see in that timeframe
... as to what's going to happen afterwards? Unless there's renewed interest in implementors, I don't see how we can move things forward
... I would need to see very clear timelines and accountabilities if we are going to go beyond svg 2

shane: think there's a tangible benefit to moving them to the sole domain of the css wg - makes it easier to resolve where neccessary

shepazu: Tab, do you see a point having svg params in our charter?

TabAtkins: no, can move to wicg and css

shepazu: so charter would be svg 2, 3 accessibility specs, svg connectors (as an accessibility spec)

Rossen: why is it not included in one of the other 3 accessiblity specs?

shepazu: it's not in scope of the other three

Rossen: what is your target for this spec in the charter timeframe?

shepazu: I would like to move the spec as it is to CR in that time, but depends on my time and interest
... possible I could put it in WICG

Rossen: think that's the best path forward

shepazu: I might get push back from accessibility people

TabAtkins: wicg isn't a place to kill things.

shane: it's a place where you can organise outside of the wg structure to move things faster

shepazu: that leaves the 3 svg next modules

Rossen: could they be published during this charter period? then we could remove them from the charter

nikos: yes we could

TabAtkins: We're cannibalising svg strokes and making it apply to css
... but I'm happy for it to be here or in wicg

shepazu: I'm happy to remove them and be totally clear about the focus of the charter

<TabAtkins> https://drafts.fxtf.org/paint/

<TabAtkins> our UD draft for stealing fill/stroke for CSS ^^^

Rossen: I'm happy to resolve to publish now and then we can remove them

shepazu: there's other language that talks about the fxtf

shane: so what will we do with the fxtf?

shepazu: we'll just shut it down and move the specs to css
... the rationale for this charter is to strip it down to what we do in the next year
... which is what we've been doing anyway
... I'll rework the charter - dramatically strip down the scope section
... the other thing I removed was svg mapping because there hasn't been much work on it
... but we will have some mention in the scope that there will be discussions about level of detail, etc
... with those changes, does this seem reasonable to everyone?
... I'm planning to use the authoring guide to drum up some interest so we can see which features people really want

TabAtkins: svg 2 lists the changes since svg 1.1 right?

nikos: yes, there's the changes appendix, but we also created a nicer list on the github wiki that people have been referring to

RESOLUTION: Accept the 2017 recharter with reduced deliverables focused on completing SVG 2 test suite and progressing the core specification and the accessibility TF work

shepazu: should I still list CSS WG as a liaison ?

Rossen: yes

nikos: yes because fxtf people will probably still want to provide input to those specs and it would be good to recognise that

SVG 2 implementation plans

https://github.com/w3c/svgwg/wiki/SVG-2-new-features

<TabAtkins> ScribeNick: TabAtkins

<nikos> http://tavmjong.free.fr/SVG/INKSCAPE/

<nikos> https://developer.mozilla.org/en-US/docs/Web/SVG/SVG_2_support_in_Mozilla

nikos: AS mentioned before, there's a changes appendix since SVG 1.1, which is super long.
... Created this wiki page in a nicer format.
... Some people have turned this into an impl matrix.
... Sebastian Zarter created an MDN doc about SVG support in Moz.
... Slowly testing how things work.
... Tav has published a page on his blog which lists Inkscape's support and plans, what they consider high and low prio.
... Would be good to get feedback from the other brwosers.
... Any plans, what parts of SVG2 they might put manpower towards, what parts they think we should forget about.
... It's been myself, Tav, and Amelia working for a while, and we'd like to get feedback from the wider community.
... If you haven't thought about it enough yet...

Rossen: Yeah, we haven't been focusing on SVG the last few releases.
... We do have some SVG interop issues we need to work on, and as part of that work I'm expecting we'll work on that.

shepazu: Anything we can to help?

Rossen: Not really, we have enough people.

shepazu: We're gonna drum up interest among devs, is that appropriate?

TabAtkins: Yes.

Rossen: Definitely. How you gonan do it?

shepazu: First, SVG Authoring Guide.
... Second, somebody was working on an SVG2 polyfill. Some things will be easy to do, some will be impossible.
... Or extremely hard.

nikos: Not sure they've considered what'll be involved really.

shepazu: And Lea mentioned doing smaller polyfills for particular features.

TabAtkins: And Houdini's Paint API went to CR yesterday, which allows polyfilling th enew paint servers.

leaverou: Tho not usable in SVG as image or background-image

TabAtkins: Yeah.

shepazu: So once we've done some of this, we plan to reach out to prominent devs - Lea, Coyier, Souiadan, etc.

<shane> ^ technically we resolved to take paint to CR pending resolution of remaining issues

shepazu: Here are the feature, they're all effectively at-risk, it's up to us to decide which is most interesting.
... So get them to write articles, raise up user's voices, etc.
... Not force it, but find out what are their priorities.
... Potentially talking to DevRel from the browser vendors, see if they're interested.
... People haven't thought about SVG being "renewed", it has a bad history of that. But if we show them new things - wrap text, don't need to make new markers for every new color, etc. Many people don't know what they're missing.

leaverou: Also I think many people generate their SVG, so they don't feel many of the lacks, but many people hand-author HTML and CSS so the lacks are felt more strongly.

shepazu: Yeah, plan to reach out for tools - D3, etc - and see what they can help with.
... Hopefully within those user communities, among people who use those libraries, we can also generate interest.

TabAtkins: Sounds good.

shepazu: With W3C's reorg, PLH has said that it doesn't look like there's interest in SVG anymore, should we shut the group down.
... Fair question, we have limited resources.
... I think it's appropriate to recharter the group - we end in one month. I think it's appropriate to invest in seeing if we can get devs helping y'all decide what your priorities are.

shane: It's also worth saying that we have been focusing in Chrome on the rationalization and rmeoval of old deprecated APIs/features - for SVG2 we'll cotninue doing that.
... Slimming down things, and targetting parts of the model that make things cleaner and simpler.
... Turning everything into a presentation attribute was part of that.

nikos: Even if all the new graphics features got dropped, there's still lots of value in the new spec tightening things up.

shane: Yes, very valuable.

shepazu: Just dropping xlink...

[general assent]

[some unminuted discussion about browser planning]

shepazu: Interesting in testing - improvements we made to the spec to increase interop. Second, new features that are already getting attraction

nikos: Which tend to be developer convenience things.

shepazu: Making sure there are tests for those so they can spread to all browsers.
... We learned that from IE - until IE implemented SVG, nobody used it, and then suddenly everyone did.
... And third, new new features - not necessarily low-haning fruit in terms of testing.

Rossen: Do you plan to solicit for testing resources?

shepazu: Two things. We're moving to WPT.
... Two, we're actively asking people in teh community, thru shotgun and targeted emails, asking to help us do testing. Grassroots stuff.
... When I say we're going to WPT, W3C is reorging, and part of it is having people focus on testing. We've already had some people come to us and say they'll help.

nikos: Paul Le Beau said he has a bunch already that targets corner cases, he's interested in helping convert to whatever format.

shepazu: We're gonna have to do some training to people to do testing.
... But the lsast resource is the impls.
... At this point we don't need valuable engineer time, we dont' need spec writers, we need testers, QA people.
... We really just wanna finish SVG2 at this point. Even if impls lag behidn the end of the charter...
... And hey, maybe the decision after a year decides that some features get punted to SVG3, if ever, and here's the SVG2 Rec.
... We want a completed SVG2, with whatever features we can.
... So when PLH asks me if this is reasonable as an SVG approach.
... Y'all would be supportive of this?

TabAtkins: Yeah.

Rossen: In the case of not continuing the charter, if W3M decides against it anyway...

shepazu: Yeah, then that's... it.
... We need dev feedback to get things stable; we wouldn't be able to do that.

nikos: Is there anything we've missed that y'all think is important?

Rossen: I don't think so, right now.

nikos: Next topic is moving along joint specs, but we already discussed that as part of the charter.

shepazu: Would it help for me to move those along, or just focus on SVG?

Rossen: The CSS/FX specs? Dont' worry about it. We have plenty of editors already.

nikos: Then I guess we can consider this meeting adjourned.

- DRAFT -

SVG Working Group Teleconference

23 Sep 2016

Attendees

Contents

Charter

SVG 2 implementation plans

Summary of Action Items

Summary of Resolutions

Scribe.perl diagnostic output