SVG Working Group Teleconference -- 24 May 2012

<trackbot> Date: 24 May 2012

<heycam> ScribeNick: nikos

Test the web forward event

<krit> http://testthewebforward.org/

CM: Are any SVG group members attending?

DS: I'll be there

<heycam> Doug and Dirk will both be there

CM: focus is on testing, it'd be good to have test repository and framework up and running by there

Tav: frame work is up but repostory isn't there yet
... other way around

krit: which account do you use to commit?

<heycam> http://www.w3.org/Graphics/SVG/WG/wiki/Svgwg.org

CM: svgwg.org account
... need to send ssh public key to jwatt or myself to get access

krit: I'm talking about repository for test suite, not spec

CM: they're on the same machine

<ed> hg clone ssh://svgwg@svgwg.org/hg/svg2-tests

CM: might be possible to set it up so that you use different accounts

krit: can we avoid sending public key for tests?
... want to make it accessible

CM: you mean to people outside the wg?

krit: yes
... we should ask css group and others to commit tests

CM: so anyone from the public could get an account and commit? would there be restrictions?

Tav: it would be like css where you can commit to your own area, needs approval to move

CM: is that done with hooks?
... or is it convention?

krit: no folder besides the approved folder is restricted

Tav: don't forget we're using the 2 step system, so it's more complicated

CM: in the CSS group do you commit directly to the w2c mercurial?

krit: yes

Tav: we should re-discuss our test format
... we have a format we came up with, new tests should follow that format

krit: can't we just do it like the css wg?

Tav: how is that?

<krit> csswg.org

krit: they have a good wiki

<krit> http://wiki.csswg.org/test

krit: in general, I think it would be great if we keep it as similar as possible

Tav: that's the plan

CM: I think we're half way there, repository is in the right structure

Tav: template is worked out

krit: only problem is to connect and commit

CM: I'll look into it
... looking a the contribute page, when you had it set up so you could commit, what authentication did yo use?

krit: sign up for wiki, then send email to Peter Lins,

CM: during this event, is the intention for people to get accounts and commit to the repository?

krit: yes

<scribe> ACTION: Cameron to look into getting repository working for test the web forward [recorded in http://www.w3.org/2012/05/24-svg-minutes.html#action01]

<trackbot> Created ACTION-3295 - Look into getting repository working for test the web forward [on Cameron McCormack - due 2012-05-31].

CM: is it 3 weeks away?

krit: 15-16 June

SVG 2 update

CM: what's the latest work that's been done?
... reminder that the plan is to publish FPWD in July, 2 months away
... I've been working on the painting chapter
... I see Erik has added buffered rendering

ED: I was wondering, a couple that I've signed up for require changes to the IDL, do we have something that parses web IDL or do we need to tweak the script?

CM: The problem is the IDL parser we're using is old. I'll have to update the parser or we'll lose some automation and have to write the IDL directly in the chapters
... at the moment, we get a lot of formatting done by the scripts. I'm not a fan of the formatting, I'd like to change it and bring the element definitions up in the chapter, rather than at the end
... I think we wouldn't lose too much automation if we included IDL in the pre elements
... manually
... in terms of other changes, I've been bringing property layouts in line with the css specs

<heycam> https://svgwg.org/svg2-draft/painting.html#StrokeShape

CM: it's a bit more algorithmic than usual, is that good or should we rewrite it a different way?

krit: the images are done with svg yeh?

CM: yes
... I did it with a dash array
... I mentioned on the mailing list, it's probably safe to use 1.1 features, but not filters or animations

krit: I'm fine with that

CM: what I might do is add links to the png renderings of the svg

nikos: we need a script where you can point to an svg and get a png

CM: use common sense and avoid things that we know aren't well supported

krit: can we convert dash array to a path?

Tav: inkscape can do that

krit: it would be a good idea for dash array so that it renders the same everywhere

CM: I've been using the coloured backgrounds when I've been reworking sections from the 1.1 text
... When it's good, I change it to yellow which means its ready for SVG WG to review

krit: It gets turned to white after review?

CM: when we publish, the group as a whole will sign off on the yellow and turn them to white
... It will be good if people look through at the yellow sections to get an idea whats in there
... most of the existing diagrams from the 1.1 spec have a thin blue border, I've been removing that and putting in the css instead

krit: I want to make sure equations have metadata

CM: I'm a bit concerned about the accessibility, the usual method with maths is to use 'mathplayer' which is ie only
... it would be good to know if there's a better solution

krit: in general we don't care how it gets displayed, as long as it can, in theory, be transformed

CM: I've used a hidden pre element with the content

<heycam> https://svgwg.org/svg2-draft/painting.html#ColorInterpolationProperty

CM: copy and paste the formula and you'll see the hidden markup
... I'm using an aria element
... I could have a hidden button that turns off the mathjax rendering

krit: be careful with hidden elements, they're not read or displayed

CM: with the annoations we're adding, should we remove them once we add the feature?

Tav: they have some historic value
... the idea is when they're published they're hidden
... 10 years from now someone might want to know why it was done that way

Referencing properties from other specs in SVG 2

CM: might not be much to talk about here
... in my email

http://www.w3.org/mid/4FB5A353.3050802@mcc.id.au

CM: in svg 1.1 there's a bunch of css properties that hav ebeen redefined
... we should reference existing definitions
... remove the tables and reference the css spec

<heycam> https://svgwg.org/svg2-draft/painting.html#VisibilityControl

CM: I've preserved most of the text, but reworded a little

Tav: that's good, you've got a short summary of what they do. If people want to know the nitty gritty they go to the other spec

CM: I'm not sure about the property links in the text, should it go to css?

nikos: no, should link to same document

Tav: I agree

CM: similarly for properties we've moved from svg to specs we depend on
... one issue with having properties moved out of svg spec is that we need to make sure the presentation attributes are allowed on all the elements
... one of the changes with 2nd edition 1.1 is for every element that's styleable is to allow all presentation attributes
... I think a hook in the main spec that other svg specs can link to would be good

<krit> http://dev.w3.org/csswg/css3-transforms/#svg-transform

krit: See sentence 'This specification will also introduce the new present...'

CM: the last sentence about parsing, I think the svg spec will need a couple of definitions like that
... other specs will need to specify if they allow extended syntax

<heycam> krit: css3-syntax spec will define syntax for presentation attributes

krit: CSS 3 syntax spec will define syntax for presentation attributes

shepazu: I've been asked to present on SVG at Test the web forward. Anyone interested in helping me?

Tav: I can

krit: I can do that
... I can talk about committing, we can share the presentation if you'd like
... however you want to organise it

shepazu: We are looking at an API for querying test results

krit: That's what shepherd is for

shepazu: the other thing I want to mention is, the event is useful because it's an examination of the idea of the community helping with tests. I'd like to crowdsource tests in the future

Removal of the Media Type Registration chapter from SVG 2

CM: I thought that because we have the media type registered now that we don't need the chapter
... I asked Chris
... He thinks we don't
... unless there are objections, I'll remove it
... it existed in the spec so the registry could point to something
... the media type won't mean anything difference between SVG 1 and SVG 2. not neccessary to resubmit

shepazu: It occurred to me that if we're aligning more closely with HTML, would it be useful to have another mime type for non XML SVG?
... A non XML SVG might be someone using SVG in HTML where they aren't quoting attributes, the HTML parser doesn't care.
... anything to do with HTML parsing can't be called XML

CM: I think there was interest in using an error correcting parser for XML
... maybe a community group started up

<heycam> http://lists.w3.org/Archives/Public/public-xml-er/

CM: not much activity since February

Do we still like solidColor?

CM: solidColour is something we've resolved to have in SVG 2 but I don't like it
... one of the reasons is that it's not that hard to have the same effect with the current elements

krit: even with changed behaviour?

tav: how would you do it in another way?

CM: until CSS variables are available use a single stop gradient
... slightly ugly but you don't have to put all the gradient attributes there

Tav: It would be so much simpler to have solid color
... suppose you have a drawing with the same colour but in 16 different places, it would be problematic

krit: you have a prsentation attribute and change it in one place

CM: my concern is that once variables is implemented everywhere people would prefer to use it over solidColor and then we'd have multiple ways of doing it, which I'd like to avoid
... I'm thinking from an authoring point of view
... If there's interest in keeping it I won't block it
... The reason I started thinking about it is because I was wondering what our plan in general is in regard to naming of attributes and elements
... whether we keep camel case
... for each camel case name we introduce, we need to update the implementations

Naming things going forward (camel case still in favour?)

[No strong opinions given on solidColour]

CM: Because of how the HTML parser works, elements and attributes are either lower case or case insensitive. There's a fix to transform to camel case for SVG.
... if we introduce new attributes with camel case, the implementations need updating

krit: I think that's fixed in DOM 4
... don't have to think about it from a spec point of view

CM: don't think so. if it's not registered it will become all lower case, and the DOM is case sensitive
... is it actually a problem to update the parser when adding the implementation for the new attribute?

Tav: camel case is painful. I wonder if adding lower case versions of the attributes that have camel case would be the way forward?

krit: what if you want one that's all uppercase?

shepazu: you wouldn't want to
... in HTML it doesn't matter
... for XML parsing, case does matter
... for HTML, it's case sensitive, but when it's parsed it's all lower case, so the DOM is all lowercase.
... if we made all lower case versions of the tokens, HTML, SVG and XML could be valid
... I know it's doubling the number of elements in the namespace, but the benefit would be an SVG that you could use case insensitively or lower case and it's all consistent

CM: in HTML parsing, if you put lower case and it becomes mixed case in the DOM, it's confusing

shepazu: I've seen lots of problems with viewBox
... I don't think Chris would like this
... but if others think there's merit we should consider it

CM: At the moment, what do we do with new things? Keep the existing scheme until we make a decision?
... I'm drawn to consistency

shepazu: with what? HTML or SVG?

CM: It's not just what people know, but what the other things in the same SVG fragment look like\

Tav: I agree

CM: it's the same sort of issue with CSS property values
... they use dash separated names if anything

shepazu: I think we should be consistent with HTML rather than CSS

CM: you could introduce a dashed version of camel case properties

<shepazu> sI think we should be consistent with HTML and CSS rather than SVG 1.1/I think we should be consistent with HTML and CSS rather than SVG 1.1/

shepazu: I used to think that consistency with SVG was more important, but now I find some of our conventions get in the way

CM: If we convert everything to the one true convention, then at that point we can decide how to name new things
... I don't want to be in the situation where we have a mix
... not something we need to decide now, but we should consider it in future
... the near future

Dropping contentStyleType and contentScriptType

CM: In SVG 1, these attributes are designed to specify the language used in the style and event attributes
... SVG 1 spec says they're deprecated, should we remove them?

shepazu: I think we can remove them

ED: I agree

shepazu: I think any solution (if necessary) will be mutual between SVG and HTML

CM: These started as HTTP headers, then became attributes in HTML
... they don't exist in HTML 5
... I understand the point that normally they're not used but because they are for extensibility they might be used privately
... Private use doesn't need something in the spec

shepazu: I've used them
... you could do something weird with script that you can't do with regular javascript

CM: Anybody against removing them?

Resolution: Remove contentStyleType and contentScriptType

<heycam> ACTION: cameron to investigate having a hidden button to turn mathjax rendering off [recorded in http://www.w3.org/2012/05/24-svg-minutes.html#action02]

<trackbot> Created ACTION-3296 - Investigate having a hidden button to turn mathjax rendering off [on Cameron McCormack - due 2012-05-31].

- DRAFT -

SVG Working Group Teleconference

24 May 2012

Attendees

Contents