SVG Working Group Teleconference -- 03 Jul 2008

<trackbot> Date: 03 July 2008

<scribe> Scribe: anthony

<shepazu> ??P is me

test-fest

AE: I made the request to discuss this
... so we could have a test fest before SVG open
... time is running out for that
... wanted to see what the group thought about a test fest

CL: So I thought we decided not to do it
... because of difficulties with people traveling

AE: Wasn't sure, because on a call you could send someone

CL: Just wondering if it's a bit late to do one
... but I can see value in doing it

AE: It's all a bit connected
... because I thought we'd have the test suite published
... it's probably too late to have travel for sure
... so maybe there's another way of doing it if we are still keen
... even then there may not be time to do that
... because we have to send a request for implementations

CL: We need to give people advanced notice
... in some ways I'd rather we organise something later
... just to give people the notice they need
... first thing is to get the test suite published

AE: We also said there would be a Beta
... with a link with known issues
... the blocking item is simply putting it up

ED: So what's the hold up with the test suite?
... it's not published?

AE: Is it something that only a team or staff contact can do?

CL: Yes it is
... planning to fix that
... once it's in the public repository
... I guess I could ask Sys Team to give access

AE: I guess Doug is busy with a lot of other stuff
... if it was in the public repository
... might be able to be published

CL: Problem is when you check it into the new place
... the revision numbering is lost

AE: Would be great if we get the test suite published
... is there a ball park date for holding the test fest

ED: Ball park date would be after SVG Open?

CL: Yeah, after it
... would the test fest be a face-to-face meeting?

ED: I guess it would be useful

CL: September was mentioned
... but I have a management meeting 15 - 17 Sep
... something around that is ok
... either the beginning or end is fine with me

DS: We are having a F2F before SVG Open
... what are we doing there?

ED: We could have some time allocated for it

AE: We'd want this test fest to be the final one
... so it could take some time
... so at the F2F we should focus on getting the test suite

<scribe> ... done

ED: So it sounds like we should work on tests
... to get the suite done
... at the F2F

AE: If we release a Beta 2 some time soon
... we could work on the test suite

CL: Would be good to announce a Beta 3 at the F2F
... I'm think the 20th - 24th Oct is the TP
... one possibility is to organise the test fest near that
... not sure if that is too far in advance

AG: Not a bad idea

CL: We'd be talking about the week around the 13th Oct
... how long will we need?

ED: 4 - 5 days probably

AE: Are people good for going 2 weeks at a time

DS: I think being worn out before the TP might not be a good idea

CL: If we did the test fest Mon - Thur
... it would give us 3 days before TPac

AE: Is the date mid Oct too late for publication of the spec

CL: Whether is too late depends on how many pass we have
... at the test fest
... traditionally they produce their implementation report
... then request to go to PR
... [Explains process]
... if we were in PR in Dec that's fine with me
... it's just a notice to say we have 6 weeks to go

DS: We should really shoot for the end of the year

ED: In that case we should go with the end of Sep

AE: As CL said we may run the risk of not having 2 implementations for every feature
... that said it would nice to have a report sooner or later
... would be good task implementers where they are up to
... but maybe if a staff person asked it would be good
... so it sounds like Sep is probably the better time

ED: Do we want to decide on a location?
... or do we want to discuss it at the next F2F

AE: Might be a bit tight at the F2F

DS: I'm happy to host or Andrew can you host?

AE: I'd be happy to host

CL: What about the week of the 29th Sep?

DS: Anytime around then is fine with me

AE: 29th sounds ok
... we can decide on another call the exact location

ED: Sounds ok

RESOLUTION: We will hold a test fest on the 29th Sep to the 2nd Oct somewhere in North America

CL: We should announce that ASAP
... and say that further details are soon to follow

AE: We should have time line for registering it for it

<scribe> ACTION: Emmons to announce that we will he holding a test fest from the 29th Sep to 2nd Oct and that we will provide further details soon [recorded in http://www.w3.org/2008/07/03-svg-minutes.html#action01]

<trackbot> Created ACTION-2080 - Announce that we will he holding a test fest from the 29th Sep to 2nd Oct and that we will provide further details soon [on Andrew Emmons - due 2008-07-10].

<ed_> http://www.w3.org/2002/09/wbs/19480/svgF2Fnuremburg2008/

svginhtml

SVG Open

<ed_> http://www.w3.org/Graphics/SVG/WG/wiki/Nuremburg_F2F_2008

DS: I would like to have something to show at SVG Open to show progress

CL: Form is now open
... for registration

ED: Do we need to something out for the public list to get people to sign up for it?

AE: We should send a reminder to the public working group list

ED: Ok I can do that

SVG in HTML

ED: I did an update to the proposal

<ed_> http://dev.w3.org/SVG/proposals/svg-html/svg-html-proposal.html

ED: you can still see the old proposal

<ed_> http://dev.w3.org/cvsweb/SVG/proposals/svg-html/svg-html-proposal.html

AE: Looks like we've taken the approach of what needs to change in HTML5

<ed_> http://dev.w3.org/cvsweb/~checkout~/SVG/proposals/svg-html/svg-html-proposal.html?rev=1.2&content-type=text/html;%20charset=iso-8859-1

ED: It's using the XML parser
... basically it's just a black box
... that says how to handle stuff the HTML parser doesn't understand

AE: Should we have a highlevel note for that?

CL: It would be good to list the requirements at the top of the proposal
... such as round trip copy and past of SVG
... to and from the document

DS: We're not trying to HTML5

ED: It would be good if people could make up dates to this
... is this a subset of SVG?
... No you can paste any SVG fragment
... there are somethings you can't do because HTML doesn't support name spaces

DS: So this does not introduce name spaces to HTML5?

ED: It does not

DS: So if have some output from Inkscape and it has all the extra metadata how does it handle that?

ED: That will still work
... it will still be put into the DOM

DS: Except it will not be name spaced

ED: It will be name spaced

AE: It's adding it to the DOM
... but wouldn't the DOM need that anyway for the Xlink serialisation

CL: The crucial thing for this is that once you've identified
... root node that you want to put in
... that all the other name spaces within it are preserved
... so we are not trying to put in name spaces into HTML5 but preserve the name spacing
... for content within the SVG element
... should have some examples

ED: I think this is not a bad start
... but it's not complete

DS: So Charles was saying that you could use foreignObject and use a switch with a required feature

ED: Problem is malformed content

CM: You could put the foreignObject at the end of the SVG
... there might be some other tricky things with Font and TextArea

CL: I think it would be useful to document this
... and Doug's fallback idea
... and give pros and cons of both

DS: So font shouldn't show up right

CM: So if we mandated elsewhere that default style sheets have that on it
... so if these things get parsed and have name spaces set properly
... then you could use that to determine where it came from

DS: I would really like to see us present this in a week or two

ED: Would be good to flesh some of this out

DS: Are you pretty confident with this algorithm?

ED: I think it will work
... What's left is to provide examples
... and authoring guidelines

DS: How about we try to have this in place by next Thursday?

ED: Sounds good
... I guess I can try to fit in some examples
... it's a good idea to keep it in public CVS and work on it there

<scribe> ACTION: Erik to incorporate what we discussed into the the SVG-in-HTML proposal [recorded in http://www.w3.org/2008/07/03-svg-minutes.html#action02]

<trackbot> Created ACTION-2081 - Incorporate what we discussed into the the SVG-in-HTML proposal [on Erik Dahlström - due 2008-07-10].

XHTML access module

ED: Haven't had time to read it

CL: I haven't had a careful read though it

DS: I'll send in my second review to this list

<ed_> http://lists.w3.org/Archives/Public/public-svg-wg/2008AprJun/0168.html

<ed_> http://lists.w3.org/Archives/Public/public-svg-wg/2008AprJun/0153.html

Issue 2004

<heycam> ISSUE-2004?

<trackbot> ISSUE-2004 -- Define SVGMatrix behavior when using float edge cases -- RAISED

<trackbot> http://www.w3.org/Graphics/SVG/WG/track/issues/2004

ED: Cameron had an action to look at this

CM: I did
... the result was that mail
... there is no need for infinity and NAN values
... so we can define that at the top of the chapter

ED: I have not issues with your findings
... looks fine to me

CM: Not sure on the type of exception that should be thrown

<heycam> http://www.w3.org/2008/06/19-svg-minutes.html#item09

CM: when one of these values is passed in

AE: I think at one point there was wording about
... what exceptions was thrown
... I know that wording was in the traits

ED: It's in A3.1
... [Reads out part]
... although it doesn't say anything about infinity

CM: We could add wording in that said

<heycam> CM: If we could replace that "signalling NaN" text with something that says if an infinity or NaN is passed in as a float, then a certain exception is thrown

ED: I would prefer not to add the wording into A3.1
... because it says "attribute property and normalisation"....
... but I guess it's ok
... if you consider a normalisation a call to any method

<ed_> s/attribute property and normalisation/attribute/property normalisation/

AG: Result?
... add that wording?

ED: Yes we should add that wording

CM: What do implementations do at the moment for not supported error?

AE: I think that with the trait tests
... we do what the current spec says
... I think it might be non-trivial to change that
... since it's in JSR226
... and that's already in the field

CM: So I guess if someone can confirm that in 226

AE: Yes I can confirm that

CM: In that case we should chose that
... there was something slightly related that I mentioned in the email
... and that was for negative zeros

<ed_> javascript:alert(document.createElementNS("http://www.w3.org/2000/svg", "svg").setCurrentTime(parseFloat("asdasd")))

ED: It seems that Opera is doing the not supported error
... so this will give you the not supported error in the console

CM: So would you be willing to change the other one
... to type mismatch

AE: The type mismatch is specifically dealing with NAN
... it's in the context of
... you're putting in what could be a float value
... but for the one you posted in that is not supported

ED: For me the changes in either one of them is very limited
... but I'd prefer if it were similar everywhere
... so do we want to put this text into the spec

<ed_> javascript:alert(document.createElementNS("http://www.w3.org/2000/svg", "svg").setFloatTrait("x",parseFloat("asdasd")))

CM: throwing type miss match error for set float trait
... and then other for the other cases

<scribe> ACTION: Cameron to fix exceptions in the specification for setFloat values [recorded in http://www.w3.org/2008/07/03-svg-minutes.html#action03]

<trackbot> Created ACTION-2082 - Fix exceptions in the specification for setFloat values [on Cameron McCormack - due 2008-07-10].

- DRAFT -

SVG Working Group Teleconference

03 Jul 2008

Attendees

Contents