SVG Working Group Teleconference -- 20 May 2008

<trackbot-ng> Date: 20 May 2008

Adobe, SVG and eBooks

<ChrisL> http://www.openebook.org/oebps/oebps1.2/download/oeb12-xhtml.htm

<ChrisL> http://www.openebook.org/oebps/oebps1.2/download/oeb12-xhtml.htm .

<ChrisL> http://markmail.org/message/mxcmubdgh7xzkccb

<ed> http://www.idpf.org/2007/ops/OPS_2.0_final_spec.html

<ed> "OPS Reading Systems must support SVG (Scalable Vector Graphics) as an OPS Core Media Type."

<ChrisL> trackbot-ng, status

<ChrisL> trackbot-ng, action-1?

<ChrisL> trackbot-ng, action-1

1.2T test suite sprint

<ed> http://dev.opera.com/articles/view/a-call-for-video-on-the-web-opera-vid/

<scribe> scribe: anthony

Circular references in use

DS: prepared a test case
... [Describes test case]
... clearly a circular reference
... do we want this to be an error?
... or do we want this to be invalid content...

ED: invalid IRI is what I was thinking
... currently says in the linking table that it's an error

AE: You can still proceed with rendering the rect?

DS: No it's sufficiently unclear

AE: We have test already with transparency

<ed> http://www.w3.org/Graphics/SVG/Group/repository/spec/mobile/1.2/1.2NG/publish/linking.html#ReferenceRestrictions

AE: and the different levels of transparency indicate how many times you've gone through

DS: What happens when we encounter this case?
... there are three interpretations for this

LR: The certain graphic operations where you can encounter an error
... this operation shouldn't affect other operations

DS: So this would happen with incorrect attribute values
... in 1.1 we would've said stop rendering if you find an error
... in 1.2 if you find an error in the value you continue on

TZ: Do you display the error on the screen?

<aemmons> http://www.w3.org/Graphics/SVG/Group/repository/spec/mobile/1.2/1.2NG/publish/implnote.html#ErrorProcessing

AE: I think the error processing model we have we should be rendering the document

ED: So the IRI has some specific handling?

TZ: If you were to serialise the DOM back do you get the same DOM?
... if you have DOM element already created how do you know it's an error?
... have to have a way to show the user that it's an error

DS: We can provide an indication of an error
... [Gives an example of authoring tool and viewer errors displayed]
... How debugging is handled doesn't effect the user experience
... it doesn't affect how the user views it

LR: Pose this from a different stand point. There is no way to put in a script that would validates onload
... this is something that author would want

DS: That is something would belong in the Window interface

AE: JSR226 has very strict error handling
... I can see the point that you load this content and there is an exception
... it would be useful to know which element
... but I'm hesitant because it increases the footprint

LR: That couldn't be embedded in the SVG?

DS: It would be in a script
... and done in the Window interface

TZ: If I was just to hand write content and put it in Opera
... I would hate for errors to be ignored
... when developing the content

ED: There is an error console and a debugging tool

DS: What if we said in C2 and C3 a UA that encounters unsupported content or an error should provide an indication to authors what the error was
... you may not want for security reasons the users to get at your files in non-debugging cases

TZ: There is no access to the DOM

AE: A way forward would be the Window interface

DS: Should apply to all HTML etc

<scribe> ACTION: Doug to adding wording about a UA providing where the error is and a reason for the error in a debugging environment [recorded in http://www.w3.org/2008/05/20-svg-irc]

<trackbot-ng> Created ACTION-2022 - Adding wording about a UA providing where the error is and a reason for the error in a debugging environment [on Doug Schepers - due 2008-05-27].

AE: So about the linking
... It should an unsupported attribute value

<ed> http://www.w3.org/Graphics/SVG/Group/repository/spec/mobile/1.2/1.2NG/publish/struct.html#UseElement

ED: [Reads out use element]

DS: Those should be reversed

CL: Would make more sense to word it that way

DS: We should say circular refs and other things the first circular ref will be treated as if it were an unsupported value

CL: Don't want to say it becomes
... want to say it's treated as

AE: Need clarification about as soon it is detected

DS: We should say something about Lacuna values
... is that what we want to happen?
... if there is an unsupported attribute value?
... I was going to put that in C2
... by putting it in here we cover everything

AE: There may be some specifics for use

<aemmons> ACTION: Doug to modify linking reference regarding circular reference. When a circular reference is detected it will be treated as an unsupported attribute value [recorded in http://www.w3.org/2008/05/20-svg-irc]

<trackbot-ng> Created ACTION-2023 - Modify linking reference regarding circular reference. When a circular reference is detected it will be treated as an unsupported attribute value [on Doug Schepers - due 2008-05-27].

<aemmons> ACTION, Doug to update C.2, point 2, that since it's 'as if the value has not been specified', that it will be as if the lacuna value was used

<ChrisL> since we already say in the spec that an invalid value is treated as if not specified

<ChrisL> ... and thus, we can state that it renders as if the value was not specified

<ChrisL> ... which is the same as the lacuna value

<ChrisL> .. and also makes it clearf that the dom does not change (the invalid value is still in the dom)

DS: If you look at my original example you see that I'm referencing a group
... at what point does it detect there is a circular reference?

CL: Need to be careful there
... depends on how linking is done

DS: We just agreed that circular references are not allowed

AE: At the first point of detection
... but the first point of detection is implementation specific

DS: We should either say something about it or we shouldn't

AE: We do say something
... we don't allow circular references and you should stop at the first point of detection

DS: In my example the use references group 2
... it doesn't render the use of group 2 but does the use of group 1 include the green circle?

AE: I'd say yes

DS: That's not what we say

ED: I think for use we would detect it when we get to group 2 the 2nd time

DS: The action is it doesn't render the action itself is we say an unsupported value

CL: There is no interoperable rendering of that
... you could go through and figure out that you had circular references

AE: I don't think you can get interoperability on this

CL: This is known to cause issues don't do it

AE: So do you test for this?

CL: We should say if you are putting this in then you'll get varying output
... I propose that we put something in
... in D.3.1

<ChrisL> Put it in D.3.1 Conforming SVG Document Fragments so that content with circular references is not conformant

<ChrisL> That makes it clear not to rely on particular rendering of it, sinjce the point at which circularity is detected can vary

<scribe> ACTION: doug to add wording to the specification that content with circular references is not conformant [recorded in http://www.w3.org/2008/05/20-svg-irc]

<trackbot-ng> Created ACTION-2024 - Add wording to the specification that content with circular references is not conformant [on Doug Schepers - due 2008-05-27].

TZ: Coming from industry I don't think this is a robust way to deal with this

ED: We could put something in the appendix for editors to give a warning

TZ: To me it is more logical to set something to (flag) say it is an error

DS: The wording already provides a way to meet your requirements

<ChrisL> http://www.adobe.com/products/digitaleditions/

<ChrisL> http://www.balisage.net/Versioning/index.html

<shepazu> rssagent, make minutes

- DRAFT -

SVG Working Group Teleconference

20 May 2008

Attendees

Contents

Adobe, SVG and eBooks

1.2T test suite sprint

Circular references in use

Summary of Action Items