<trackbot> Date: 11 September 2009

<ChrisL> zaki, +39 is me

<glazou> hi Bert

<scribe> scribe: ddahl

<scribe> meeting: HCG

Extensibility in HTML

<ChrisL> http://lists.w3.org/Archives/Member/w3c-html-cg/2009JulSep/0073.html

chris: skip action items, and accessibility workshop, move straight to HTML Extensibility

<ChrisL> http://lists.w3.org/Archives/Member/w3c-html-cg/2009JulSep/att-0075/Distributed_Extensibility.pdf

tony: slides are actually on "distributed extensibility"
... idea is to allow third parties to create markup that could be used in HTML

(tony presents slides)

adrian: Dojo framework has support for incorporating widgets into HTML
... standardizing this approach would make it easier to create authoring tools
... coming up with a proposal to HTML
... today will talk about what IE does now, issues, proposal will be related to but not identical to what IE does now
... would like to capture thoughts and feedback now

tony: what IE does looks somewhat like XML namespaces
... IE also uses this to allow binary extensions as well
... in HTML today an unknown element would not be processed by the browser

adrian: this can be used to put data into the page that might be processed by a script, or by a browser addin
... typically within line of business applications

steven: on slide 5, "org.my" is a URL?

adrian: no, just an identifier

<Liam> [so, "namespace" identifier not constrained to be a URI

<ChrisL> I'm hearing that its a string, can be a URI or a URN

<Liam> [and known html elements are treated specially]

tony: differences from XML, here the default ns declaration is ignored. this is because some sites have already incorporated default ns's into HTML, intended to be interpreted as HTML

chris: people intending to extend HTML must not conflict with existing elements

<ChrisL> so authors need an internal list of all html element names, to avoid

<Steven> nor with future elements

steven: or could be problems if HTML adds elements

<Bert> (Only if you accept that proprietary extensions should be allowed...)

adrian: this is current IE behavior, we're not saying that it's necessarily right. there is a lot of existing markup that does this

chris: is it still the case in IE that there are predeclared ns's

adrian: those are always declared

chris: are these declarations scoped or are they assumed to be global

tony: the prefixes currently are global, can only be declared once, on root element

adrian: this is an implementation restriction, not necessarily the best approach

chris: what would other browsers do in HTML? Opera, Safari?

tony: Opera supported ns's up to version 7, but don't think that other browsers supported ns's

sam: some resistance to supporting ns's in HTML

tony: ns's don't apply to attributes in IE, ns is just part of attribute's name

adrian: in today's implementation, you can't nest default ns's
... this is just a restriction on IE's implementation

<Zakim> Steven, you wanted to ask about future additions

chris: which features are well-motivated and which just didn't have time to do

tony: some restrictions had to do with compatibility

adrian: most restrictions were just due to not having time
... e.g., not nesting
... different treatment of attributes and elements

liam: recent discussion in XML community on ns's
... wrote proposal on "unobtrusive namespaces"
... if you implement ns's, hope that you would implement XML ns's

<Liam> http://lists.xml.org/archives/xml-dev/200909/msg00068.html

adrian: one issue around SVG and ns's being discussed in HTML is that we're talking about elements embedded in HTML
... or might behave differently within an SVG document

steven: from RDFa experience, it's not too much of a problem for elements not to be in a namespace

<rubys> http://dev.w3.org/html5/spec/syntax.html#coercing-an-html-dom-into-an-infoset

<Zakim> rubys, you wanted to discuss section 9.2.7 in HTML5

sam: explains rule
... colon isn't allowed in an element in XML, so has to be converted

<Zakim> Liam, you wanted to discuss well-formed

chris: haven't discussed whether ns's are the appropriate extensibility mechanism

liam: may be flexibility from the XML community, rules about ns's well-formedness are a bit fuzzy, XML Core WG might post a clarification if helpful

chris: changing both XML and HTML is a possibility

<Zakim> Bert, you wanted to suggest vendor prefixes

<rubys> tool for viewing how browsers currently work: http://software.hixie.ch/utilities/js/live-dom-viewer/

<rubys> tool for viewing how html5 is intended to work: http://livedom.validator.nu/

bert: you can't get full ns support in HTML, so could you do something like what we did in CSS, like use names starting with "-" for extensions

<Liam> [Bert's suggestion is ilke the "pragmatic namespaces" proposal (find it with google) to use element names like <org.w3.xforms.input> ]

steven: JavaScript libraries are starting to use XML to extend HTML, JavaScript mutates that into something that works in the browser
... could we standardize on something like that?

janina: ARIA is host-language neutral, including HTML, so have preference for incorporating HTML as a standard extension

<rubys> http://livedom.validator.nu/

sam: if you get Firefox you can turn on Henry Sivonen parser in preferences
... in the future colon will be mapped into that weird sequence, this will affect a number of scripts and tools

plh: browsers are not dealing with namespaces?

sam: correct, only way to get ns's now is to use XHTM

chris: where did escaping of colon come from?

<paulc> I have to depart early today.

sam: discussion in the group

chris: where are we? we have a description of what's currently done, have some links to suggestions. it would be useful to find out what different libraries do. what else?

paul: they have an action to take something to the HTML WG by the end of September

chris: what discussion or actions here would be helpful?

paul: when the extensibility discussion starts, HTML WG chairs will make sure that HCG is aware of start of discussion.

liam: some kind of ns proposal from the XML community will be sent to the HTML WG
... also a proposal about ns's in XPath

paul: there's a proposal to have a joint meeting at TPAC to discuss this, need to make sure this is actively planned

chris: any other comments? adjourn

rrsagen, format logs

This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/S??/Sivonen/
Found Scribe: ddahl
Inferring ScribeNick: ddahl
Default Present: +1.345.1.aaaa, Debbie_Dahl, wiecha, glazou, Kazuyuki, Sam, +39.524.9.aabb, raphael, Chris_Lilley, adrianba, tross, Steven, Janina, [Microsoft], Lofton_Henderson, Bert, Liam, Art_Barstow, Plh, Cooper
Present: +1.345.1.aaaa Debbie_Dahl wiecha glazou Kazuyuki Sam +39.524.9.aabb raphael Chris_Lilley adrianba tross Steven Janina [Microsoft] Lofton_Henderson Bert Liam Art_Barstow Plh Cooper
Regrets: JohnB
Found Date: 11 Sep 2009
Guessing minutes URL: http://www.w3.org/2009/09/11-hcg-minutes.html
People with action items: