Liaison with OIPF

Opent IPTV Forum

giuseppe: open networks and managed network deployments
... how to properly reference mostly WIP sources
... How have other organizations addressed this problem
... They are asking us for advice
... a) reference specific version with date
... Means regular review and update. This is a bit messy
... b) reference current draft without a date
... need to define a way of working to minimize changes
... c) only reference proposed recommendation
... d) delay any work until specs are at proposed recommendations level

<kaz> liaison letter (member-only)

giuseppe: e) reference latest version and decide which version to use later
... f) define some profile that can be referenced directly and keep track of change
... g) automatically use latest version available at time of publication
... h) any other option?
... Looking at 2D transforms, 3D transforms, HTML5 canvas, CSS changes
... Good to get feedback firsthand in this group
... want to also get this out to other groups (e.g. process group)
... floor open for comments
... Kaz mentioned this problem with other specs.

<kaz> EPUB 3

Kaz: There is a similar case in EPUB 3
... They used undated URL
... This is a working draft specification
... They need to change their URL if this changes

<kaz> Geolocation spec

Kaz: Geolocation is another examples
... Even though HTML5 spec is not stable, they will add note within spec
... The HTML5 spec will let them know when features are stable
... Geolocation also refers to WEB IDL
... They need to change the reference to WEB IDL - which has changed a bit
... Each section might change

giuseppe: one option is to use the section name rather than number

kaz: it requires HTML5 or CSS3 editors to use name rather than number references

giuseppe: glen on queue

<giuseppe> ack

glen: a couple comments
... the official language on status of these documents specifies inappropriate to refer to these other than as WIP
... need to put WIP somewhere in reference
... in the case of EPUB, the URLs are generic TR+shortname reference that would not be required to change
... would not anticipate process committee changing this requirement
... before doc is in CR stage it is questionable whether it should be referenced at all
... unlikely to get consensus on what "stable" means.
... I would expect significant changes to HTML before PR status

giuseppe: People have different opinions, but useful to collect examples of how others are handling the problem
... we can collect a few of the best options.
... could probably reduce the scope to a few possiblities

glenn: I would also follow up
... in a recent interview with Jaffe
... changes to accelerate the standards process
... they should be looking at this problem and formulating policy
... we should pass any input we collect on to them

<corvoysier> http://news.cnet.com/8301-30685_3-57390399-264/jeff-jaffe-lights-a-fire-under-web-standardization/

giuseppe: it's going through us because the liaison was established through us

glenn: I'm working with one standards group that decided to indirectly reference them through html5 specification
... This in turn has references to other specs
... this reduced the number of open ended references

giuseppe: Will be an issue solved for basic DOM and some other areas.
... any other opinions?

jan: Really appreciate the insight that drafts should be consistently referenced
... That will help external bodies do references

<glenn> link on Jeff Jaffe discussing accelerating process et al

<glenn>

giuseppe: We can run this through w3c management again before responding to OIPF
... I will draft first letter to them

kaz: thanks for the proposal
... suggest we include PLH and Ralph

thx

giuseppe: people can reply on the list if they agree with my conclusions
... next point

Profiling criteria

giuseppe: only thing we have agreed on is the name
... I created the online document

<kaz> Giuseppe's proposal

giuseppe: it was interesting to get feedback on what should be included in Profiling docuemnt
... how do we address specs with overlapping topics
... Should we include profiles for different conditions
... Any feedback on any of these points?

glenn: if we try to consider individual features, we will never reach closure

giuseppe: probably one way forward is to have some consideration inside specs on whether they are stable (especially important if various features are stable and not stage)

glenn: try to create a fuzzy document that is more generic in areas of technology defined by W3C
... should divide the idea of a profile into generic (abstract) and concrete
... concrete may specify individual features
... abstract can specify more general featues
... may be useful to try to get consensus on an abstract profile

giuseppe: I suggested "meta" profile. Maybe that is the same idea

<glenn> if we can't reach a consensus on an abstract profile, we'll never conclude on a concrete profile

giuseppe: any other opinions on this?
... for the reference itself, I used a TR version (current approach). Other options include editor draft

jan: if we have abstract list of requirements, how do we deal with that over time?
... is this a continuously updating reference?

glenn: I think thats TBD
... if we define an abstract profile then we could make reference to abstract pages.
... w3.org/TR/CSS has "CSS Snapshot 2010" which references CSS levels 1 through 3 and other items
... premature to make conclusion on that yet.

jan: maybe can let it progress
... then see what is needed. freeze when group decides it is appropriate

giuseppe: 1st step is to try to delimit which technology to use
... feature testing and performance testing are two important items
... 1) determine if parts are missing
... 2) test performance
... This is just getting started in W3C. What are best practices?

giuseppe: could be in scope for profile work
... another area that is covered is what to do about areas where HTML5 is providing technology, but relying on external specification
... should we undertake this topic (e.g. reference to metadata defined elsewhere)
... trying to understand what we want to put in this profile

corvoysier: ... anything specific?

giuseppe: in-band tracks is an example
... details of what is in-band depends on several factors defined elsewhere
... same information can be exposed differently depending on geographic location or transport method
... i don't have a concrete proposal yet
... people need to be aware of the functionality
... there are different levels we can reach. We should talk to these other organizations.

corvoysier: will need liaisons

giuseppe: I'm talking about work that may not be done yet
... I will provide more examples later online
... looking at the agenda again...

glenn: do we have a conclusion on what to call a profile

resolution: title will be Web Media Profile

giuseppe will post on wiki

<giuseppe> sibtitles: Guidelines for integration of interactive media services in a browser-based environment

<kaz> draft document

CSS specs we may want to reference

<kaz> draft document

glenn: which list are you referring to?

<giuseppe> http://dvcs.w3.org/hg/webtv/raw-file/tip/tvprofile/tv.html#css

giuseppe: now just section headings and link.

glenn: I would not agree to including this particular list
... in notes it should say that the contents of this section and which specifications are mandatory is under discussion and TBD
... I'm not sure this is the final list to be supported
... just to prevent misunderstanding

giuesppe: I will put in a note
... want to make sure it is relevant for that industry area

glenn: CSS properties only lists CSS 2.0 and not 2.1, Box model not being actively developed, etc.

giuseppe: I am working on an update as well
... to kaz, do we have an update for bugzilla account?

kaz: Looks like it has now been created, I will pass on

glenn: CSS3 referring to high-end typesetting features that may not be relevant. View mode? most everything needs a lot of review and discussion.

giuseppe: we have to pinpoint which specs we're interested in

glenn: do we need any references to SMIL?

giuseppe: put links in to provoke reaction so people can react on why they think is relevant

glenn: should support HTML5 and syntax, CSS, DOM4, - that sort of high-level abstraction that may not require mapping to specific spec.
... raises question: why do it unless it is useful
... need to choose balance points to make useful

giuseppe: 1st step at specification level
... need to know what to implement and support
... I'm open to input and will try to go through this
... participate in list to make your comments known

kaz: another topic
... got comment on how to handle OIPF reference
... has letter from OIPF
... They know of this reference problem and can discuss on mailing list

<inserted> kaz: so I'll forward the liaison letter to the liaison ML, and ask W3C Team mates for opinions

use case functionality

giuseppe: use the wiki or mailing list to provide your input
... suggestion to do survey of use cases and consider coverage of recent web specs
... we can do this and use our web space

next call

giuseppe: will continue with monthly call (first monday in general)
... there is call info on wiki
... I will send reminder
... any other business?
... closing call... see you next month

[ adjourned ]