<scribe> scribenick: kaz
Chris: We have 3 topics today: Bullet Chatting TF, Media production use cases, Web of Things.
<cpn> Slides
Chris: Anything else for today?
(none)
Chris: We've had a couple of
discussions during the monthly calls,
... in Jan, Aug, and TPAC during the MEIG meeting, breakout, and
TTWG as well.
... Building momentum over last few months, there's now a CG.
... The people involved have been quite active, there are two main
documents,
... a use case description, which was presented during the MEIG
call,
... and API proposal looking at implementation
considerations.
... In conversation with people involved, we thought it would be
worthwhile for the gap analysis and exploratory discussion to be
done within the MEIG.
... We can create a sub-group as a M&E IG Task Force.
... Want to get to a clearer description/understanding about what
is needed.
... Three main questions: Is a new data interchange format, e.g.,
TTML/WebVTT, or extensions to those formats?
... Do we need a client/server API for delivering the bullet
chatting content?
... And for presentation, what web APIs or changes to HTML are
needed for rendering and layout?
... We propose a short-term activity through the TF,
... and then further spec development by the Bullet Chatting
CG.
... Here within the MEIG we could do discussion for gap analysis by
broader participation, to help with the gap analysis.
Song: I can give a status update,
then try to answer those questions.
... Bullet chatting refers to comments that float over a video at a
specific time point in the video,
... it can bring a very interesting experience for viewers.
... Bullet chatting is very popular with video services like
Niconico, from DWANGO, who are in the CG and will contribute to the
TF.
... In China, there is Bilibili and AcFun, who are also CG
members,
... and others such as TenCent also provide this service.
... That's the reason to standardise the technology and make it
easy for developers across Asia and elsewhere.
... After TPAC 2019, we created the Bullet Chatting CG, where we're
working on the use cases and APIs.
... Chris, etc., also joined the first telco of the CG. We gave a
demo at TPAC in September.
... During the call in November, we discussed a plan for
standardisation.
... We plan to start with use case clarification, then generate
requirements, and find the gaps in existing standards such as from
TTWG and other APIs, as well as data formats and protocols.
... One thing we may want to standardise is CSS Animations with
synchronization, and another is a data format and client/server
API.
... And now we're planning to create a TF within the MEIG as we
thing this is the place where companies can exchange ideas beween
content providers, distributors, and device manufacturers.
... There are several devices manufacturers in the CG to work on
this topic. The goal for the TF is to make bullet chatting easier
to implement,
... and to apply it to broader applications. Chris described
mission for the TF.
... For the specific questions, feedback from our last meeting was
that a data format could be useful for more use cases, for things
not supported in HTML yet.
... Regarding a new API, we discussed applying an animation to a
media timeline, which could be useful.
... Regarding rendering and animations, we're collecting use cases,
e.g., 360 video and games. We think a solution like CSS animation
with synchronisation can work.
... That's as what I can say at the moment.
Chris: Any thoughts?
Kaz: We've generated some initial Charter for the expected TF, right?
Chris: Yes, there is also an issue on
GitHub as well, a CfC. Can we take the resolution today to work on
this?
... Then we can set up a wiki page for the TF.
... Anybody has any objection?
(none)
<scribe> ACTION: The MEIG would create a TF for the bullet chatting discussion.
Chris: I want to avoid confusion
betwee the TF work and the CG.
... I propose we use the existing CG repositories to capture the TF
outputs, and contribute to the existing documents, to keep
everything in one place.
... The MEIG will do the gap analysis, and the CG can develop the
APIs, which the MEIG is not chartered to do.
... Song, what are the next steps?
Song: We'd like colleagues from TTWG
to join a TF call,
... There is some overlap already identified between the TTWG specs
and bullet chatting, but some capababilities may not be supported
yet.
Pierre: TTWG has focused interchange
formats, not APIs.
... So it's critical to answer the question whether an interchange
format is needed. If not, there'd be reduced involvement from
TTWG.
Song: We have a draft exchange format, but it's not been defined as a standard format yet.
Pierre: Is that a priority?
Song: It's not the top priority yet, as we still need to clarify gaps.
Pierre: So the top priority is an API?
Song: Yes
Chris: TTWG input to help clarify the gaps around interchange formats would be great.
Pierre: If we know we need an
interchanging format, TTWG would be the right place, so could just
do the work there.
... But we'd need to know that from the Bullet chatting proponents,
it's more of a use case question.
<nigel> +1
Chris: We may still need some exploratory work to answer that.
Nigel: I agree with Pierre's
comments.
... If the priority is to do an API first, then working out how to
serialize the data needed for that is much easier after you have an
API that works.
... It could make sense to ensure the requirements are captured in
the API, then work out how to serialize that, then you have
something concrete for a gap analysis.
... Doing both at the same time may not be such a good idea, as
there is a risk of duplicating the requirements capture and getting
different results, particularly if there's different groups of
people.
Pierre: I would like to hear from the proponents of bullet chatting whether they think a data interchange format is needed. It's a critical question, and MEIG can help answer it.
Chris: Right, but not necessarily to answer right now.
Pierre: Definitely
Chris: That can be the first item for
the TF to discuss, and that sets the direction for the next
stage.
... I propose to schedule a call to do that, will follow up with
Song about that.
Pierre: At TPAC we had discussion on
media production use cases.
... It looked like there were a couple of people interested in the
topic, or had prepared similar input to the IG.
... Garrett started working on a gap document.
... What's the best approach for this topic for 2020?
... I'll summarise the background and problem statement. Just my
perpective for now, but should be a group excercise so open for
input.
... [Background]
... Professional media assets are increasingly being stored in the
cloud rather than on-premises servers.
... Once content is in the cloud, the next step is to build web
applications around those assets, to author, manipulate, quality
check, etc.
... The web platform has made great progress in the last 10 years
for media playback, professional authoring applications require
additional capabilities.
... They may be incremental capabilities, precise synchronization,
support for HDR and wide gamut color images, higher fidelity timed
text, etc.
... [Problem statement]
... What limitations of the web platform prevent the development of
professional media apps, and how should these limitations be
addressed?
... Any comments?
Garrett: I think this is a good idea,
collecting the use cases and gaps would be valuable. We as media
producers have seen the gaps, but I expect there are others we've
not thought of.
... Collecting the use cases and seeking input from publishers
would show it's a problem. It's a valuable thing for use to
have.
Pierre: any other input?
Barbara: One approach is fixing
existing problems that have been there for a long time,
... another is looking at emerging areas, e.g., streaming and
editing live sports.
... A high level question is whether you want to tackle existing
stable usages, or new and emerging use case?
Pierre: I would say all :)
Nigel: I think there is a clear need.
Implementers tell me there are problems that they solve in some
locally constrained way, but it's not clear what those problems
are.
... I would like us to gather information about the problems
... which make media production web applications difficult to
implement.
Chris: I agree with Nigel, but want
to add that the frame accurate seeking issue in GitHub,
... which talks about seeking and rendering aspects,
... generated a lot of interest. Not only from IG members but the
wider community, more than any other topic we've covered.
... So I think there is something here to follow up.
... [Media production use cases: Plans for 2020]
... https://github.com/w3c/media-and-entertainment/issues/4
GitHub issue (issue 4)
Pierre: It's not clear to me what all
the gaps are,
... maybe more importantly whether we have critical mass within the
MEIG, e.g., producers and people who develop professional media
applications.
... How to proceed? We'd want to poll external parties to draw a
clearer picture.
... We could invite input via GitHub, but not sure this is
productive for that community.
... Another option is maybe organizing a workshop, invite external
people.
... Simply organizing the workshop would provide useful information
on who's interested and help refine the problem statement.
... Does the concept of holding a workshop sounds interesting? Or
other approaches?
<Barbara_H> Gaming did a workshop and it was effective
Garrett: It's valuable to put
together a list of gaps, and holding a workshop would be good
figure out what's missing.
... How would you recommend we but get right people to attend?
Pierre: We have a lot of contacts, collectively. First step would be to define a scope for the workshop and each of us could informally talk to potential participants to gauge interest.
Chris: We've talked offline about a few ideas, whether it should be a virtual workshop or F2F.
Garrett: Possibly both so that people who can't participate in person can join the discussion.
Pierre: We could try, but it's hard in practice.
<Barbara_H> Target audience - Professional versus Casual?
Nigel: One option is to write some
kind of public communication to let people know,
... a W3C blog post that describes the problem area and invites
input in a more public way that we could do something, of we have
the right inputs.
Pierre: I like this idea. Does anyone think running a workshop is not a good idea?
(none)
<Barbara_H> Media is a pipeline - capture, production, distribution, consumption/rendering. Production is part of the pipeline
scribe: maybe the question to ask is whether the concept of having a workshop would be good or not
Chris: An alternative we talked about
was to invite some specific people to share input during our
monthly calls.
... It wouldn't be as productive as a workshop.
Barbara: Thinking of media as a
pipeline, production is part of that.
... The open question is whether we've met the needs of the
production audience.
... You talk about professional production, but there's a growing
area around amature production, e.g, for YouTube.
... Question of whether to go for the high-end or the broader
market.
... I agree that we need to understand the needs of the
audience.
... Some companies may want to move to the web but there may be
something holding them back.
... As an example, one compamy we work with couldn't move to the
web until new technologies like WASM came out.
Pierre: Professional vs non-professional is a good question, it's a grey area.
Chris: I like Nigel's suggestion,
having something written down that we can share is a good
approach.
... Unless we have everything covered among existing IG members, in
which case we could go ahead and do use case/gap analysis.
... I also agree with Garrett's point that seeding it with some
initial thoughts, so I think we should do both.
Pierre: Maybe we can combine that with our outreach messages, as a concrete example.
Chris: So the next step is to write something.
Garrett: Happy to contribute to that.
Kaz: I'm happy with this discussion, but before taking a resolution we should think about our needs, writing a concrete messages would be the next step.
Kaz: There was some discussion at
TPAC, we invited Chris to the WoT meeting to discuss possible
collaboration between MEIG and WoT.
... From a MEIG viewpoint, possible technical topics should include
event handling and time synchronization,
... and from a WoT viewpoint, TV as an edge device for IoT purposes
and video streaming data.
... It should proide a good collaborative discussion, so I propose
a joint call. There's a Doodle poll. It seems the best slot would
be Tuesday 4th Feb 2020.
W3C WoT Standardization Overview
Kaz: WoT's main purpose is
interconnecting the IoT silos, e.g., CoAP or MQTT or HTTP based
communication, lots of IoT standards and alliances.
... We attempt to connect these using web standards, e.g., the WoT
Thing Description. It's a simple JSON-LD based data model to be
used as metadata for IoT purposes.
... Devices advertise capabilities. We can talk about the details
during the joint call in February.
Chris: We need some of the right people, including device manufacturers?
Kaz: Yes, also NHK have been working
with WoT and have a Hybridcast based demo. We should include
them.
... If we hold the joint call on Feb. 4, we'll have many
participants including NHK, Siemens, Fujitsu, Intel, etc.
Pierre: We could extend the IG call to cover the other topics as well, e.g., starting 30 mins earlier.
Chris: OK, so we cover other IG
business as well.
... Any final thoughts for today?
Rob: I'm particularly interested in
the WoT idea, there's a crossover with WebVMT and the data sync
aspect.
... It's a proposal, but it's now in GitHub. I'm interested in
hearing comments about that.
<RobSmith> https://github.com/w3c/sdw/issues/1141
Rob: There may be some common ground with bullet chatting too, in terms of the CSS animation, also thinking about for WebVMT.
Kaz: Interesting.
Chris: Jan. 7th, will send out an
agenda.
... Thanks for joining!
[adjourned]