<atsushi> cwilso, that's it, I think
<cwilso> scribenick: Leonard
<scribe> Agenda: https://github.com/immersive-web/administrivia/blob/master/meetings/wg/2019-10-22-Immersive_Web_Working_Group_Teleconference-agenda.md
<cwilso> https://github.com/immersive-web/webxr-ar-module/issues/28#issuecomment-540811746
Brandon: How to treat immersive
AR
... browsers are shipping with immersive VR
... What are the top-level things that can be asked for at
top-level. What sensors (especially with VR != AR)
... what is the text that is put on the "button" to initiaite
the XR session.
Leonard: What is the definition of "immersive"
Brandon: "Immersive" means ... not defined in the spec. In this case "immerisve" means going to the headset, tracking, etc.
Headset ==> rendering. Phone ==> full screen with split screen (a la cardboard)
<Zakim> klausw, you wanted to say currently, handheld AR is considered immersive-ar
Leonard: All encompasing viewing
Klaus: b: Additive blend mode
==> what does that mean.
... a: <missed the question>
Chris: Deeper question on "immerisve" to be done off-line
<Manishearth> https://github.com/immersive-web/webxr-ar-module/pull/38 + https://github.com/immersive-web/webxr/pull/869
Brandon: Push AR enum into VR
document because enum definitions must be complete
... discussion between screen/phone/headset
<cwilso> https://github.com/immersive-web/webxr-ar-module/issues/29
<cwilso> https://github.com/immersive-web/webxr-ar-module/issues/29
<cwilso> https://github.com/immersive-web/webxr-ar-module/issues/9
Manish: See above links for
discussion points
... Handheld AR allows phone orientatiuon tracking. Is this a
different kind of AR? If same, then can
tailor content. If different, then headset and phone content is different. It's important to be able to distinguish
scribe: between the two. E.g.,
DOM overlay on screen, world space on headset (see #9
above).
... Include type of AR session as part of mode. If not, should
DOM overlay be a pseudo-requirement?
<Zakim> bajones, you wanted to channel Blair and Rik
Brandon: Allow developers to do
content for the platform that works best. How to communicate
DOM overlay capability
... What level/when should that decision be made. Phone AR or
Headset AR? Developer to test and generate appropriate code
paths
... OR does the UA provide an interface layer to handle the
differences.
... [channeling Blair and Rik] best allow developers to do
content and not devices
... argues for single AR mode
Nell: Multiple points. 1:
Principal of API designs are balancing developers for specific
content & UA implement features for experiences for their
users
... e.g., Reference spaces. UA only obligation to support only
basic level. Developers can do lots of more expansive
things.
... Is the functionality based on screen-space UI? Developer
choice, but the UA provides the underlying support.
... DOM overlay API supports HTML, CSS, and WebGL canvas
(non-limiting list)
<Zakim> klausw, you wanted to say distinguish use cases?
Nell: if there is no DOM overlay API, does immerisve AR on handheld exist and does a headset ...
Klaus: In favor of exposing, but
targeted to what developers need to do. Concerned about
limiting developers. Synchronization with DOM ...
... and 3D display. Should there be a set-aside/dedicated
screen space (or similiar)
... split between hand-held & immersive: Needs to go back
to what developers might want to do in the future
Nell: Meant to require API shape. Reviewing Q: For HH, would not having an DOM overlay prevent useful content?
Klaus: Doesn't think that all content requires DOM overlay.
Manish: If DOM is requiremed, then still expose the type of overlay (real, emulated, no support)
<Zakim> cwilso, you wanted to answer Nell's question from historical data
Chris: Historical data (1+ year), HH AR mostly requiremed DOM overlay -- primarily for text, but others too.
Nell: Clairify one concern with
no overlay. Build experiences that target HHAR and draw stuff
right into buffer. Not possible for
... UA to handle when the UA is a headset. Create a pit of
failure. Shouldn't depend on flag because it gets incorrectly
used.
... Wants to see an API for DOM
Chris&Leonard: To Nell, please review the previous to make sure it is correct.
<klausw> specific pit of failure: content assumes GL viewport == screen rectangle, and draws directly into it. This will look awful on a headset.
Chris: Looking for comments from Microsoft
<klausw> Nell: content also should not make assumptions based on checking that "number of views equals one" implies handheld AR
Nell: Create strawbody to draft
shape of API to avoid rehashing conversation
... Suggest Manish to create strawbody API because thinks he
has the best understanding of all sides in screen vs. head
space
... What would the overlay API look like
Manish: Agrees
... Finish enum move sooner rather than later. HH decision
should not impact the enum "much"
https://github.com/immersive-web/webxr-gamepads-module/issues/18
Nell: Have PR and is super-urgent because it is a potential breaking change. This behavior is contrary to behavior of Gampad API.
<Zakim> bajones, you wanted to talk about the latest movement from the gamepad group
Nell: Difference is mostly in the "liveness" of the returned objects. Please Review #18 with PR #22. Requests Chris to get Gamepad people
<bajones> https://github.com/w3c/gamepad/issues/8
Nell: to see this
Brandon: See Issue #8 and
multiple other conversations, Gamepad should actually move to
live updates. Listed changes and
... do additional work to get there. CHange spec text to match
implementions in FF and Chrome.
Nell: No further discussion. Just do the above work.
<Manishearth> https://github.com/immersive-web/webxr-input-profiles/issues/82
Nell: Look at #82 and associated PR. Review/question before next call.
<Manishearth> https://github.com/immersive-web/webxr-input-profiles/issues/79
Nell: #79: No call discussion,
but read and review. If you have the fallback profile array, is
it a requirement that the order is an subset/inclusion.
... The PR will finish by end of month.
<Manishearth> https://github.com/immersive-web/webxr-input-profiles/issues/14
Nell: Moving on to #14... Problem
is sort of a mess. What happens when threre is not a clear
answer about the gamepad.
... Perhaps the device is completely new to the UA, but may be
compatible with another existing example.
... May also be a UA instruction to prevent/avoid
fingerprinting. What about squeeze events?
Need to finish this work this month.
Chris: Will ask Trevor to include this as a topic next week.
Nell: Wants to make sure WG people who care know that it will be discussed.
Chris: Review the Help Wanted section at end of the agenda. These are items that need some work.
Adjourned.
This is scribe.perl Revision: 1.154 of Date: 2018/09/25 16:35:56 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Succeeded: s/viewports/views/ Present: cwilso Leonard Manishearth bajones ada kip NellWaliczek Found ScribeNick: Leonard Inferring Scribes: Leonard Agenda: https://github.com/immersive-web/administrivia/blob/master/meetings/wg/2019-10-22-Immersive_Web_Working_Group_Teleconference-agenda.md Found Date: 22 Oct 2019 People with action items: WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]