This is a first stab at organizing the breakout sessions that have been proposed so far - please get in touch with public-tpac-unconference@w3.org if you identify problems with the proposed schedule. See also how to help building the unconference agenda.
Argos D-F, Floor 1
Navis-A, Floor 1
Navis-B, Floor 1
Navis-C, Floor 1
Kashi, Floor 1
Sumire, Floor 3
Sakura, Floor 3
Hishi, Floor 3
Kei, Floor 3
Yoh, Floor 3
Boardroom, Floor 3
Bay Penthouse, Floor 34
– Argos D-F
A target privacy threat model for the Web – Navis-A
JS Built-In Modules – Navis-B
Results from MDN Developer Survey – Navis-C
Introduction to W3C – Kashi
A Non-Technical Discussion on Decentralized Identifier (DIDs) & Self-Sovereign Identity (SSI) – Sumire
DataCue and "Time marches on" in HTML – Sakura
Anti-Homograph-Attacks – Hishi
Authentication by Communication Network – Kei
Improving Spoken Presentation of Content – Yoh
JSON-LD 1.1 Update – Boardroom
Japan Language Requirements Task Force: Evolving the JLReq document – Bay Penthouse
Lunch – Argos A-B, 1st floor
Web of Things PlugFest – Argos D-F
Efficient audio/video processing – Navis-A
Bullet Chatting – Navis-B
What is the Future of W3C – Navis-C
Standardizing user activation behavior – Kashi
OpenJS Foundation Collaboration – Sumire
New Module types: JSON, CSS, HTML – Sakura
Portals (status and next steps) – Hishi
Introducing: Audiobooks! – Kei
web-platform-tests update & discussion – Yoh
Privacy Budget – Boardroom
Introducing aria-virtualcontent – Bay Penthouse
Mini App Standardization – Argos D-F
XR Accessibility – Navis-A
Ad Measurement and Privacy – Navis-B
Next Generation TextTrackCue – Navis-C
Voice assistants - opportunities for standardization? – Kashi
Web stories – Sumire
WebAuthn network transport discussion – Sakura
Spec editing infrastructure – Hishi
Multiscreen/foldables – Kei
A new approach to large DOM and virtual scrollers using rendersubtree – Yoh
WebTransport status and next steps – Boardroom
Linked Data Security – Bay Penthouse
Break and W3C Groups Demos – Argos A-B-C, 1st floor
For a More Capable Web—Project Fugu – Argos D-F
DIDs (Decentralized Identifiers) Q&A – Navis-A
Supporting privacy-focused ads selection – Navis-B
Process 2020 – Navis-C
WebRTC combined with IMS – Kashi
Input for workers/worklets – Sumire
HTML 3D Element & Native glTF – Sakura
WebCodecs – Hishi
Getting text direction into RDF – Kei
ReSpec - so many new features! – Yoh
Web Packaging – Boardroom
Next steps for Web Monetization – Bay Penthouse
Registries at the W3C – Argos D-F
WebGPU – Navis-A
Building a Smartweb Powered by Blockchain – Navis-B
Edge Computing for the Web – Navis-C
Personalization of Web Content – Kashi
Images on the Web – Sumire
Github tools and Bots to assist Chairing – Sakura
Standardizing 360 video – Hishi
Digital Transformation, the Sentient Web and Cognitive Agents – Kei
UndoManager API – Yoh
Trust Tokens – Boardroom
Next steps for Web Monetization – Bay Penthouse
Reception – Argos A-B-C, 1st floor
We found broad consensus to build a target privacy threat model in PING, starting from https://jyasskin.github.io/privacy-threat-model/. We also got several concrete suggestions for things to ensure that document covers, which are recorded in the notes.
The group discussed how to expose timed events and timed metadata within media streams, e.g., ID3 tags or emsg events in CMAF. The discussion focused on whether the API should simply expose an ArrayBuffer, per the HTML5 specification, or expose structured data according to the scheme of the event. Next step is to continue the discussion in a session in the Media Working Group TPAC meeting.
The group also discussed the accuracy of firing of TextTrackCue events during media playback, when seeking, and when adjusting the timing of cues. The HTML spec does not specify a timing accuracy, and implementations vary. The conclusion was to report issues to improve particular implementations, and also add a note to the HTML specification on accuracy.
We showed the updates to JSON-LD since the 1.0 release in 2014 and described our transition to CR expected in a couple of weeks.
There was discussion about reactivating the JSON-LD CG to serve as an advocacy point, and to solicit input on a new JSON-LD timeline for subsequent changes.
One area noted was the non-resolution of text direction in both JSON-LD and RDF, which will be discussed in another session. If a solution is found without modifying the RDF data model, potentially the WG could adapt some syntactic sugar to extend the JSON-LD value object to represent text direction more natively.
So far, three implementations (Ruby, JavaScript and Python) are targeted to be updated to 1.1, and we’d like to see more. One implementation area of specific interest is C++, and developers are sought.
In room Argos A-B, 1st floor
After a short introduction to set the context, the group reviewed current performance of media capture use cases and reflected on experience gained on similar topics in Web Audio and Firefox OS project.
The WebCodecs proposal, coupled with hooks to it in dedicated APIs as proposed in the Insertable Media Processing proposal, was hinted as a possible solution in that space, even though WebCodecs does not currently aim at solving raw media processing scenarios for now. More thoughts should be thrown into designing a generic architecture for media processing. There was interest to continue discussions, possibly in the Media & Entertainment Interest Group or in a Community Group.
We talked about the current (broken) state of the Web and how Chromium recently changed its user activation model to address related problems. Then we addressed three specific comments our pull request received in the past: scope of the change, testability, and defining activation triggering events. By the end of the session, everyone was aligned with the proposal although there were some concerns about compatibility in edge cases. Chromium's rollout helps assuage some fears but not all. We decided to proceed with caution, and also laid out a longer term plan. See also a more detailed report.
This session opened by introducing proposals for JSON, CSS, and HTML modules, in which the JavaScript module infrastructure is expanded to allow developers to 'import' content of those types in addition to JavaScript. The new modules would participate in the module graph and share all of its benefits: static resolution of dependencies, parallelized loading without extra work by the developer, clear definition of components/API surfaces, and assurance that duplicate dependencies are only processed a single time.
The following 3 open design questions were then introduced and discussed:
An overview of WAIs current work in the XR space was given, such as in the Research Questions Task force and the APA, including introducing the draft XAUR User Requirements document. We discussed some the current accessibility challenges such as a lack of declarative semantics for accessible XR content, as well as issues with rendering environments and performance.
The Accessibility Object Model (AOM) was discussed, there was a sense that there are limitations with the current AOM, including a lack of implementation of the virtual DOM tree. Other options such as using having native support for glTF in the browser that can give 3D content a DOM for current accessibility APIs to function in XR space were mentioned, as well as how sensor enabled technology may contribute to accessible XR experiences. The need to better understand keyboard and switch requirements was also raised, as well as use of haptics.
Roughly 20 people gathered for this lively session that began with a short introductory set of slides, including a video from MIT's Auto-Id lab concerning the proposed Open Voice Network. After further introductory remarks, highlighting a variety of use cases and interests, there was an active discussion. The session was attended by staff from both Microsoft and Google who were close to, but not directly involved in, work on Cortana and Google Home.
The conclusions were:
We heard from the 16 participants how they were drawn to the Web, their hopes and fears for the future, what get us going, etc. Among ourselves were a few people who were born after the Web. We all come from all walks of life and not one of us had a similar journey leading us to the Web. But every single one of us was amazed by the Web and continues to be in awe of its potential. A majority of us expressed how they hold “making the Web a safer place for people” very dearly.
The W3C Communications team will get in touch with the participants as we may wish to use their stories, anonymously or not, to tell a compelling story to the public, possibly even as part of an upcoming crowdfunding campaign.
We discussed the feasibility of a network transport for WebAuthn, and found that version 2 of Google's caBLE proposal also solves many of the same problems. We got clarity on our understanding of the properties provided by authenticator proximity, and decided that we would attempt to extend caBLE to support a network transport once a cryptographic binding has been established.
Discussed the Window Segments proposal and the foldables problem space. Attendees provided great feedback:
Next steps:
The talk covered the proposed rendersubtree attribute and how to use it to make large sections of DOM perform well without virtualization. The presentation (including videos for all demos) and IRC minutes are available. There were a lot of questions and clarifications on the details of rendersubtree.
Things not captured by the spec/explainer or the presentation included:
In room Argos A-B-C, 1st floor
Together with Anssi Kostiainen from Intel, John Jansen from Microsoft, and Thomas Steiner from Google, we jointly organized a breakout session for a more capable web under the umbrella of Project Fugu 🐡. You can see our slides online. In the session, we argued that to remain relevant with native, hybrid, or mini apps, web apps, too, need access to a comparable set of APIs. We briefly touched upon the APIs being worked on by the cross-company project partners, and then opened the floor for an open discussion on why we see the browser-accessible web in danger if we don't move it forward now, despite all fully acknowledged challenges around privacy, security, and compatibility. You can follow the discussion in the excellent(!) session minutes, courtesy of Anssi.
Jeffrey gave an overview of the Web Packaging proposals, implementation, and deployment, and answered questions from the attendees.
Bruce Lawson presented slides describing the problem statement and motivations for creating a standard.
Following this, Adrian Hope-Bailie walked through some slides describing the proposed API and the architectural decisions that are still being discussed ito how the feature should be deployed to browsers.
This was followed by discussion between the proposers and audience. All of the discussion has been captured as issues against the explainer.
This session focused on digital transformation as the adoption by organisations of advanced digital technologies as a means to achieve increased efficiency, improved flexibility and greater speed at adapting to changes in business conditions. This emphasises the strategic importance of data and the need for enterprise-wide data management and governance. The business challenges for digital transformation include the need to capture knowledge about information and processes in a machine interpretable form, and to bring people along rather than to impose solutions from above.
From a technical perspective, the integration of information services across organisations is greatly eased with the use of graph models for concepts and relationships. The second part of breakout looked at ideas for simplifying the use of graphs for the average developer, and the increasing importance of machine learning to address scaling issues. Experimental work was presented for an approach based upon decades of progress in Cognitive Psychology that embraces both RDF and Property Graphs. This seeks to build upon the latest models of how the human brain works, and the analogies with Web architecture. Help is sought for gathering use cases and data sets as a basis for driving further progress.
Some links: https://www.w3.org/Data/events/tpac2019/digital-transformation.pdf https://www.w3.org/Data/demos/chunks/chunks.html
Bruce Lawson presented slides describing the problem statement and motivations for creating a standard.
Following this, Adrian Hope-Bailie walked through some slides describing the proposed API and the architectural decisions that are still being discussed ito how the feature should be deployed to browsers.
This was followed by discussion between the proposers and audience. All of the discussion has been captured as issues against the explainer.
In room Argos A-B-C, 1st floor
In room