For a More Capable Web—Project Fugu

18 September 2019


A_ton_of_people, Anssi_Kostiainen, boaz, flaki, jeff, JohnJansen, Thomas_Steiner

Meeting minutes

<Thomas_the_fish> Feel free to ask questions in the IRC!

<diervo_> anssik just voluteered xD

tom: Google, Intel, Microsoft part of the Project Fugu effort
… we can do stuff on the web today, but we're missing building blocks to build "Tokyo" on the Web
… users want features and they flock to alternatives, e.g. super app-platforms WeChat, Cordova
… another example: Slack app based on Electron

[slide showing a number of APIs available to web apps today]
… mission statement: Enable web apps to do anything native apps can, by exposing the capabilities of native platforms to the web platform, while maintaining user security, privacy, trust, and other core tenets of the web.
… go to goo.gle/project-fugu to learn more
… we have a white paper available at bit.ly/powerful-apis

tom: all the work happens in the open, including design discussions
… the process we use from idea to a new API is as follows:
… identify need > write explainer > solicit feedback > formalize spec > ship it
… at formalize spec step we launch an Origin Trial in Chrome to give people ability to test these ideas with real users
… we don't want to end up in a situation we ended up with vendor-prefixes
… we iterate before shipping, improve the design based on feedback from real users and developers

current features in progress and under consideration are at bit.ly/fugu-bugs
… Recently Shipped or Origin Trial APIs
… you can find information on Origin Trials of these capabilities at developers.chrome.com/origintrials

tom: Image support for the Async Clipboard API

[showing demo of the Async Clipboard API]

tom: File sharing support for the Web Share API

tom: File receiving support for the Web Share Target API
… Badging API
… Contact Picker API

[demo of Contacts Picker API]

tom: Periodic Background Sync
… these were APIs that have shipped or are in Origin Trial today
… we have more explorative APIs in works
… Barcode Detection API
… Native File System API

tom: Next Steps
… SMS, Local Fonts Access, Notifications Triggers
… just a small number of features showcased today here, for the full list see goo.gle/fugu-api-tracker
… more resources available for you to learn more
… Web Capabilities page: goo.gle/capabilities
… Fugu Capability process: goo.gle/fugu-process

[Showing a tweet that shows some Fugu <3]

<Zakim> torgo, you wanted to emphasize the need for more interoperability with other browsers and other engines (as highlighted by the results of the MDN developer needs survey presented this morning).

<JohnJansen> can someone else scribe?

torgo: Dan co-chair of TAG, I LOVE this work AND we heard this morning in the MDN preso that developer pain point is cross-browser compatibility, we see these APIs coming to TAG
… I'd love to see even more emphasis outreach-y type of activities to get more other browsers onboard, for more convergence

alex: we're here to talk to you, we do everything in the open and welcome y'all to take part in the work from the very beginning
… if the tracker looks scary, just talk to us and we're happy to help

torgo: we looked at SMS Receiver at TAG F2F, other implementers might have different priorities and interest, so there's some gray area where I'd like to see more engagement, actively soliciting and listening other browsers

axel: SMS API is going to Origin Trial, so we do not let not yet ready experiments to escape the lab
… diversity of approaches is considered as part of the process, looking at alternatives

reilly: not a lead on SMS API, but working on Fugu, I found this API interesting since Safari has a different way to do a similar task [with SMS]
… maybe Safari iOS approach is a better approach, we have had very good discussion with WebKit engineers on this exact issue at this event and continue that discussion

thomas: we have Fugu syncs so invite Apple, Mozilla, others to join us
… we want collaboration

vincent: I agree with the points made, we had some of the earlier projects with Notifications, and some of our efforts, had interest early e.g. WebBT, WebUSB
… at that time 2014 onwards we put together proposals to gather feedback from community

[scribe missed that comment...]

<wseltzer> pes: have you thought about two-tiered permissions?

<gsnedders> ... and have a "web page" and a "web application" privilege model?

alex: all permission gates use async APIs as to allow UI innovation by browser vendors
… UX research to be done, to understand trust relationship of users with apps that are "installed" vs. infrequently user sites

john: windows perspective, we work with Shell team to ensure permissions are properly surfaced in Windows adhering to the platform conventions

wseltzer: thank you, one of the things that is interesting to me with my W3C Strategy hat on: this is a great collection of capabilities the web could look like in the future. Would like to hear how we can use the Strategy Funnel, w3.org/Strategy/Funnel to help gather and gauge community interest.

jeff: W3C CEO, maybe variation Wendy's, CGs are for incubation, Rec Track for standardization
… we're getting better at this, timing incubation and formal standardization
… worry is if incubating too long, we do not get the right participation by right people, not just implementers
… we need to make sure they are all in the conversation
… Rec Track is often a signal for the wider community to get engaged

alex: the reason to do things in incubation is to allow rapid iteration and fail fast as needed
… WGs are hard places for API design work, we're looking for guidance on transition from incubation to formal standardization
… our goal is to get to Rec Track with this


gsnedders: interop in MDN survey as a key pain point, question becomes whether vendors with less resources should focus on fixing interop issues or new features

alex: eventually not implementing stuff becomes an interop issue if enough browsers ship it and websites use it, and we do specs that have design docs and allow interop implementations

rmondello: Rick from Apple, I want to thank Dan for a request for ourreach
… re SMS, we've had productive design discussions around that feature
… looking at values behind approaches, UX
… the framing of an SMS feature vs. credential/password-type feature, encourage folks to get together and hash this really out

heycam: many of these APIs are interfaces to device-level, testing story is interesting, need Web Driver extensions

reilly: I've made noise in w-p-t, we've trying to build Web Driver extensions for these features
… that is 1-way mechanism so interested in evolving Web Driver protocol to better support device-related API testing

john: lead of PWA capabilities for Edge, we meet tomorrow with Web Driver group and want to discuss this with folks and are open to suggestions, we've had long discussions during the week

james: if we believe in multi-vendor web, we can't really improve the web faster than we reach consensus

vincent: example, canvas2d has been extremely valuable addition to the web, it was led by a single vendor and other came along, in the end we had success
… we can use a similar model in this context, we have Pointer Lock, Fullscreen in similar situation, 5-year gap between implementations still success

james: canvas2d was one API, in scope of this project there are 10s of APIs

torgo: one of the promises of the web to the user is, it is more ethical, TAG has documented principles for more ethical web
… discussing with people involved with Fugu, impressed with thinking around mitigations to expose possibly dangerous APIs in a secure, privacy-preserving manner

<rmondello> To clarity the notes around what I said earlier: We've (Safari/WebKit) has communicated that we think the "SMS" framing is either unideal or harmful, rather than a credential, authentication, and one-time-password framing, and have concerns around the message template format. We're hopeful that these differences can be overcome.

tom: thanks you all for joining us, let's keep on moving this effort forward together!

<tomayac> The slides from the breakout session are available at https://‌docs.google.com/‌presentation/‌d/‌1QPuoAWsk1BQfkg73x7IExvxi3VjIU9DGE0uIylprfWk/‌edit?usp=sharing.

For a More Capable Web—Project Fugu (presentation slides)

Minutes manually created (not a transcript), formatted by Bert Bos's scribe.perl version Mon Apr 15 13:11:59 2019 UTC, a reimplementation of David Booth's scribe.perl. See history.


Succeeded: s/in the future/in the future. Would like to hear how we can use the Strategy Funnel, w3.org/Strategy/Funnel to help gather and gauge community interest./

Succeeded: s/we do specs that/eventually not implementing stuff becomes an interop issue if enough browsers ship it and websites use it, and we do specs that/

Succeeded: s/Present+ John_Jansen//

No scribenick or scribe found. Guessed: anssik

Maybe present: alex, axel, gsnedders, heycam, james, john, reilly, rmondello, thomas, tom, torgo, vincent, wseltzer