Video and Slides
Outline: Web is Communications
Dominique Hazael-Massieux, W3C, Web Technology Expert & Software Engineer
Beyond WebRTC, the Open Web Platform expanding capabilities are helping make Web browsers a key component of existing and future communication systems: this presentation will review the ongoing work in W3C that paves the way for what future Web-based communication systems may look like.
Presentation Review
You can ask Dom questions in the comments section of this weblog, or contact Dom directly, his info is in his presentation.
With one web server you can reach 4B people! The web is an amazing platform. Yes, there is the awareness issue, let’s assume your service goes viral 😉 Permissionless innovation is made possible, you can just do it on the web!
Dom introduces the W3C (430+ organizations and 60 staff) that works to keep the web as open and accessible as possible. WebRTC is a big achievement in web communications, and there’s still much to do in opening up WebRTC to support innovation across codecs, architectures, and protocols. And better supporting features such as screen-sharing, background blurring/removal, and taking advantage of WebAssembly and WebGPU.
Security is critical, and anchoring trust in identity is a hot topic, with WebAuthn, Decentralized Identifiers, and Self Sovereign Identity. We heard yesterday from Guillaume from TeleSign on the importance of the convergence of offline and online identity.
The Web Neural Network API (WebNN) will be important for many communication applications such as noise suppression, transcription, translation, and recognition. Regardless of ability, real time communications should be accessible to everyone.
With Web Transport, QUIC, and HTTP3 there a many new ways to optimize communications performance over the internet. One topic Dom mentioned is scalable video coding; matching coding to the screen. I worked on a European project nearly 30 years ago on SVC, back then it was broadcast and contribution quality SDTV and HDTV. Not the explosion of formats we have today.
The scope of W3C is vast, also encompassing progressive web apps, payments / monetization, and the Immersive Web (AR/VR). The Web is Communications, and any company involved in programmable communications should be part of the W3C.
Thank you Dom for an insightful presentation of how the Web is Communications 🙂
Thank you Dom for an excellent review of the importance of the Web and W3C to communications. I have a few questions:
1) Will the next version of WebRTC take as long, or do you think the many planned WebRTC improvements will see more incremental adoption?
2) Identity authentication is something we’ve covered at TADSummit for several years. DID and SSI have come to the fore with the need for electronic COVID vacation and test certificates. What’s your view on European Self-Sovereign Identity Framework (ESSIF)?
3) A point made by Guillaume Bourcy, another TADSummit Asia presenter, is the need to converge off-line and on-line identities. They focus on the mobile phone numbers as the link between those worlds. Do you think W3C has a role here, or is that too deep into implementations?
4) There are so many cool features in chat and video communications / sharing. You showed an AR example in your slides. Where should the W3C draw the line on how far standardization goes?
5) WebNN is interesting, the power of users’ devices to implement algorithms that are time sensitive, like noise suppression and could have a significant benefit for green compute. But how does WebNN know there’s enough local resources to implement the algorithm?
6) What do you think are the main accessibility challenges the W3C must focus on?
7) At FOSDEM there was an interesting presentation on the Greener Web (https://fosdem.org/2021/schedule/event/webperf_building_a_greener_web/) what is W3C doing here?
8) What are the main benefits for a programmable communications company to join W3C?
Taking them one at a time…
> 1) Will the next version of WebRTC take as long, or do you think the many planned WebRTC improvements will see more incremental adoption?
The first official standard of WebRTC was indeed long in coming – we started some 10 years ago! Of course, the reality is more complex than what that figure might imply, since WebRTC has been available in various shapes in browsers for many years as well.
In any case, with that milestone behind us, the W3C WebRTC Working Group where the specifications are being developed has already started using a much more incremental approach to its next rounds of improvements to the technology stack. This means defining smaller specifications that we expect to see adopted faster among implementations.
The work I mention in my talk on WebRTC Scalable Video Coding, or WebRTC Encoded Transform (the key to end-to-end encryption among other things) for instance are defined as small extensions to the main WebRTC standard.
> 3) A point made by Guillaume Bourcy, another TADSummit Asia presenter, is the need to converge off-line and on-line identities. They focus on the mobile phone numbers as the link between those worlds. Do you think W3C has a role here, or is that too deep into implementations?
W3C is already quite active in the space of identity management – as my talk alludes to, our work on Decentralized Identifiers and Web Authentication are the most current illustration of this, but there are also early discussions on improving the flow of federated login through new browser APIs.
The ongoing work most closely related to Guillaume’s focus on managing identity toward reducing of fraud (in particular in the payment ecosystem) is our work on Secure Payment Confirmation which combines Web Authentication with some of the dedicated payment flows made possible by the Web Payment APIs to create a very smooth user experience with high guarantees against fraud, as illustrated in a recent experiment run by Stripe https://www.w3.org/blog/wpwg/2021/03/26/secure-payment-confirmation-stripe-experiment-and-next-steps/
On the specific point of getting online and offline identities to converge, I would say W3C is taking a different approach: with the very strong focus of our community on privacy, and the recognition that in general, people don’t have a single “off-line” identity, but instead of kaleidoscope of complementary identities, positioned on a spectrum of “on-line-ness”, we are instead ensuring users have as much control as possible on controlling what identities get exposed where, which is driving a lot of our work on Web Advertizing and some of the changes being considered in terms of cookie management, fingerprinting prevention. The work I allude to above on Secure Payment Confirmation is an illustration of that – it’s not that a merchant or a payment service provider needs to know your name to have confidence on the transaction you’re pursuing – they need to have confidence you are the person with the legitimate control of the payment instrument.
Phone numbers certainly have a role to play in identity management, but they have a number of limitations that would make me wary of making them central to it: they convey both an identity and an intrusive communication endpoint, their security model has historically provide insufficient against targeted attacks, their ownership change overtime in ways that are difficult to detect, and the properties of the devices they attach to are both constrained and undetectable – I have recently had to deal with account recovery linked to a phone number that can’t receive SMS, and that has not been a particularly pleasant experience!
> 4) There are so many cool features in chat and video communications / sharing. You showed an AR example in your slides. Where should the W3C draw the line on how far standardization goes?
W3C’s mission is “to lead the Web to its full potential” – in other words, the sky is the limit!
More pragmatically, W3C standardization requires getting agreement among people and organizations with very different needs, wants and backgrounds – that’s what draws the line of how far standardization goes. The fact that so much has been standardized and is being standardized in W3C is a recognition that the value all these contributors have put into that work has so far well exceeded the cost, and I’m optimistic there is room for a lot more of that!
To me personally, the Web is unique in its openness (both in terms of how it is being designed and how anyone can make use of it), and so its continuous harmonious development is key to providing a resilient technology infrastructure for our societies – we need it more than ever.
> 5) WebNN is interesting, the power of users’ devices to implement algorithms that are time sensitive, like noise suppression and could have a significant benefit for green compute. But how does WebNN know there’s enough local resources to implement the algorithm?
A big part of the ongoing design of WebNN is about addressing that question – the Web has always been designed to be device-independent, and WebNN definitely needs to accommodate very different level of hardware capabilities.
The high-level approach is that developers can either ask for a specific type of hardware acceleration (CPU, GPU, other *PUs), or let the browser determine what it thinks will work best. But the exact details are likely to evolve significantly based on feedback we get from early adopters – WebNN is still very much in its early design phases.
> 6) What do you think are the main accessibility challenges the W3C must focus on?
W3C tries as much as possible to address all disabilities, not a subset – it does this by bringing guidance that can apply across the board (as done in the W3C Content Accessibility Guidance) or technologies that expose information that assistive technologies will need across the board.
The high level challenge of making that possible is that the underlying platform evolves continuously and sometimes toward very significantly different user interactions: I organized late 2019 a W3C workshop on making augmented and virtual reality work for people with disabilities, and the challenges in that space are immense. And as with most other immense challenges, we expect that a combination of iterative improvements, advocacy and design with intent will help make them more and more approachable.
> 7) At FOSDEM there was an interesting presentation on the Greener Web (https://fosdem.org/2021/schedule/event/webperf_building_a_greener_web/) what is W3C doing here?
We don’t have a structured program in this space yet; there have been previous attempts at figuring out how we can design Web technologies to make their impact a net positive to the climate crisis, but they have not gelled yet on specific improvements yet. There is a clear recognition of the need as illustrated in the Ethical Web Principles identified by the W3C Technical Architecture Group, the senior technical board in W3C https://w3ctag.github.io/ethical-web-principles/#sustainable
From my perspective, a particular challenge in this space is ensuring the right focus: we are facing a world-wide emergency with a lot if parameters, and there is a danger to focusing on the wrong problems (e.g. reducing energy consumption to the small-scale level when each new blockchain project wipes it out in a minute).
It’s definitely a topic very close to my interests, so I would be extremely happy if people got in touch to discuss ideas for systemic improvements in this space.
> 8) What are the main benefits for a programmable communications company to join W3C?
The position I advocate in my talk is that the Web is a critical platform in any strategy for programmable communications, because of its openness, its universality and its magic “one-click” distribution model.
Not everyone will agree with that position, but for programmable communications companies that do agree, joining W3C is then the logical path in ensuring this critical platform fulfills the need of your product or service, and the needs of your customers.
(the alternative is to leave that to your competitors – I’m not a business strategy consultant, but I don’t think that’s what one would usually recommend)