TAG in Paris!
For our first face-to-face meeting of 2025, the W3C Technical Architecture Group met in Paris. It was my first face-to-face with TAG and a great insight into the group's processes, dynamics and decision-making. The group is very varied, and everyone has their own expertise, so while I was able to learn some new, interesting things, I was also able to share my thoughts and expertise. Here's a rundown of our first F2F!
Societal Impact Questionnaire - I’m a co-editor!
Part of my running platform for TAG was to contribute to and advance the Societal Impact Questionnaire previously edited by Amy Guy. In fact, my co-editors (Sarven Capadisli and Tristan Nitot) also included the questionnaire as part of their running platform, so we'll be working together to refine and develop what's there now. It was clear that the purpose of the questionnaire could be easily misunderstood; it's possible spec authors will see this as a checkbox ticking exercise or another document to fill out before submitting their spec for design review. That is not the purpose of this document. The Societal Impact Questionaire is less of a rule book and more of a guide to encourage spec authors to think about the impact of their spec during development. Ideally, the questionnaire would be referred to at different points in the spec creation process, and we'd help with the incorporation of the questionnaire into the spec creation process, but we're still fine-tuning the wording and language. Sarven, Tristan and I are all passionate about the impact that technology has on our world and society. As we've seen recently, tech is not neutral, and it's important for us to think about who/what is (positively & negatively) impacted by what we build.
We Talked to French Developers (in English)
On the Monday, Tristan (who very kindly hosted us) organised a developer meetup to allow developers to hear about what we're working on and for us to get feedback from them on what we should be working on. We had a really good turnout and some really good insights. I spoke to a number of people after the session about working in tech as a Black woman, accessibility, interoperability of AT technologies, and more, but two things stood out to me.
A Desire for Versioning in Baseline
A developer expressed (kind) frustration at understanding Baseline. At the moment when a web feature is considered baseline, there isn't any data that says when it became baseline.
For developers considering backwards compatibility (which is all of them), this presents a problem. If something became baseline in browser version 4, a developer doesn't know that; instead, all they know is that the feature is baseline, and they may interpret that to mean it's available in all browsers regardless of version, which isn't true. The good news is that this is something the W3C WebDX group is currently working on, so this will change soon!
Accessibility Interoperability Data - Who is Responsible for it?
Currently, Browser Compat Data (BCD - the dataset that shows web feature availability information on MDN and CanIUse) shows what ARIA features are available in what browsers, just as it does with any other web feature. However, some web developers I spoke to don't feel that's enough. The difference with ARIA features is that there's another dimension to them, assistive technologies, mostly screen-readers. Web developers don't just need to know which ARIA features are supported in the browser but also how the different screen-readers interpret them. And while MDN, CanIUse and others primarily deal with the web only, I agree that this data should be available on these resources as it's crucial to building accessible sites, apps and services on the web. The ARIA Authoring Practices Guide (APG) has some data on the availability of the patterns in each screen-reader/browser combination, and myself and various colleagues have tried to see what it'd take to include this data in BCD, however, since the focus of APG is patterns, it's not quite a fit. However, Michael Fairchild (Senior Technical Program Manager for Accessibility at Microsoft) has built Accessibility Support, which shows exactly the data we'd want. Getting that data into BCD is still a mammoth job, as the shape of the data is different, but it's something the WebDX group will explore.
We Reviewed Some Specs
I won't say too much on this because discussions are ongoing internally and we typically like to update spec authors on the appropriate GitHub channels. However, we were optimistic about how many design reviews we wanted to get done, and as these things go, we didn't do as much as we'd liked. Updates will be coming for the specs we reviewed in the coming weeks.
Web User Agents & High Performance Baseline CG - I Led a Breakout!
Web User Agents
TAG is working on a Web User Agents Findings document to define the responsibilities of a web user agent. We primarily worked on the language in the document and making sure it conveys what we intend it to. This work is important as we think about the diversity of user agents in general and what engineers should be considering when creating web user agents specifically.
High Performance Baseline Community Group
This is a brand new community group that's concerned about diversifying user agents. Xiaocheng was a representative from the group and filled us in on the background, purpose and questions the group has. This was a more advisory session. first up, we advised the group to consider a name change because it's easy to confuse with the already established Baseline. We also advised the group to do some exploratory work on what new web user agents (such as Ladybird and Servo) consider the most important web features needed to get up and running. This will help the group understand the needs of web user agents and what features may have the highest returns for users and developers.
I Learnt New Things!
The Web in China
The W3C is a predominately Western organisation, and we create and build from that lens, despite the things we create intending to serve a global audience. We assume that everyone engages with the web the way we do and then, for better or worse, make principles based on those assumptions. Xiaocheng led a session on the web in China, and it was really enlightening to learn that most people in China do not engage with a web browser; in fact, less than 40% of people engage with a web browser. Instead, they engage with Mini App Platforms (also referred to as Super Apps), which operate as apps, e.g. the chat app WeChat, but also as platforms where other apps can be hosted. For example, you can order a Starbucks drink through WeChat. These platforms are different from mobile platforms because they’re built with web technologies.
Digital Credentials
This really was an introduction to Digital Credentials session for me, as most of my understanding has mainly come from the news and random software updates/proposals. The purpose of this session was to understand what is needed to write a finding for DCs, this is an opportunity for TAG and the W3C to be an authoritative voice in the space, encouraging regulators to take what we say into consideration. Our priority is making sure DCs are safe and don't pose too much of an additional surveillance risk. An authority shouldn't need to be able to create a profile on you based on what you purchase, watch, drive to, etc. I'm aware that there are other technologies that enable these things today, but DCs could possibly be used to link all these things together to create a master profile. It'd be bad if an authority could see that I purchased a can of coke, drove up to see my partner and we watched the latest episode of Abbott Elementary, just because I used my DC to buy the coke in the shop, showed my DC to a police officer who stopped me on the way and used my DC to log into the streaming platform. This is a benign example, but I hope it's clear to see how this could easily be exploited.
I'm Back Home
The 3 days in Paris were productive, enlightening and empowering. I've come back with a whole bunch of PRs to review and submit, issues to open or comment on, and projects to start and continue, and I'm excited to get cracking!