Breakouts 2025: Collective
The other day was W3C Breakouts Day: Effectively, an opportunity for people to create sessions to talk about or work on anything related to the Web. We do this twice a year. Once entirely virtually (last week) and then again during hybrid meetings in association with W3C's biggest event of the year, called "TPAC".
This year there were 4 blocks of sessions with 4-5 sessions competing concurrently for the same time slot. There were 2 in the morning that I'd chosen, but then realized that I had calendar conflicts with other meetings and wound up missing them.
In the afternoon there were 3 sessions proposed by Igalians - unfortunately two of them simulteneously:
- Vectors of Neglect - about SVG, led by Eric Meyer.
- Revisiting in-band text tracks in MediaSource Extensions, led by Alicia Boya Garcia
- Collective funding, governance and prioritization of a browser engine projects, led by myself and Stephanie Stimac
There were actually several in that timeslot that I wanted to attend but the choice was pretty straightforward: I'd attend mine and Eric's. Unfortunately I did a bad job chairing and forgot to assign a scribe or record a presentation.
Eric's session told the story of SVG and how, despite it getting a lot of love from developers and designers and other working group participants, it just didn't get a lot of love or investment from the browser vendors themselves. Amelia Belamy-Royd's (former editor/chair) told us about how things ultimately stalled out around 2020 with so much left undone, burnout, disappointments, and a W3C working group that was almost closed up with a note last year before being sort of "rescued" at the last minute with a very targeted rechartering. However, that rechartering still hasn't resulted in much - no new meetings, etc. We talked about how Igalia has been investing, along with companies like Wix, in trying to help move things forward. Then the question was: How can we move it forward, together? You can read the minutes for more, but one idea was a collective around SVG.
I mention this because while might they seem at first like pretty different topics, Eric's session and my own are really just two things raising the same overarching question: Why have we built the infrastructure of the whole web ecosystem to be fundamentally dependent on the funding and prioritization of a few private organizations?. And - how can we fix that?.
My session presented some things we've done, or tried so far.
Things we're trying...
First, just moving from proprietary projects to open source has helped a bit. Instead of one organization being 100% responsible, the stewards themselves today contribute "only" ~80 to 95% of the commits.
My company Igalia is among the top few committers to all of the web engine projects every year for at least half a decade, and we have many ways that we try to diversify work upstream. What's interesting about attempts to diversify funding is that there is a kind of a spectrum here relating to both the scope of what we're tackling and how well understood the approach is...
On the left is a single sponsor with very specific/measurable tasks. This could be a list of bugs, or implementation of a pretty stable feature in a single engine. You can contract Igalia for stuff like this - The :has() implementation and proof-of-concept that moved the needle in Chromium, for example, was sponsored by EyeO. CSS Grid in Blink and WebKit was sponsored by Bloomberg Tech. We know how to do this. We've got lots of practice.
Moving slightly more 'right', we have things like our Open Priortization which is trying to decide how to prioritize and share a funding burden. We know less of how to do this, we learned things with the experiment, but we've not repeated it because it involved a lot of "extra" work.
Then there's our efforts with MathML, funded by grants, then partnerships, then a collective aimed at generally advancing and maintaining MathML support in browsers. That's now involving more decision making - there's not a tight list of tasks or even a specific browser. You need to take in money to find out how much budget you have, and then try to figure out how to use it. In this case, there's a steering committee made up of the biggest funders. Every 6 months or so there is a report of the work that was done submitted for payment. Is that a great model?
We also have Wolvic, which attempted to have partnerships and a collective, with a model where if you put in a kind of minimum amount every 6 months then you would get invited to a town hall to discuss how we'd prioritize the spending of available funds. Is that a great model? The thing I like about it is that it binds things to some kind of reality. It's great if someone gave $500, for example, but if 10 other orgs also did so and they all want different things, none of which we can do with the available funds... At least they can all help decide what to do. In that model, the input is still only informative - it's ultimately Igalia who decides.
Or, you might have seen a new thing hosted at the Linux Foundation for funding work in Chromium. That's more formal, but kind of similar in challenges. Unfortunately there was no one attending our sessions who could talk about it - I know a bit more, but I'm not going to break news on my blog :). We'll have a podcast about it when we can.
Or, a little further right there's Open Web Docs (OWD), which I'm pleased does seem to be succeeding at sponsoring some real work and doing real prioritization. Florian from OWD was there and was able to give us some real thoughts. They too are drawing now from grants like Soverign Tech Fund. They produce as part of this an Impact Report.
Or, maybe all the way to the right you have a general browser fund like Servo. There it is the Technical Steering Committee that decides what to do with the funds.
A lot of this thinking is currently still pretty esoteric because, for example, through the collective, Servo isn't taking in enough money for even a single full time employee. I also mentioned our work with Interop, and the fact that even with that level of effort to priortize, it is very difficult to not wind up taking resources away from actual work in order to figure out which work to potentially do! It's hard to not add admin costs that eat away at the overall progress. But that's exactly why we should be talking about these things now. How can we make this work?
The Actual Discussion
As I said, we didn't have a scribe, but there was some discussion I can attempt to summarize from the queue I can still see in IRC.
Q. Do you think the W3C has a role to play?
That's why I brought it here. My own desire would be yes. There are some venues where working groups and interest groups are encouraged to pool funds and so on - but the W3C isn't currently one of them and I'm unsure how others feel. It's a thing I have been trying to pitch for a few years now that we could figure out.
The Wolvic model that I described above was born from thoughts around this. What if W3C groups had the ability to put money into a common fund and then at the end of the year they could decide how to deploy it. Maybe sometimes that would be for tools. Maybe sometimes that would be for fixing bugs. Maybe sometimes that would be for new implementation, or last implementation, or supporting spec authorship. But it would always be grounded on something real about collaboration that puts at some degree of power beyond the implementers. Maybe that would mean that print CSS would inch along at a snail's pace - but at least it's better than not at all, and it shows a way. At least it doesn't seem like the group isn't even picking up the phone.
Q. Are engines like Servo and Ladybird filing implementation reports? WPT?
Basically... Not exactly. The main thrust is that both engines do make themselves available on Web Platform Tests and closely track their scores. That said, the models differ quite a bit in their approach and Ladybird, for example, generally is more focused from the outside in. It's possible they only start with 25% of the standard required to load a certain page. So, at what point would they submit an impact report? Both projects do currently open issues when they find implementation related questions or problems, and both have the ability to expand the test suites in response to those issues. So, effectively: Nobody has time for that level of formality, but they're following the spirit of it.
Q. Doesn't it help to have a thing you can measure if you approach someone to ask for money?
Yes, that is the whole point of my spectrum above. The further right you move, the more of a question this is. On our podcast, for example, we've asked people "Why don't you donate to Mozilla?" and the main answer given is always "I don't know what they'll spend it on, but probably not things I would agree with". The more people putting in, the more there is to balance or require trust. Currently a lot of these work for grants or more specific/finite problems which do reports of success - like the one I provided above for MathML or the implementation report in OWD.
But does it mean there shouldn't be more general funds? Currently the general funds come from default search and wind up with a single org to decide - is it what we want?
So, that's all that was discussed. All in all, it was pretty good - but I'm slightly disappointed there wasn't a little more discussion of how we could collectively govern or priortize, or even if we all agreed that that's something that we want. I'd love to hear any of your thoughts!