-
Notifications
You must be signed in to change notification settings - Fork 6
2024 Funding the Web Ecosystem
- GitHub issue: https://github.com/Igalia/webengineshackfest/issues/40
Servo tweet: https://social.tchncs.de/@Blort/112550943514436964
Stephanie
- continuation from yesterday
- last night Servo post on Mastodon got a bunch of donations
- ppl want to fund
Brian: There are many different standards bodies. Some of them, like Khronos actively encourage joint funding. That's a really interesting thing we should look into/learn from in other standards bodies. Would take a while to integrate that. Could lead to prioritization questions, engine questions... just as interesting. Interop project is a good example where we tried to do that together with the companies that already have massive funding. Gets into the thing Stephanie was talking about. This is the MathML fund, this is the SVG fund, etc... we could do that kind of thing. If you make it any bigger, you have to talk about how to prioritize the money, which is super interesting on its own.
Dietrich: One question I had is can you post a link to the mastodon post about Servo? I also wanted to ask a question: One thing that is interesting is "Who would that be?" Companies that have millions and millions of dollars - but if you start listing the stakeholders (5 billion) vs who would actually write checks - it's interesting. Standards bodies are an interesting case of organizations already writing some kind of checks - so is membership of IETF, for example.. No fees, but paying to participate there. Thinking about that is an interesting exercise
Stephanie: I've been involved with web We want for a long time - the idea was to help create a roadmap of features developers want, but one gap has been finding the companies to help fund and make that a reality. Like, we have a member from Salesforce who has been active in those discussions. Marketing, in a sense. This person on Mastodon said "This is a non-google controlled browser" - it comes down to the right call to action/the project
Dietrich: One of your earlier slides - what happens when one company fails... Google is maybe not the best example because they have the resources to stay around for the next few hundred years. The smaller attempts are interesting where do they have any runway is... What about the cases where that pyramid is upside down - if they fail that whole area of the web could disappear. There would be outsized impact in even a small amount of funding
Stephanie: Talking about single points of failure, I know I've seen open source maintainers announcing they are going to stop maintaining - it takes so much to keep it up. That's a really big problem
Vaadim: Yesterday I realized the Linux Foundation doesn't maintain any funding for Servo - what is the point?
Rego: When Mozilla stopped working on it, they donated it to the Linux Foundation. They have a few kind of projects - they have some hosting stuff, the trademark. There are some that have a board, different kind of 'membership' agreements, a TSC. Then there are bigger ones, like AGL which have many companies putting money and a board and way to decide how to use the budget, etc. With servo the idea was to have more organizations, but it hasn't really happened yet, so it has mainly only fairly basic benefits of being in Linux Foundation. There are some other things, events, opportunities to have talks and so on.
Dan: bloomberg is a member of the openjs foundation, part of the linux foundation. two independent 501c3, foundation with service contracts b/w linux foundation and openjs foundation, only employees are in the linux foundation. they provide services, for example: linux foundation helped with visa letters and logisitical support for an event in europe. regarding project founding, although the linux foundation doesn't pay for nodejs or jquery, it facilitates the transfer of money from other organizations to contractors that work on node. it provides trademark and helps with legal/money issues. With the fork of teraform (related to licenses), there was a fork, then teraform was upset, linux foundation provide legal help with negotiating the fork. My concerns about linux foundation is related to their prioritization and their efficiency. one final thing they pay for in the case of node is the computers where the CI runs. but programmers have to be at the core of any open source sustainability solutions
Valerie:
- re programmers being source & govt funding
- sovereign tech fund - german govt fund
- funds dev of open source infrastructure
- advocated a dev-first way of managing
- seek out people
- request proposals
- very general - but only libraries
- not browsers, but maybe web standards? unsure
- good direction, but goodness of hearts of govt of germany
- socially funding is good
- bureaucratic orgs have challenges
- participatory budgeting at govt level
- newer approach, not tech specific
- spending the peoples money in a way that comes from them
- mostly at city level, not country
- proposals come from citizens, and refined
- chosen democratically before execution
- should apply to open source
- maybe LF could be using approaches like this
- how to keep dev-led, not top-down
links! This a non profit consulting firm which helps cities that want to do participatory budgeting: https://www.participatorybudgeting.org/ Here is a doc that describes the process: https://www.participatorybudgeting.org/wp-content/uploads/2023/05/ScopingToolkit2017_v1.1-1.pdf
Andreu: You mentioned the Sovereign Tech Fund, goodness of government's hearts -- should we be lobbying towards gov't funding in open source and browsers and web specs at the same time as we market that idea to people so that they can vote on it in a democratic fashion?
Valerie: Yes!
Dan: Working with gov'ts, taxation-funded things is good. Solves the free-rider problem. A company like Bloomberg, we like to say we're taking the long view so we fund things -- we're not the size of a big tech company, can't shoulder the whole world's funding problems. We're seeing these cuts done at the large companies, has to be a shared responsibility. So we can do a bit where we can justify particular business goals, but doesn't extend to everything. Should be a bit taken from all the companies, taxation accomplishes just that. Could be taxation specifically of large software companies. Public funding is good; we should be conscious, though, that public funding is likely to be military-oriented. Happening all over the world. In the US there are some analogous structures to the STF, some parts of the military fund software. In Europe, there's an exception to the freeze where they can spend more on the military. We could lobby truthfully that funding these projects solves military goals of improving our defensive security posture. US gov't has come around, insecure software is unsustainable. Led to OSS supply chain security stuff. We in open source have successfully inserted ourselves as a key dependency of militaries. They will fund us to the extent that they see it as in their interest. Per yesterday, many of us are anti-capitalist, we have to decide what to do w/ this reality. Possible to work with while maintaining ethics as long as we take a careful look at every action we take while receiving this kind of tainted funding. Adtech money isn't necessarily better. The beautiful participatory gov't thought, the military, it's the same budget. We have to get increasingly careful about ethical decisions, which projects to take on. Not just whether a funding source feels good or bad. A bad source could trick you into doing a project that's destructive. Be realistic about what funding sources exist.
Brian: For browsers, we have -- Google, clear reason for Google to invest in Chromium/V8. Good reason for Apple to invest in webkit and JSC, good reason for Mozilla to invest in Gecko/SpiderMonkey. Talking about orgs that are not those orgs investing, it's not as clear what they should invest in. When we go ask a gov't or some other org, how do you make the case about what to prioritize and why? Also an interesting question. When we did our open prioritization experiment, it was difficult, had to pick things representative in each engine and each area in the web platform. Like Dan was saying, there's a lot that gov'ts have to say, or like Val said, there's a lot they have to say when it comes to security. Security angle makes it easier to get funding. Other needs of the web platform, you can't really tie to that. What kind of need is it and how many people can you get interested/involved? Who's going to fund MathML? Clearly important, but who's going to fund it. SVG, a different angle, when we were shifting to wider color gamuts and things, different formats, who's going to pay for all that? Has to be done across platforms. So I think there's a lot of interesting questions, curious if anyone has any thoughts. Would like to see a browser project that's more diverse somehow.
Keith Cirkel: I work for Github, trying to make the case that corps should fund some of this work. We did some projects around integrating MathML. Mostly relied on polyfills rather than putting up the money to do browser dev. I find it difficult to make the case. I've put energy into it but I'm not sure what the right message is to the company. It seems difficult to make the case, mostly because of conflicting needs. A company wants to deliver a feature within quarterly timelines, which doesn't match the length that browsers integrate. If I could wave a magic wand and say we can deliver things faster, would be an easier sell, but I don't know what the alternative is since that doesn't work. Any suggestions? I want to make the case that companies like Github should be paying for it.
Valerie: Just to be a little pessimistic, I don't think we can ever expect companies to reliably pay for stuff. Companies aren't people, they're entities looking for profit, it changes on a whim depending on market pressures. Sometimes they'll fund things for a little while. Your efforts are noble, but I don't think it'll ever be a reliable source of money. Thinking about gov'ts a better way to go. One of the things that will force companies to put more funding into open source is if it improves their image. For example, Meta open sourcing AI. Their image was starting to go pretty down and now they have a lot of positive press. That's cool b/c there's this social pressure to provide back to OSS infra. It'll only be done performatively, but the more that we do push open source and make it more clear to society at large or whatever, the importance of these things, then companies will be more incentivized to look good and give back. Andreu, you said something along these lines about work in the public space, will also get gov't support if public is more interested.
Bernd: I disagree with the Meta and OSS perspective. If you're ahead, you do closed-source, but if you're behind you do open source.
Valerie: Also a good point
Bernd: Gov't approach is way better. ... I think we have to go into institutions. Lobby them, people have to go in there with the right spirit. Have to know a little bit about it and feel it. If you're part of the institution, say a director of a tech fund, it's easier. It's not easy to get in, you have to be in the institution all your life. Sunday, European elections are coming up; important to go in the right direction there.
Brian: A way to unify these ideas -- tax breaks for companies? Insulate yourself from whims, tax break is an incentive. Tax breaks for companies to participate in the thing. I've talked to a lot of companies and they have occasionally made that point to me: if that was a thing, they'd have an easier time making something happen.
Dan: It can be a little hard for recipients of funding to prove their effort is worthy of a tax break. As it is, you could set up a nonprofit. Anyway, private funding is unreliable. In particular, we have a US corp culture about short-term results. Not nec. in alignment with long-term value of companies. That's how these big orgs operate for the most part. One thing they're responsive to is regulation. This has been a big move recently, lots of gov'ts -- US gov't has been saying security is important, and getting concrete about it. Now companies are working hard at that. If that notion is sufficiently elaborated in productive directions, it'll be helpful. Keep an eye on the underlying thing -- this is about the military, but that's ok because it's aligned in this case with our goals. We want projects to be secure as well. That's how gov't justifies it to itself. If we want to get these multiplier effects - -can get them from justifying things on the basis of maintainership and security. If we want to make new features, like wide color gamuts, maybe it could be funded by a particular source but backstopped by broadly looking at testing and code reviews and these things, which are security-critical, have to be done well, if those maintainership things could have stable funding, which they currently don't, then feature dev work will be able to rely on that multiplier.
Andreu: What you were mentioning, how does it play into things about making ... (Didn't hear)
Dan: Making sure that software has good maint. practices is a goal of lots of different orgs, including Linux Foundation. We have OpenSSF software scorecard. More points if you have multiple people signing off on releases, etc. Issues with the methodology but it's something. Once we have a way of evaluating how well-maintained things are, that's a metric we could target with funding. Esp. once we have ?? visibility into what software dependencies are. Orgs with a mandate to have security can examine that. Can be used to direct funding towards most vulnerable dependencies. Ultimately there will always be people in these systems, no full solution, but as long as you make sure that there's multiple people and it can be their job, people can be held accountable by formal systems to do good maintenance. I think this stuff is solvable. We have certain outlines for a solution; lobbying we do can take place in some of the imperfect institutional frameworks that exist.
Dietrich: Managed a contract about open source (space stuff???) - how software that is open source will be used. We knew some of the people who were using it. I wanted to talk about a company that paid igalia to work on web platform, the shifting windows of change at company, the amount of money goes up and down, the sustainable funding you need for an actual open source project is constantly disrupted by company. I had to fight a lot of keep that funding. I left the company, but still work on the web platform parts. I started a new org that I'm using as a nonprofit umbrella to getting funding, p to p web techs. There is some renewed interest right now. But until we get large nation state level -- best option is w3c and other industry consortiums are better than the companys interest fluctuating. Now I try to get my old company and other company thats depend on this web stuff to pool their money for these standardization work. Hopefully, next year, I can tell a story about whether that worked or not.
Dan: active military funding for Tor...
Valerie: Brian, you've talked in the past about the W3C maybe funding implementation work. I'm surprised you haven't brought it up yet in this discussion. I'm wondering how those conversations have gone in W3C
Brian: I talked about this before W3c members funding things... not the w3c itself. I know for epub, they have an open source conformance checker, they needed some software maintainance stuff, two 2 years to raise 40 thousand dollars. But there are interest groups within the W3C and those groups of companies might make sense for them to pool their money. some groups like Khronos do, we should find a way to encourage that. We could use that w3c to do that.
Valerie: That all sounds really great. Have you been talking to W3C staff about this? Infra to encourage companies that are in W3C to have a special way to donate - W3C is nonprofit, manages the payments to developers?
Brian: So, it's only now, that the w3c is a 501c3, as far as when it was restructuring. Every group that joins has one representative. The AC as various meetings and we have a mailing list, and I've brought it up several times, and gone to several groups to encourage them to do it. Like, the print group, I've been trying to go to each of these people -- image if you had just put it towards developing, what you would have. The business part of that which is that it is hard, that is not the norm, they don't want to change. As far as the w3c, that needs time to gel, people who aren't just me, you keep saying that same thing.
Dan: This is great, we can probably fund some things through sturtures like this. the linux foundation has structures like this, your membership fees are already spoken for, BUT you can donate money these other ways! What does that amount to? sometimes things get funded, but you still have the free rider problem. Igalia is running on the open source world which is funded in all of these unstable ways. BUT the LF is not a non-profit, so maybe we will see more donations to w3c since its a tax break.
Dietrich: open source devs are not business people, to accept money for this work, is an insurmountable barrier because of complexity. Opens to do this as a nonprofit is even harder. Open collective tries to help, for open collective europe... (missed). I've been trying to find an org/tool to get fund public goods with tax benefits internationally. There isn't a lot of documentation about this. (seems like Dietrich knows about a swiss company that does this and a Liechtenstein (government org?), he says it's "dark arts level stuff")
Dan: I work with a Swiss association if anyone else wants to be in touch offline.
Jonathan Kingston: (From Duck Duck Go) I wanted to talk about what Dietrich mentioned. Barrier for new people to compete is really high. Need 300M to just get off the ground (in browser engines). ... Using WebViews... a few sets of problems, standard WebView implementation, huge compatibility problem, drastically all very different, each one has their own quirks, WebView 2 for example is almost an entire browser, includes telemetry and things you don't want in a browser... then Android that's quirky in different ways. Not really ?? -- An area I think would be similar to user journeys or developer experience. If we could think of the company pipeline and how you make it easier for companies to -- Firefox for example used to have an easy way to embed a browser, and they removed that, unfortunately. No other ways to cleanly integrate a browser besides take a fork. Then you're battling however many Chromium engineers to maintain the decisions you want to make. You want to contribute to a standard, things like privacy, just to do that you have to invest so heavily in taking a thing off the shelf. Huge investment to fight Chromium or all these weird
Dietrch: this has been about funding opensource/web platform. We could spend another hour on the business of making browsers. What $$ you need to make a browser, its a barrier to entry to new browsers
Brian: We ran a break out session about funding browsers specifically, we talked about how search is the natural funding model for the web. We also talk about regulation. One idea we pitched, you could set a level on search, so that you have to pay if you have search that you pay into for browser dev....??/ but what about governence and how you spend the money. If you have companies like duck duck go together to create an arch that is more welcoming to the sort of thing that browsers what to do. If they bought into a fund to fund that and lower their cost. They would have a collective advantage to compete with a google.
Dan: its been discussed a lot -- media orgs make the case that search is destroying their business revenue. There is a lot of different, not software maintenance, that need to get paid for and aren't getting paid for. Like journalism.
Stephanie: A comment on search, I use microsoft edge, bing is the default search, I get points every time I used bing search, and you can get starbucks gift cards. I wouldn't mind that going back into the web instead.
Brian: Just wanted to throw out, all kinds of interesting experiments -- you can remove the technology from it. Brave has this idea that your eyeballs are worth money and they're worth money to them as well. So they sort of pay you to look at ads. You turn on ads, they sort of pay you. I think that conceptually is an interesting idea. I know I mentioned yesterday that traditional consumer television -- you are the product. The reason it appears to be free is you're consuming ads and that's worth something to companies. A long time ago when we had to pay for long distance, a company experimented with that where you would call up and listen to a few ads and you would get some free long distance calling. Interesting experiments, I don't know how far they can go, but they make me think about things a little differently. Very early on I thought Brave as interesting and set it up -- I want to put my money back into the ecosystem, I don't want to get paid, but I want it to go into reinvesting for creators. I'll put in $5 and whatever money you pay me, I'll add it to that. They were completely unprepared for that, the money just got lost because they didn't think that's what someone would do. Basically that's what Stephanie was just saying. If there was a way to reconnect the payments they're already paying, instead of routing it circuitously from Google to Apple to the WebKit team. This is what you're worth to us. If Google said you're worth $12/year to us by seeing these ads, and there was a way for you to opt in and say, I like paying for the web because somebody has to. I'm curious what would happen if there was an easy way to do that.
Dan: but for search, how do assert that we are worth it more than other things? Something about funding planting trees. About brave, it was financed by a crypto scam, so it's idealistic and not yet proven.
Brian: Could remove a lot of things from the equation, but it's the idea that's interesting
Bernd: It's interesting how search will change with large language models
Vadim: Igalia experimented with Open Collective, how did that go?
Brian: there are different levels at open collective, a super group called open prioritize. We have done many experiments, we picked 5 or 7 different standards that were in different phases of the standard cycle, one had zero implementations some had two, different browser support, different techs. There were several points, one was to raise awareness and directly include developers. Shed light on problem of prioritization, you think that people will pick the same thing as you. One the shop talk show, they talked about this, one host said that one standard would win, but he was wrong another one won. We did raise enough money on :focus-visible, alice worked on and rego implemented, it was successful. Since then, we have tried, like mathml and wolvic. For wolvic, we are trying to get all the companies interested in XR and XR devices, if you contribute x amount in 6 months you get to help collectivey prioritize what you can do with that money. Once you are in that meeting you can see how much money that actually gets you, raise awareness. We also have one for servo, people like servo, for some reason it has the biggest collective interest among individuals. I'd like to do more, if you have ideas.
Dietrich: I have a question about the granularity of those donations?
Brian: I think you can donate directly to the top level, but no one has done that.
Dietrich: if you talk about legislation, you hit the problem of allocation. Stephanie brought up, impact evaluator. Thinking through scoping level of stakeholders, impact evaluators. Framework for articulating your needs and allocating based on needs.
Brian: Eric and I are on the committee for the interop project. The needs are the entire universe, and the submissions are who saw it within the month that is open. 150 things get open. how do you chose between those? it's complicated, it takes resource to evaluate. The public only sees the outside of that, not the evaluation. You see public signals like github emojis and stars, for example, one of the most starred things was mathml -- so why was it not funded? Even though it had the most stars, I don't think mathml is actually the most popular thing. Even if we can collect pots of money, how do you choose? Val asked about w3c, trying to get working groups toward this problem. if they managed to collect 300 thous, but you can't do all the things, they'd have to work together to build consensus, and prioritize. You need money as part of the discussion to keep the prioritization real.
Stephanie: One thing I saw on edge team before we switch to chrome. We had "user voice", people were encouraged to upvote things they wanted. We would get peaks because someone who had a large following because an influential person wanted it. Github stars and upvotes are not a key signal, easy to manipulate.
Brian: to throw in there, there is a signal, but some things have clear business but are important. like accessibility, internationalization, math, how do we get those things done?
Brian: I want to go back to keith's question about time frames, I've seen people get involved briefly in standards and then burn out. How do we improve that situation somehow? there is an aspect of standards that need improving? how important is it that something become a standard rather than you go outside of the standard process?
Keith: my motivation for putting things in the browser is to avoid doing the same things over and over again. I've worked on lots of design systems. The contribution model is a real struggle because browsers are very intimidating, and its not very clear how one contributes.... I've been fortunate enough to turn up to these people and show up and get stuff done. Other people don't have that privilege. I wonder if there is something we could do to standardize an easier way for third party contributions. Advocating for more outside contributions.
Brian: Dietrich, is there a different way to do things that is not quite standards track?
Dietrich: we need to do things at protocol labs to figure out differences in browsers, some was standards, some was just interop.
Brian: there is an aspect that is more R&D oriented, that can happen in extensions or in minor tweaks to arch, now openui is discussing a global design system. How to fund htis kind of stuff
Dietrich: you are describing social discussions, and an env where the relevant folks can make that decision. the pre-standard, is years of conversation. how do you get folks excited about this? One project I did was the web platform pipeline accelerator. We mapped out the stakeholders, all the decisions, the discussion forums, mailing list, we identified all the features that contributed a specification and implementation 3 major browsers. That has got to make it seem imsurmountable. This could be better codified.