Hacker News Reader

Comment by mrweasel | Hacker News Reader

The mentioning of DoD contractors made we wonder: would the EU help fund the development of Firefox, in order to secure competition in the browser market?

Mozilla would need to restructure and move Firefox to the foundation, but MDN, Firefox and possibly Rust could be funded by donations, corporate sponsors and perhaps the EU, assuming the money could be directed correctly. It would require a management skilled in seeking out these funding options and a brutal focus on openess, but the latter IS something Mozilla knows how to do.

View on HN

I think part of this deal must involve radically simplifying the web and reflect on the EU's sponsorship of W3C via ERCIM. We've seen web tech becoming more complicated all the time, with self-acclaimed "standardization bodies" (WHATWG) acting as fig leaf for Google's take-over of the web. Now with Firefox on the brink to drop out, like all other "browser vendors" (Opera, MS) except Google before, and failing to create a new browser engine in about 10 years of development (!) with highly competent developers, it must be clear to everyone there's something very wrong with trusting the web as the universal digital medium for accessing nearly every piece of information. It's also becoming extremely hard to find pages that used to be ubiquitous only 2 years ago, so we're heading into an information crisis (on top of information warfare). W3C may have acted in good faith, but in the end contributed to irresponsible CSS complexity (which, in turn, is the result of having no standing/statue and no representation but rather a "pay-as-you-go" sponsorship model). WHATWG have consistently denied including HTML/CSS profiles for simplified subsets (as eg suggested by MS). Time to stop playing this naive web march of progress game only benefitting very few US companies, at the expense of everybody else.

How would you "radically simplify" the web without causing breakage?

Couple ideas:

- specify profiles that can be targetted by web developers and browser developers; demand web assets to be delivered in sane and portable formats (rather than a mish-mash of JSON configs and plugins)

- in public tenders, demand that web sites for public information don't become apps: announce sites requiring JavaScript as application/html

- promote, invest into new web frameworks for apps (not sites) that abstract/isolate from HTML+CSS as underlying rendering mechanisms such that these can run stand-alone (like applets used to); I kindof like React, but unfortunately it ties apps to browser runtimes which isn't adequate for apps most of the time IMO

- transfer specs to standardization bodies with fair representation and transparency

- push for energy-efficiency; require that sites with not that great information density can render on easily-produced devices before the end of times

- push distributed web tech

Going for complete backward-compat is impossible, but humans used to do just fine without today's web of shit. HTML is rooted in markup technology which had highly organized content representations since the late 1960s.

> announce sites requiring JavaScript as application/html

You assume sites would deliberately make their sites stop working in browsers because you don't like JS? This is not going to happen. What would be the incentive?

Don't know; a noscript browser that doesn't suck?

> - in public tenders, demand that web sites for public information don't become apps: announce sites requiring JavaScript as application/html

That would be bold. Unfortunately I dont see any lobbyist being motivated to inform policy makers accordingly and no politican himself having the knowledge and cochones for this step.

Start with a markdown-optimized browser that makes it easier to publish new static content with minimal attack surface.

Let's imagine a browser (UA) that supports basics of HTML (language portion of HTML5) and CSS (e.g. v 2.1).

Yet it has WebVM instead of JS. WebVM is a virtual machine similar to JavaVM but with DOM, CSSOM and layout APIs exposed in it as a runtime.

So if some site needs flexbox in its CSS then it can be included as:

    <link href=".../css-flexbox.bytecodes" rel="extension">

If JS then

    <link href=".../es6-compiler.bytecodes" rel="extension">

and so on.

This would allow a) keep browser small and maintainable and b) open room for innovations in Web technology beyond browser vendors and comities.

As an example: TypeScript can be compiled to WebVM bytecodes instead of JS.

Not bad an idea at all! Hadn't considered bytecodes for layout algorithms, though I guess it would work with regular js as well, and I AFAICS Project Houdini is about providing the necessary core rendering API.

> would work with regular js as well

In principle JS may work but it is too slow for that. Function calls are too pricey, no integer arithmetic, etc.

At least in my Sciter (https://sciter.com) element's layout controller must have three methods:

    void calc_min_maxes() // calculates min-content and max-content
    int layout_width(int px_new_width);
    int layout_height(int px_new_height);

so it is quite easy to add new layout modes. This way I've added ()

    flow:grid(...); // a la display:grid
    flow:horizontal; // a la display:flexbox  
    flow:stack; // no alternatives in standard CSS
    etc.

Modern Java (and C# for that matter) are almost ideal for such things, especially considering modern effective JITs and VMs.

What is needed in some ability to identify the extension, so that the user may substitute their own, or may want to cache it instead. This would also improve speed, since it can avoid download it for each separate web page, and can also substitute a native code implementation.

I don't think it's fair to make Google particularly responsible for WHATWG and the over complicated web. The other vendors were on board and used the sentiment against XHTML from those who directly did write-once HTML, right before the web out grew their manual skill level and became capable of offering tooling. The results have naturally been a hodge-podge since WHATWG was about using the sentiments of everyone against a standard, and the results are everything in every direction except the track the W3C standard was on.

I'm all stoked for publicly funded open-source software, but somehow The EU and most governments seem to be pretty bad at doing this. I recall numerous initiatives that aimed to create open-source ecosystems with massive amounts of public money (billions of euros for some projects) in the past, each and every one one of them ended up as a big disaster.

The latest attempt at publicly funded software is Gaia-X, the EU's (actually mostly France and Germany's) plan to build a European data infrastructure. I really don't want to be negative but I have a really strong feeling that it will fail as well (of course not before all the money has disappeared).

It would actually be cool to see what a organization like Mozilla could do with a sizeable public grant, they can't possibly spend that money worse than it's being spent now.

Why would the EU give even one cent where the C level execs and board members collect cushy pay-cheques and fire the "hard" skilled people?

Who appoints the board of Mozilla? What will it take for them to be replaced? Can they even be replaced? It's apparent they're protecting a CEO who in any other organisation would have been canned years ago.

https://news.ycombinator.com/item?id=24164834

> In the 2+ years I've been at Mozilla, I think there's been almost complete turnover at the C-level.

Interesting. What was the reason for the turnover? Who evaluates their performance?

Yeah for some time I had high hopes for FiWare but it turned out it was just subsidies for the tendering companies without ANY real incentive for a sustainable business model past the funding period.

I was responsible for securing the funding for such tasks from the EU for Mozilla. This can work, e.g. [https://browser.mt/]. However, the latest layoffs have basically gutted this plan.

Is there any chance of the EU funding a Firefox fork that is optimized for integration with other EU initiatives, e.g. data sovereignty and individual data ownership, https://www.technologyreview.com/2020/08/11/1006555/eu-data-...? If a fork is too big a bridge to cross, how about extending the roadmap of an existing-but-neglected Web Standard, contributing reference implementation code to Firefox?

Mozilla wasn't built in a day, and neither will successors. Step one is to get funding for one person with a forward-looking vision and experience with the non-public history that got Firefox to this point. There is a narrow window to preserve a subset of the brain trust for core features (Servo?) before developers get scattered to the winds of other companies.

Finally, if Firefox will eventually lose to Chrome, perhaps we need to step back and look at core protocols (like the Gopher reboot) and new security models (e.g. running each browser tab in a hardware-isolated virtual machine). Extensibility was a core offering of Firefox. There may be new optima on the trade-off chart of security vs. power, when running on newer hardware.

As to a fork, doubtful. It’s too large a project.

Generally the EU wants to fund open projects not companies.

So funding the implementation of a Web Standard in a Firefox, say, is “out”, but implementation of a browser extension is “in” as it’s (more-or-less) browser independent.

This would likely require the bulk of Mozilla to be based in Europe. It could be achieved for sure, but it would take years to transfer the relevant knowledge. I don't expect MoFo and MoCo leadership to be willing to entertain such possibilities, considering how "valley" their outlook seems to be.

>but it would take years to transfer the relevant knowledge.

A) The people with the knowledge just have been fired and are free to move and to be hired

B) Thanks to corona, remote work is a fully accepted option

Something new can be running next week or even tomorrow.

I think remote work is missing the point here. It's politically difficult for governments to spend their tax revenue by paying a largely foreign workforce.

Not just politically difficult, it's also generally unwise. In the event of deteriorating relationships, you'll suddenly lose access to what you've been funding, because the US would declare it "a matter of national security" and outlaw the export. You can still work with the sources you have, but it'd be hard to keep it running smoothly. Much better to not have that risk if you're going to spend billions on it.

People don't move abroad easily, particularly very experienced ones with the historical and institutional knowledge that is really key for a project the size of Firefox - uprooting their families to a foreign (and likely not even English-speaking) country, is a difficult choice. This goes double for Bay Area / SV people used to massive salaries that are hard to justify and achieve elsewhere, and people moving to Europe where immigration bureaucracy is not trivial. Would you take a pay cut and force your family to emigrate just to keep working on a browser, when you've probably plenty of easier and more rewarding options?

As for remote, the political support to fund a project effectively based in the US would be zero.

You also forget that EU institutions don't typically enjoy the freedom to spend a few hundred millions on the spot - they have processes and rules and parliamentary oversight to ensure agreement between 27 governments. It's not like the US president cutting a check, more like the US Congress approving funds - to get any significant amount, it would take some time.

I'm not saying it's impossible, but it would realistically be "next year" (at least) rather than "next week".

It makes so much sense for the EU to start investing in an open web and open source software. A relatively cheap option to move towards a society with better data protection and online civil liberties.

I’ve never took the time to really dig into the details of what the EU is exactly doing right now, but it seems regulation is higher on the list than investment and innovation.

They've actually funded quite a few projects through their Next-Generation Internet initiative [1] [2]. Relatively small grants, but many, and available without too much red tape, which I think is a huge boon in getting projects off the ground - if just a few of those are a success, so is NGI.

(Disclosure: I'm the recipient of one of those grants.)

[1] https://www.ngi.eu/ [2] https://mastodon.xyz/@NGIZero

EU research grants have an incredibly high documentation burden and Europe in general still has a problem considering software development as research. It's usually everything that leads up to the development and then the development itself is mostly an afterthought.

The fine European-academic tradition of considering actual development as grunt work that is an unworthy thing to do for the great minds. You can see where it got the European software industry as compared to the US. It is endlessly frustrating.

"Software Engineering" in Europe is the other kind of bad, learning methodology without knowing what it is applied to, and learning some "enterprise grade" rubbish.

It is as if no one who has anything to say in academia here knows how good software is actually created. And that is probably the case.

</rant>

Depends on the funding program nature, and grant structure (eg is there a research consortium). Not necessarily that high.

Reply on news.ycombinator.com