Maintaining an Independent Browser Is Expensive

I think you're focusing on the costs and neglecting many of the benefits. For example, just think about how much easier it is to learn to code nowadays, thanks to the proliferation of helpful websites that take a variety of approaches to teaching, many of which were enabled by web technologies. GitHub, JavaScript 30, Khan Academy, RunKit, Flexbox Defense, Codecademy, just to name a few. And think about how many more of these sites exist today than would have in the past, because more people are building them, they're much easier to build (again, thanks to new web technologies).

And this is just learning to code.

"The complexity of the modern Web platform is a problem. Mozilla developers feel this acutely, believe me."

I believe you. Do they want to take action to address it?

"The problem is that if the open Web platform does not expand to meet the needs of modern applications, then modern applications will simply be restricted to single-vendor platforms like iOS, Android and Windows, and over time the relevance of the open Web will atrophy. That is not an acceptable outcome for Mozilla."

I understand. Why does the "open Web" need to stay relevant? Honest question. The answer to this question is really the core issue, IMO.

"If you can convince Web developers to build sites that work on cut-down browsers, great. I don't see any way to make that happen en masse though."

Not sure where the "en masse" part comes from. That was not in the original comment.

The comment was meant to draw attention to denial of choice in user agents given to users. The illusory scarcity of browsers that will "work". (Thus web developers design for browser implementations instead of according to open web standards, which should be implementation-agnostic. Browser developers are the ones who are seemingly in control of what is and what is not a "standard" in the mind of the web developer. Too often, companies are the ones writing the "open web" standards.)

I believe in the antithesis of the "en masse" notion.

That users do not need to converge en masse around a small selection of known, complex web browsers because most sites, e.g. those that simply present information, already "work" with cut-down browsers.

I use a non-graphical user agent. Most sites work fine for me. I get what I need. And this is without the web developer even contemplating the software I am using. To the extent they are following certain open standards for HTML, it all works anyway.

On the same day as the OP was posted, it shared the HN front page with a re-post of a 2012 ACM submission by PHK.

In this ACM submission, PHK describes building Firefox on FreeBSD.

He cites something like 122 dependencies, and a requirement for some binary to run plugins.

I have build Firefox on BSD myself using pkgsrc. It takes longer than building kernels.

Is there a "Firefox Lite" where a user can opt out of various features at build time?

Why not?

"FWIW Mozilla obtaining a monopoly would be inconsistent with their mission and I don't think actual Mozilla developers are aiming for that. Partly because it's not a realistic outcome!"

Nor am I suggesting that providing options to users should result in any "en masse" behaviour by users or web developers.

It is not a realistic outcome.

The goal of the experiment might be to provide some software to users that the market share leaders will not provide. Is this realistic? Mozilla has taken risks before. Not every project it has sponsored has attracted a mass following of users.

For example, imagine a version of Firefox that optionally has no Javascript engine. This software would be less complex (and intentionally less functional). But believe me, as a text-only browser user, it would still work.

If a user complains about a major browser as "bloated", if she complains about ads and is forced to use NoScript or an ad-blocker, if she is pondering how complexity makes her browser more vulnerable to attacks, then it would be interesting if there was another option she might try, which came from a well-known source such as Mozilla. She might build Firefox without the features that lead to these problems.

What happens after that is anyone's guess. As it is with any experimental Mozilla project.

But at least one can say that an option was presented to move away a state of from increasing complexity in addition to the option of using Mozilla's full-featured Firefox browser competes for market share with the other major browser vendors.

Internet commentators sometimes reference a quote from Steve Jobs something like: users do not know what they want (until Apple gives it to them).

While this may be self-evident to developers, I believe that users do know what they do not want. These are programs features, modifications and behaviours that users are familiar with because they have been using the software for years. Alas, rarely do developers of graphical programs give users the option to remove features. As such, we can only guess what might happen if they had that option.

A large portion of the web does "work" when using user-agents that lack the latest features. It also works with the recommended browsers. The question is whether Mozilla can acknowledge this is true and release software that takes advantage of it. Mozilla can still pursue its mission, including keeping up with the Joneses.

Why would you want your published version to exclude up to a years worth of fixes for known issues?

Thanks for the thoughtful comment.

> Do they want to take action to address it?

Within the constraints of the assumptions I've outlined, yes, and they have.

One example is the fight for asm.js/Webassembly against PNaCl+Pepper. We fought that fight in part because Google was fine with introducing a whole new platform API in the form of Pepper, and Mozilla instead wanted to minimize the additional complexity by reusing existing Web platform APIs.

> Why does the "open Web" need to stay relevant?

I want a platform that isn't controlled by any single gatekeeper, through which app/content developers can reach all users and all users can reach the apps and content they want. (A sort of corollary is that the platform should be implementable in free software.) The Web is by far the closest thing we have to that.

One problem with a niche browser is that when you have a browser with very low market share, Web developers don't test against it, and with the state of software development technology today, stuff that's not tested in a platform implementation tends to not work on that platform implementation.

Another problem with a niche browser is that by definition it impacts a small number of users, and therefore only a relatively small investment can be justified.

Fortunately there is nothing stopping you or some like-minded group from doing your experiment. Tor has already done something similar and that's gone quite well.

I have been discussing the debt-based economy and Google and Mozilla recently with BrendanEich on Twitter, including a debt diagram.