[jholly]

Locked out of iOS. Under heavly competition on Android. Put the metro port on the back burner. Increasingly fighting to stay on feature parity with Chrome on the desktop. FirefoxOS still to make an impact.

I love Firefox. I'm just worried for it. With the world going mobile, I wonder if there's space for Firefox. It would be such a shame not to see it around 5 years from now.

I wonder if it makes sense for a small organization to spend so much time and money on a custom rendering engine. Is it such a horrible idea to fork and contribute to blink? Just thinking out loud. Mostly out of despair.

[magicalist]

> Is it such a horrible idea to fork and contribute to blink? Just thinking out loud

A fairly bad one, yes. There would be virtually no benefit to doing this, and a serious detriment to web browser diversity (which would likely be a detriment to the health of the web as a developing ecosystem).

> It would be such a shame not to see it around 5 years from now

They still have plenty of revenue, and the amount still keeps going up every time they renegotiate their search engine contracts. They'd be more in danger of irrelevance, but while they had a slight dip in browser marketshare (depending on who you ask), it seems relatively steady state now (just like their major competitors). Meanwhile no one is wringing their hands over Microsoft's irrelevance in the browser market even though Windows Phone remains a tiny contender.

Your worries seem a little unfounded.

[jholly]

Microsoft isn't a fair comparison. They have other, multi-billion dollar cash cows to rely on.

I don't think they're unfounded. Nearly every developer I know uses chrome, and chrome dev tools. This wasn't always the case. Market share can be a lagging indicator, i think it's a little short sighted to suggest they're okay. They're not!

Forking isn't a sin. Diversity often a result of forking. Take a look at this graphic[1]. There's a good chance the browser you're currently using is a result of a fork.

Mozilla has to really examine what they stand for. The open web doesn't depend on implementation internals. Just specs, tests, and sensible governing body, and a set of willing participants that don't have unreasonable misaligned interests. They can still promote an open. Perhaps even do a better job of it with the extra resources.

[1] http://upload.wikimedia.org/wikipedia/commons/7/74/Timeline_...

[pcwalton]

> Forking isn't a sin. Diversity often a result of forking. Take a look at this graphic[1]. There's a good chance the browser you're currently using is a result of a fork.

Sorry, I think there's absolutely no way to spin the idea that abandoning Gecko for Blink would increase diversity. It's self-evidently false.

Forks lead to more rendering engine diversity than straight-up reskins, I suppose, but much less diversity than actual from-scratch engines. (Incidentally, since Blink is explicitly not designed to be embeddable, "adopting Blink" essentially means becoming a Chrome skin.)

[JohnTHaller]

Firefox has equal share with Chrome and most of the techies I know still use Firefox. Not having every private test URL of stuff you're working on sent to Google is reason enough. I do have several dev friends that use Chrome. They're usually mobile devs unconcerned with privacy (they have everything in Google).

[camus2]

    > Firefox has equal share with Chrome and most of the           
    >techies 
    >I know still use Firefox. 
    > Not having every private test URL of stuff 
    >you're working on sent to Google is reason enough. 

And where does Mozilla money come from ? tell me.

[JohnTHaller]

Firefox has a separate search box from the URL box for a reason. Every keystroke of the search box is sent to Google so the auto-complete results come up. Chrome has a single box for both URLs and search, so every character of every URL you type is sent to Google. That's the difference.

[throwawaay]

Put up or shut up:

Please provide evidence that Chrome sends private URLs to Google.

[Curmudgel]

If you have omnibox predictions turned on, every keystroke is logged.

https://support.google.com/chrome/answer/180655?hl=en

[pekk]

What is unhealthy about shared components? Should we insist that browsers also be built with different languages or compilers or it is unhealthy? We are not talking about a proprietary component here.

There may be reasons to prefer Firefox not to use blink, but "detriment to the health of the web" is a really vague and pointless argument. What do you really mean by that, if not just that Chromium/Chrome are somehow bad for the web?

[magicalist]

> There may be reasons to prefer Firefox not to use blink, but "detriment to the health of the web" is a really vague and pointless argument. What do you really mean by that, if not just that Chromium/Chrome are somehow bad for the web?

Er, huh? I don't think my phrasing was that confusing: I said that the loss of diversity in implementations would likely lead to a loss to the future development of the web. Multiple implementations of a spec ensure that the spec matches reality and often make for better specs, as feedback from implementations over very different foundations bring out fundamental architectural and performance issues that may not have been obvious otherwise (e.g. you write a feature so that it's efficient on platform X, it means that in 5 years someone implementing a new browser will have to write it very much like X to get decent performance).

This is actually enshrined in the W3C technical report development process: for a specification to become a Proposed Recommendation, "the Working Group should be able to demonstrate two interoperable implementations of each feature"[1].

There were lots of arguments about this when Opera stopped developing Presto to move to Webkit. I think Chromium/Chrome have been, on the whole, great for the web.

[1] http://www.w3.org/2005/10/Process-20051014/tr#cfr

[bzbarsky]

> Should we insist that browsers also be built with different languages

In a perfect world, yes. Then you might not get standards written entirely around ease of implementation in C++, leading to crazy behavior just because it happened to be the simple thing to do in that language.

> if not just that Chromium/Chrome are somehow bad for the web?

What's bad for the web is monoculture, because then people start building sites that rely on quirks of that monoculture that are not codified in a standard. And then if something better comes along it's a long uphill battle to get that something better adopted, because it has to duplicate those quirks.

For example, there is no way something like Servo would be able to happen if we had a Blink monoculture on the web, and I seriously doubt that Blink will be able to parallelize to the extent that Servo is planning to and that I think the web needs in the future.

For another example, once you have an engine monoculture, the developers of that engine become gatekeepers for what's possible on the web. We saw that once already with IE. While Google is not likely to shut down development of Chrome in the near future, they have their own priorities (e.g. getting video DRM implemented is a lot more important to them than anti-tracking countermeasures), so unless you think those priorities are perfect for the web you should be against them having a stranglehold over what ends up in the web platform. Just like you should be against any other entity having such a stranglehold. It was bad when Microsoft was there, it would be bad if Google were there, and it would also be bad if Mozilla were there.

[justinreeves]

jholly's post reads like a concern troll.

[zanny]

> Locked out of iOS.

So is Chrome, at least in the sense the iOS Chrome is slow as hell because Safari gets special privilege. But that is the case with all iOS software - you are beholden to Apple. I wouldn't consider success or failure there something you can control at all.

> Under heavly competition on Android.

True, but mobile FF has the best UI out of all of them, I think.

> Put the metro port on the back burner.

Because nobody used it. At all. It is senseless to make and maintain a product nobody will use. All this says is that Metro is systemically panned.

> Increasingly fighting to stay on feature parity with Chrome on the desktop.

Unless you mean things like MSE, I don't see where FF comes up short. Since version ~8 their engine has been on par or faster than Chrome and they support emerging web standards like RTC overnight.

> FirefoxOS still to make an impact.

It isn't targeting the first world, it will never make an impact here - it is a sound decision, it would be impossible to take Android on directly, and they really don't have to, since FF on Android has their marketplace. It is getting sizable deployment contracts in India and such, so I'd call that a success.

> I wonder if there's space for Firefox.

Of course. It is the only major web player pushing openness without an agenda (like Google).

> It would be such a shame not to see it around 5 years from now.

Even if Mozilla the company were a shell of its former self, I imagine the community would pick up a lot of the slack to keep Firefox at least usable, if not bleeding edge, like with how Thunderbird is doing.

> I wonder if it makes sense for a small organization to spend so much time and money on a custom rendering engine. Is it such a horrible idea to fork and contribute to blink? Just thinking out loud. Mostly out of despair.

Does it make sense to have half a dozen python interpreters? Or 3+ binary cross platform compilers? How about how many GUI toolkits are out there, from Wxwidgets and EFL to more popular cross platform ones like Qt or Mono.

And they are developing Servo as the solution to Gecko being really old at this point. And Rust is an amazing language, and I'd definitely look to develop in it as soon as it hits a 1.0 (probably when Servo is ready).

[pcwalton]

> And they are developing Servo as the solution to Gecko being really old at this point. And Rust is an amazing language, and I'd definitely look to develop in it as soon as it hits a 1.0 (probably when Servo is ready).

Thanks for the kind words :)

We expect Rust to be at 1.0 before Servo becomes a production browser engine.