Mozilla is building an operating system | Hacker News Reader

Mozilla is building an operating system

wiki.mozilla.org

140 pointsvilya85 comments

Impressive how the most voted comments are rants about the browser.

If that's insightful or useful comments hit me...

I really don't understand this hate towards Mozilla that has been going around lately.

Edit: Why am I being downvoted? It's a sincere opinion of a person who doesn't see a relation in critizising a project of an organisation in an article that talks about other projects.

If you dont like Firefox its fine, but it's a wonderful project with millions of users and that cannot be denied.

Sending my opinion to oblivion because of pointing how unrelated the firefox rant comments are to the article makes no sense.

Building an os is hard stuff, and if you can't produce a browser without memory leaks (or a caching strategy I can't possibly start to fathom), you're not headed in the right direction.

http://code.google.com/p/chromium/issues/detail?id=64804

http://code.google.com/p/chromium/issues/detail?id=36142

http://code.google.com/p/chromium/issues/detail?id=68304

http://code.google.com/p/chromium/issues/detail?id=90484

...etc... Guess Google shouldn't be writing an OS either?

I have three points with this post. My first is that, despite the extremely good marketing, Chrome has faults just like any other large piece of software. The point is that they have good engineers on the problems. So does Mozilla.

My second is claim is that a Firefox-based solution is good for the web. Even if it isn't as successful as Android, it will open up the web more! Android is a farce of open source (I cannot get the kernel code running on my phone. That is a blatant GPL violation. Nobody is doing anything about it. I cannot get the Android userland code running on my tablet. That is just not open source.) At the moment, Android is scarcely more open than iOS.

My last claim is that many of the "memory leaks" in Firefox are either caching--memory which is not being used is memory which is being wasted--and those which are truly memory leaks tend to be from binary extensions. The new extension SDK should mitigate this, since it is sandboxed JS code instead, from what I understand (disclaimer: i do not work on the extension API, nor do I write extensions, and I could by way off here.)

Unfortunately it's the new black to hit on Firefox/Mozilla. I fear this leads to a Google-monopoly that would be bad for everyone.

Honestly yes. Google should not be writing an OS either.

But lets be fair here, Google has waaaay more resources than Mozilla (hell, most of Mozillas resources come from Google...), and a comparison between Chrome and Firefox shows that.

There is a law of decreasing returns when it comes to "resources". That is to say, an engineer only makes so much, and throwing additional engineers at a single project only gains you so much, and the gains of each additional engineer are far less.

And the comparison between Chrome and Firefox shows... nothing? In my usage, Firefox works fine and Chrome swaps like hell and crashes even when I can get it working. Clearly other people have different experiences, but I can't reproduce them.

And you would not argue that IE6 was better than Firefox 1.0, I am sure. Microsoft had waaaay more resources than Mozilla then too. That is not indicative of failure.

"And you would not argue that IE6 was better than Firefox 1.0, I am sure. Microsoft had waaaay more resources than Mozilla then too. That is not indicative of failure."

At the time, Microsoft was not providing shit for resources to the IE team (by that time they had already creamed netscape), so really that's just another example of how insufficient resources will starve a browser project.

I'd have to trust Mozilla to keep up in the browser space before I considered using an OS developed by them.

I think it's more like "tough love" than hate. If I bitch about Firefox, it's because I want to love Firefox, but I can't. That pisses me off. If I didn't care about it, like the way I don't care about Safari or Konqueror, or Opera, I wouldn't talk about it, like I never talk about Safari or Konqueror, or Opera.

I checked your comments and you certainly aren't a regular Mozilla flamer, so I'll take what you say here at face value.

The thing that frustrates me is how many other people come out of the woodwork on any Mozilla or Firefox related post to give the story, "I used to love Firefox but it uses too much memory so I switched to Chrome and now I'm so happy". My opinion is Google has some amazing marketing and social skills to power that kind of sentiment shift. The big thing that rarely comes up in those threads is that Mozilla is a community. There are memory bugs that have been open in bugzilla for years, but there have been various people in the community working on them, sometimes in their spare time, for years as well. You look at the scope of the project, and it isn't small. It certainly isn't a hack project someone can commit a quick patch to in a day.. There are certainly paid Mozilla employees working on the code base every day, but the number doesn't compare with any of the other major browser vendors. The FirefoxFlippers™ rarely ever follow up with something like, "I tried to look at the problem and realized what a difficult thing memory management is so I took the easy path and decided to let some profit based vendor do it for me and hope that they treat me well."

My personal and probably bias experience is that I have never felt a memory issue. I tend to run a lot of add-ons and I only rarely keep a lot of tabs open because I prefer to follow a pattern of opening up a bunch of things then working my way back down to zero. I feel that most memory issues are related to bad add-ons, bad plug-ins, or bad websites, and I am very happy with the new about:memory stats that show up in Aurora. I just hope that some of the FirefoxFlippers™ who aren't just trolling for Google might take a moment to try out their easily repeatable excessive memory use-cases and drop us a line so we can work together with the community to get rid of them.

I'll now take my downmodding for ranting on a post with a comment that is only tangentially related to the topic. :/

You seem to be under the impression that the difficulty of putting out a working product is something the consumer cares about. They don't.

Results are what matters, and for all too long Mozilla's results have been poor compared to Google's.

There are a lot of things that "the consumer" doesn't care about. Some of those are things that Mozilla feels are very important and we spend time trying to show people why.

Mozilla doesn't make a browser to maximize their profits by taking as much from "the consumer" as they can get away with. If we try to compete with the other vendors on those terms, I agree we will fail.

I hope that there continue to be enough people whose goals align with Mozilla, and that together we can work on improving not only Firefox but the web. As we do that, then we will be able to help more of those uncaring consumers.

This sits among the more passive-aggressive posts I've seen in a while.

Mozilla makes a browser. It is evaluated in comparison with other browsers. It does not matter who makes those other browsers or why--and it's not like the primary contenders are "maximizing their profits" off of it anyway! The WebKit guys aren't "maximizing their profits" by writing WebKit. Google isn't "maximizing their profits" by using WebKit in Chrome. There may be some knock-on benefits for Google, but it's certainly not directly making them cartloads of money. But what they have done is written a very, very fine browser.

I was a Mozilla user well before Firefox--heck, when Mozilla Suite was EOLed I was a SeaMonkey user for a while. And then I left. Why? Because Mozilla's browser stopped being worth my time. The competition blew past XULRunner (lol) and Gecko (ouch) and Mozilla simply has not caught up. Maybe the memory issues have been fixed in recent years; I wouldn't know because the reason I won't go back now is performance and compatibility. Lofty goals that we poor benighted consumers don't "get" are nice to have, but bringing them up because your browser can't hack what WebKit can doesn't excuse it from not being able to do so.

It is notable and both sad and funny that the only time you get the "people don't understand what we're doing!" spiel is when you don't have people singing hosannahs to your greatness. If you want them, compete successfully. Don't come around whining that people switched away because the things that matter to them were inadequately addressed by your development team. Because that's solely and singularly on you guys. I want you guys to do well. I truly, honestly do. Right now? You don't compete, so I avoid Firefox. Sorry.

I was trying not to be aggressive and still make some points that were important to me. Reviewing the post, I'll concede it has a passive-aggressive tone. I apologize.

You admit to not having used Firefox in recent years, so how do you know that Firefox "doesn't compete"?

Because I have to target it. It's installed; it is not my first or second browser of choice except to make sure CSS3 and JS don't hork.

>There are memory bugs that have been open in bugzilla for years

As one of those who used to love Firefox (back when it was marketed as being "light and fast") and now has switched to Chrome, I think this is really where the hate towards Firefox comes from. Mozilla apparently makes enough money that they can continually increase the scope of their company, while at the same time years-old memory bugs are left to linger.

Firefox became popular precisely because they were the anti-microsoft and anti-netscape browser. They have succeeded in taking netscape's place as the slow bloated web browser.

I understand what you're getting all, and all I can say is... <sigh /> Actually I don't even know where to start. I am a big Firefox advocate in many ways... I've been on the Mozilla bandwagon since they were numbering releases as M1, M2, M3, etc. I think I started using Mozilla at about M2 or M3 or back in like 1999. I don't take the idea of abandoning Mozilla lightly, but by the same token... I have too much going on, and not enough time, to desire to jump into the Firefox code myself and start hacking. My perception is that it's a huge, complex code-base, that would require a lot of historical domain knowledge, and/or a huge effort in time to understand, before one could make a meaningful contribution.

Coupled with the attitude of a lot of the top Mozilla leadership have demonstrated over the years (that is, arrogant, dismissive, abrasive, and unwelcoming), I don't feel much desire to jump in and help, as much as I'd otherwise like to.

Then Chrome comes along, and - for whatever reason - doesn't seem to have the kinds of memory issues that Firefox has become notorious for. It's awfully compelling to say "let me just use this thing that just works, as opposed to fighting with Firefox."

My personal and probably bias experience is that I have never felt a memory issue. I tend to run a lot of add-ons and I only rarely keep a lot of tabs open because I prefer to follow a pattern of opening up a bunch of things then working my way back down to zero.

To be fair, this is part of what makes this difficult for the FF team. There are always people like you who say they never see these issues, and there are always people who do see them, and are so frustrated with them that they give up. And there doesn't seem to be much rime or reason to the whole thing. I've heard the "bad plugins" explanation before, I've heard "you don't understand the output from top," I've heard "it's not a memory leak, that's by design... it's the way we cache images," etc., etc. But at the end of the day, for some people, in some releases, Firefox has consistently had a problem with sucking up, and holding onto, memory like it was going out of style.

For a while, when the going word was "it's memory fragmentation that's causing the problem" they did seem to make a lot of progress. I remember a bunch of FF 3.x builds over the span of a few months, where the problems did seem to go away (or at least mostly). But somehow they always seem to come back eventually. <sigh />

I dunno... I'm rooting for the Firefox devs to get this figured out for once and all one day. And if they do, I'd probably switch back to Firefox. But to get back to the main point of this whole discussion... I'd rather see them spending their cycles on this issue, than adding features, building an OS, and all this other stuff.

I hope I made it clear enough in my first reply that while your post was the one I chose to thread off of, I do believe that you weren't griping just for the sake of it and I take your interest in the Mozilla mission and frustration with the memory issue as sincere.

I believe there are lots of reasons that different people run into a memory issue, and I believe that while the Firefox devs have tackled some of them, there are plenty of other opportunities. I hear what you are saying about the difficulty in communication, although I would hedge that just a bit by saying that a lot of the abrasive communication isn't from true "leadership", and even the leadership people can get on edge from having to deal with too many people complaining and not contributing when they see the corp of volunteers working as hard as they can. :)

I'll reiterate, given my specialization in metrics, I love the new about:memory feature and also the Telemetry project. These two projects will give us useful tools to be able to accurately measure memory consumption on a variety of platforms for users who feel the memory problems as well as those who don't.

If you have some time, try out a nightly or aurora. If nightly, then opting in to Telemetry will submit the data to us. If aurora, then you can visit about:memory and either submit a bug or feedback with what you find. I and the firefox devs would greatly appreciate it.

Cool, I'll look into that. I've been doing some "have two different browsers running at once" stuff lately, so I could test being logged into something as two different users... so that's given me an excuse to at least fire up one Firefox window. I'll look into grabbing a nightly and play with that little. Maybe I'll even gradually start using it more if it's stable enough.

Well, I wouldn't trust the nightly to be very stable. I'm sure there will be oopsies (internal Firefox joke) to be had on occasion there.

I've been running Aurora since Firefox 5.0 was released though, and I haven't had a crash yet and it has been smooth sailing. YMMV of course.

Honestly?

Because Firefox performance is atrocious, and that is the single most salient point from the Mozilla project.

Have any evidence that Firefox performance is "atrocious"? Firefox JavaScript is faster than any other browser except Chrome (and beats Chrome on some apps like JSLint), and it has GPU acceleration while Chrome does not.

It runs like a lame dog, in total.

No, I don't have benchmarks. I was astonished at the relative speed Chrome had a good while ago, and I switched over.

It might be as simple as good UX engineering to enhance perception of performance while other things happen in the background. I have no idea. But on {Win|Mac|Linux}, Chrome just zooms.

I am in complete agreement. And quite honestly it grows tiring actually refuting the various anti-Firefox rants with actual fact -- it is a bit of a hobbyhorse of some people while they cheer on whatever team they're waving the flag of.

It honestly must be emotionally exhausting to work for Mozilla projects. For all of the incredible work they have done (Firefox is the reason Apple still exists...I'll argue that point if anyone would like), there are so many anti-Firefox fanboys who appear in every single Mozilla discussion to express their ridiculous anecdotes or to make uninformed, naive performance observations.

I wonder what your theory is on how these "anti-Firefox fanbois" are created?

I used to be a "Pro-Firefox fanboi", had the apparel and everything. Now I no longer use it and on occasion like to publicly comment about my new dislike of it.

Since the facts are apparently not on my side, what is going on here?

Firefox was once the "I'm ahead of you" browser of choice for people who need to be propped up by their choices (of software or hardware). Eventually it got popular enough that some subsection of those people had to move on, generally to Chrome. Eventually the same nonsense will happen with Chrome (the loudmouths deriding it publicly for being yesterday's news).

So, essentially, browser hipsters.

Cool theory, except my bicycle has a respectable number of gears and I think Lady Gaga is pretty good.

Face it, people bitch about Firefox because they legitimately had bad user eperiances. Your attitude doesn't suprize me though, it seems to be one held by Mozilla devs too.

I am a Mozilla dev. I certainly don't hold the attitude that there's something wrong with you if you don't like Firefox. Rather, I'm interested to hear about your bad user experience so that we can improve.

I don't have anything to say that hasn't been said thousands of times already. There should be no mystery to what is bugging so many people.

I wish they'd concentrate on making a browser that does not need to be restarted after every hour or 2 of browsing lest it become more and more unresponsive.

I wish they'd concentrate on making a browser that can start scrolling before all the Javascript on the page has finished running.

(I'm using FF 7 on Snow Leopard with 1 gig of RAM. Would I have to restart FF less often if I had more RAM?)

I think you need minimum 2 GB of RAM, ideally 4GB+. More than 4-5 tabs on 1G machine will deteriorate performance very easily.

I find this a bit silly. I could understand, maybe, if I had four tabs open with Flash going in each one, or perhaps a very JS-intensive UI. But if I'm doing mostly "normal" browsing and my browser needs for than a gig of RAM to operate normally, I'd say something is wrong.

Running FF5 on Win7, 2 GB of RAM. I leave it open all day with no issues. 5+ tabs at all times.

You have caused me to resolve to buy more RAM. Thanks.

1 GB is pretty meager for general use these days.

To me, it's clear that they have spent more time making a rendering engine that is more compatible with CSS and Javascript. Every other popular browser(including Opera) has some serious Javascript and CSS rendering issues where it counts. From my development experience Firefox is the most developer friendly browser, by miles.

I wish they'd concentrate on making a browser that does not need to be restarted after every hour or 2 of browsing lest it become more and more unresponsive.

Amen, brother. I have been a Mozilla / Firefox / Seamonkey / etc. fanboy for years, but even I finally gave up and moved to using Chrome primarily, and it was mainly due to that very problem. For years now they've battled memory issues (I say "issues" since they continue to insist that there aren't any "leaks" per-se) that kill performance... one release will work great, then the next release, it's back to more of the same... next release, things get better... another release, and it's time for a RAM upgrade again... it's like they seem to be totally incapable of truly, finally, getting a handle on the memory usage situation. :-(

Firefox has made memory overhead a serious priority in the Fx7 & Fx8 development cycles, with some impressive results:

https://bugzilla.mozilla.org/show_bug.cgi?id=666058#c31

Too little, too late. It took them literally years to admit that they had a problem. I've long since migrated, and I know many other people that have as well.

Maybe for you.

For me, I try Chrome out very often but keep coming back to Firefox because of the fantastic extensions, and because it just feels right in a way that Chrome doesn't.

But that's OK: I'm assuming you've migrated to Chrome, and that's a very good thing. The competition between Firefox and Chrome has only served to make both browsers better.

Two open source browsers with a significant user base? In 2001, we only could've hoped to be so lucky ;)

FF 7.0a2 (got Jul 24 from the Aurora channel) is less responsive than FF 3.6 was on my machine -- although I cannot be sure that the unresponsiveness is due to memory pressure rather than, e.g., insufficient parallelism of the UI relative to the part that executes Javascript.

I think you can help them in this -- they're starting to gather UI responsiveness metrics. If you're seeing especially bad UI lag, they'd probably like you to run those metrics to help them figure out what is going on.

http://blog.mozilla.com/ted/2011/06/27/measuring-ui-responsi...

Possibly disable hardware acceleration.

Disable: Google Updater Google Toolbar FlashBlock if you have NoScript

Disabling hardware acceleration (it's in the General tab of the Advanced pref pane) seems to have solved the horrible problems I was having with Firefox 7. Thanks!

I am not on snow leopard yet so I dont know how much 1gig of RAM is by Snow Leopard standards but for reference how does Chrome / Safari do in comparison?

I haven't used Chrome enough to say because if you increase the text size on Chrome to where I like it then the text spills over the right side of the window (at least on 13.3-inch Macbooks) with the result that one has to scroll back and forth horizontally for every line of text. I admit that my preference for large text is unusual though.

Safari was unresponsive (spinning beach ball) even more of the time than FF has been.

Well, as a user you have the option of using another browser. Though, admittedly, with 1 gig of RAM I can't think of any decent one that wont give you problems after a few hours of browsing -- maybe chrome?

As a web developer, I wish they'd concentrate on making a browser that lets you change the line-height of a button/input element. Such a simple change, yet the problem has been around since 2006, if I recall correctly.

It's not a simple change, because sites actually rely on not being able to _decrease_ the line-height. I tried allowing sites to change the line-height and had to back that out because of sites breaking. You can read the whole sordid story at https://bugzilla.mozilla.org/show_bug.cgi?id=349259

Note that the change to allow changing line-height specifically on <button> did stick, and will ship in the next Firefox release in 3 weeks or so.

Thanks for the reply.

My opinion is that relying on line-height being unable to decrease a height of an element is a really bad idea and the owners of those sites should just be forced to fix it. Of course, I understand how from your perspective making a change that breaks a site like youtube (and probably a decent percentage of sites in general) is probably not an acceptable solution.

Oh well, I personally only needed line-heights on button elements, so the mentioned change will actually solve my UI problems. Thank you.

Yeah, if I could force site owners to fix their sites my job would be way easier. ;)

As it is, if no browser allows it, then sites will depend on not being able to do it, and any browser that does allow it will be perceived as buggy (adding to the resistance to fix the problem on the site's end).

Try disabling all add-ons, it got me to really low RAM usage and I had AdBlock & Firebug only earlier.

Will disabling Flashblock improve performance?

(The only other extension I use is Adblock plus.)

Should I disable the plug-ins the come with FF?

Np plugins ship with Firefox.

I wish they'd concentrate on making a browser that can start scrolling before all the Javascript on the page has finished running.

That's weird, that's what happens to me in Chrome, not Firefox.

My first question is why? Do they have a real vision other than "me too"?

Does anyone else remember when Phoenix was first released to fight bloated browsers? Now it is that bloated browser and running an OS developed around its technologies isn't too appealing to me.

Find your roots before you start off on some other project if you hope to have any significant success.

What is your definition of bloat - size of download, application startup times, number of features, speed while being used, memory usage?

Firefox 5 without addons does not seem to be bloated to me by any of the above definitions. In terms of featuresets just for reference - Netscape communicator used to ship with a mail client, browser, news client, html editor and a calendar.

not so much the size of the download, but all of the rest, yes.

I feel that Firefox is nearly at the point Netscape was at the point Phoenix (the original name of Firefox) came into being.

5.0 does seem a little faster.

For what it's worth, I still run Firefox (5.0) on a regular basis, including as I post this comment.

How can "Firefox be nearly at the phone of Netscape" when it doesn't ship with a mail client, irc client or any of the other related stuff that no one needed? People hear words like bloat and jump on the bandwagon. That's disappointing for a site such at this. I personally see more memory usage for the same workload (number of tabs) on Chrome over Firefox. However Chrome's memory is distributed across more processes so it's easier to hide in plain sight.

That said, I'm starting to come round to the train of thought about a user being able to control what they install. Not all users require the developer tools, in the same way that not all users require the ability to read news feeds and in the very same way that some users want enhanced share options in their browsers.

Did you read the article?

The vision is "users, not the device phone vendor, should be able to control what apps they can install and run".

Note that's what being built is not an OS. What's being built is a set of APIs to allow web apps to do more of the things that only native apps can do right now, so that you're not stuck using native apps that the device vendor has a stranglehold on.

Already being discussed here: http://news.ycombinator.com/item?id=2805708

Boot To Gecko is not a project to build an operation system.

IIUC, Mozilla is planning to reuse Linux and core components of Android. If anything, I would think Boot To Gecko would be better thought of as a shell around the OS. Using process separation (underway for a while now in the Electrolysis project) bugs/leaks/crashes in Gecko should be isolated and not significantly more disruptive than in normal Firefox.

I think the major value provided by (and work required for) Boot To Gecko will be on creating open standards (in the open) for access to the device so that web platform can be a first class citizen of the device.

How well is it going to be received compared with the level of effort required? ChromeOS is pretty bad and Google has been working on it for a long time.

It seems strange to me that they announce this without even mentioning ChromeOS as if B2G was conceived in a cave.

You sir have hit the nail on the head. The article reads as is they invented something revolutionary. I do think there is great potential in browser os market and a little competition is all that's required to pave the way for the 'next big thing'! This might just do good for the ChromeOS.

The point here is not to build an OS. The point is to improve web APIs so that web apps can compete with native ones, and do it in an open fashion so that all web browsers can implement the new APIs. ChromeOS has done some of this, but neither far enough nor openly enough (what ChromeOS extensions are working on becoming cross-browser standards?).

In that case, wouldn't it be more effective to throw Mozilla's weight behind an open source mobile Operating System that actually has a chance of going mainstream ala MeeGo?

It depends on the goal. If the goal is to have an open source mobile operating system, then maybe. If the goal is to have it not matter which exact mobile operating system you use, so new mobile operating systems have an easier time entering the market.... then no.

In case it's not clear, Mozilla is not building a kernel or anything here. They're just building capabilities for web apps to be able to do whatever mobile apps can do now.

I wonder if they're planning to write it in Rust? It would be interesting to see how an OS written in a more modern language than C would turn out - especially since other attempts like Microsoft's Singularity seem to have gone quiet.

Wouldn't it make more sense to replace gnome/kde with gecko? Isn't that essentially what the win8 demo was all about? So many questions, so little cloud...

The wiki page doesn't do a good job explaining it, but what they're aiming at is mobile devices. Essentially they want a phone that replaces every native functionality with html/javascript. So a web application would be your phone dialer, a web application could access your NFC chip. All of the capabilities they're going for is stuff that's in the pipeline in W3C, they want to create an OS to accelerate the process (and create new standards as well).

You mean like webOS?

No, I meant like web applications.

Essentially they want a phone that replaces every native functionality with html/javascript.

I would love to hear more about the security model.

Browser asks the user for permission to do stuff, the really sensitive stuff is pushed back to happening on click events, the even more sensitive stuff runs through code validation. What extra security is needed than what we already have with the above?

The security model is one of the hard parts that needs working out, yes. This is just an announcement that there is a plan to work one out in case people want to participate instead of just having stuff dumped on them from on high when it's done the way Google and Apple do it.

The project is DOA unless they also go and buy spectrum. Because no one is going to allow this phone on their network.

They need to buy some 4G spectrum and then this becomes interesting. Until then its just a failed project.

I realize they can do more than one thing at once but how about getting threaded tabs into a current build of Firefox first. Coders talented enough to make an O/S certainly should be able to make threaded tabs happen.

They just hit a pretty important milestone for that.

http://felipe.wordpress.com/2011/07/20/initial-electrolysis-...

I'd like to see what Mozilla comes up with. Bring it on Mozilla!

Telephony Apps and api natively in the browser- yay!!!

Does Mozilla have any mobile patents?

I can't wait until Safari starts building its own operating system...

Will it also run Internet Explorer?

tl;dr -- Mozilla thinks it's a good idea to write an operating system in Javascript.

Reply on news.ycombinator.com