Taking It Back: Why This Xoogler Is Joining Mozilla(medium.com)
medium.com
Taking It Back: Why This Xoogler Is Joining Mozilla
https://medium.com/mozilla-tech/taking-it-back-why-this-xoogler-is-joining-mozilla-5c2a5d2cab63#.jyzt6hmwq
47 comments
Not all of Mozilla uses the MPL; Rust uses MIT/Apache 2.0, for example.
So much fluff. Long gone are the days when Mozilla wanted to build the best browser. You simply have to use the browser everyday to know where they are putting in a their efforts.. it is certainly not in the browser.
Rust and servo have noble visions but are irrelevant to the future of the web. UX is what drives browser market share and unless Mozilla refocuses on this, they will contjnue to lose maket share. TBH, it will take years if they start now to convince back all the people they lost.
Rust and servo have noble visions but are irrelevant to the future of the web. UX is what drives browser market share and unless Mozilla refocuses on this, they will contjnue to lose maket share. TBH, it will take years if they start now to convince back all the people they lost.
I don't agree. In my opinion, even on my high-specification workstation, the performance of the modern web is below my expectation for 2017 (having been using the web since 1997). Browser and CPU performance have not kept pace with web site bloat. CPUs have gained more cores but they mostly sit idle when using a web browser.
So focusing on more concurrency in the web browser is extremely compelling in my view. I don't believe we're going to see widespread trimming of bloat from the web at large. It's something a few conscientious companies think about, but the majority just keep throwing frameworks and third-party scripts at their sites with abandon.
Web performance in 2017, even on high-capacity hardware, ranges from acceptable to bad. I'd prefer it ranged from superb to acceptable. If Servo moves us in that direction, I am a big fan.
Perhaps to your point: UX includes performance. And in that sense, UX may help Mozilla market share.
So focusing on more concurrency in the web browser is extremely compelling in my view. I don't believe we're going to see widespread trimming of bloat from the web at large. It's something a few conscientious companies think about, but the majority just keep throwing frameworks and third-party scripts at their sites with abandon.
Web performance in 2017, even on high-capacity hardware, ranges from acceptable to bad. I'd prefer it ranged from superb to acceptable. If Servo moves us in that direction, I am a big fan.
Perhaps to your point: UX includes performance. And in that sense, UX may help Mozilla market share.
I am absolutely flabbergasted by how every fully featured browser is capable of bringing my modern workstation to its knees. Late last year I took a screenshot of a 32 kb text document running in a tab managed by chrome process with no other tabs to take care of occupying over 1 GB of ram. And when I say text document, I mean a file with a .txt extension with no javascript or anything else going on. The world is wrong!
The management process, every tab, every extension, and a pool for registered offline workers are all running by firing up chrome... I'd presume there's similar overhead for FF, but not process isolated, that was work in progress iirc.
It's not necessarily a bad thing, but can all add up, and it's hard to measuer/add up the separate processes.
It's not necessarily a bad thing, but can all add up, and it's hard to measuer/add up the separate processes.
This sounds like a bug. If you can reliably reproduce 32kb txt file taking up 1GB, would you mind filing a bug at https://crbug.com? Or I can file a bug if you let me know how to reproduce this.
Of course, UX includes performance. If you go to the landing page of a website and it loads very slowly or has janky animations.. they are screwed. It's part of the UX.
> Long gone are the days when Mozilla wanted to build the best browser.
Come on. Do you really think I come into the office on weekends just for fun?
> Rust and servo have noble visions but are irrelevant to the future of the web. UX is what drives browser market share and unless Mozilla refocuses on this, they will contjnue to lose maket share.
Fortunately, Servo is experimenting with UX as well, via browser.html!
Come on. Do you really think I come into the office on weekends just for fun?
> Rust and servo have noble visions but are irrelevant to the future of the web. UX is what drives browser market share and unless Mozilla refocuses on this, they will contjnue to lose maket share.
Fortunately, Servo is experimenting with UX as well, via browser.html!
> Come on. Do you really think I come into the office on weekends just for fun?
There is a difference between working hard and working smart. Working smart means to fix today's problems that all your users are complaining. People want to use Firefox everyday but cannot. The solution to today's Firefox performance problems is not Servo or Rust. It's basically engineers putting in time to fix it in the _current_ code base.
To be more constructive: Why don't you do a "Ask HN" about what stops them from using Firefox every day and you will have the answer. In fact, just reply to this comment and I will do this for you (It just seemed for me to do this as an "outsider").
On a personal anecode: I don't use firefox everyday for "fun" either despite having so many performance issues. I believe in Mozilla's original vision and it's sadly now like a "religion" for me - I use it despite the pain it causes me.
There is a difference between working hard and working smart. Working smart means to fix today's problems that all your users are complaining. People want to use Firefox everyday but cannot. The solution to today's Firefox performance problems is not Servo or Rust. It's basically engineers putting in time to fix it in the _current_ code base.
To be more constructive: Why don't you do a "Ask HN" about what stops them from using Firefox every day and you will have the answer. In fact, just reply to this comment and I will do this for you (It just seemed for me to do this as an "outsider").
On a personal anecode: I don't use firefox everyday for "fun" either despite having so many performance issues. I believe in Mozilla's original vision and it's sadly now like a "religion" for me - I use it despite the pain it causes me.
> The solution to today's Firefox performance problems is not Servo or Rust. It's basically engineers putting in time to fix it in the _current_ code base.
If you're so convinced that the correct answer is to hack more on Gecko (as if hundreds of engineers haven't been doing exactly that for a decade), then it should be easy for you to explain to me, in detail, technically, why Gecko's architecture is suitable to achieve best in class performance in the long term.
> To be more constructive: Why don't you do a "Ask HN" about what stops them from using Firefox every day and you will have the answer.
Don't deflect. I want you to defend the choice of continuing to hack on Gecko's C++ codebase instead of using Rust. Preferably describe specific subsystems that you think Gecko is doing better than Servo is, and what the problems you think Servo's architecture has. This should be very easy for you to do, as you seem to think it's so obviously correct that the only conclusion is that none of us even care about making the browser better.
If you're so convinced that the correct answer is to hack more on Gecko (as if hundreds of engineers haven't been doing exactly that for a decade), then it should be easy for you to explain to me, in detail, technically, why Gecko's architecture is suitable to achieve best in class performance in the long term.
> To be more constructive: Why don't you do a "Ask HN" about what stops them from using Firefox every day and you will have the answer.
Don't deflect. I want you to defend the choice of continuing to hack on Gecko's C++ codebase instead of using Rust. Preferably describe specific subsystems that you think Gecko is doing better than Servo is, and what the problems you think Servo's architecture has. This should be very easy for you to do, as you seem to think it's so obviously correct that the only conclusion is that none of us even care about making the browser better.
I'm not the original poster but I think this is dismissive of a very real problem with FireFox. As a user I've seen perceived performance regress significantly over the last few years. From the outside Servo looks like a distraction - and the results tend to confirm it.
It may well be that Gecko just is that terrible and the only way to fix it is with Servo. If that's the case you guys really need to hurry up because you're loosing significant mindshare. Your organization's predecessor has also made this mistake before very publicly in the past with the Netscape rewrite.
It may well be that Gecko just is that terrible and the only way to fix it is with Servo. If that's the case you guys really need to hurry up because you're loosing significant mindshare. Your organization's predecessor has also made this mistake before very publicly in the past with the Netscape rewrite.
> From the outside Servo looks like a distraction - and the results tend to confirm it.
Then show me those specific results.
I'll pose it to you, too. If it's so obvious, tell me what the problems with Gecko are, specifically, and how Servo cannot solve them.
> If that's the case you guys really need to hurry up because you're loosing significant mindshare.
This is the first time I've been called lazy by a random HN poster. I guess there's a first time for everything.
> Your organization's predecessor has also made this mistake before very publicly in the past with the Netscape rewrite.
Firefox was possible because of Gecko in the first place! Do you want to be using code descended from Netscape 4, which had no dynamic restyling at all?
Then show me those specific results.
I'll pose it to you, too. If it's so obvious, tell me what the problems with Gecko are, specifically, and how Servo cannot solve them.
> If that's the case you guys really need to hurry up because you're loosing significant mindshare.
This is the first time I've been called lazy by a random HN poster. I guess there's a first time for everything.
> Your organization's predecessor has also made this mistake before very publicly in the past with the Netscape rewrite.
Firefox was possible because of Gecko in the first place! Do you want to be using code descended from Netscape 4, which had no dynamic restyling at all?
With respect it's not the job of the user to be an expert in the architecture of your application. The fact you won't admit there are problems identifies a real with issue with your organization.
Firefox is slow, I don't know why. I was hoping the engineers working on it would know - it's scary that they do not.
Firefox is slow, I don't know why. I was hoping the engineers working on it would know - it's scary that they do not.
So you don't know what the performance problems in Firefox are, but you're confident that the solution has to be a short-term fix in the existing C++ codebase.
Do you realize why that does not make sense?
Do you realize why that does not make sense?
I suppose a better question is how did the existing code base get this bad? I work on a product 10 years older than FireFox and we've been able to make do with upgrades in place. Over time large pieces of architecture change.
The difference is our customers see improvements over time not large periods of stagnation.
The difference is our customers see improvements over time not large periods of stagnation.
Most of the hard problems are extra hard because of the existing codebase, and the nature of it. Servo/Rust are definitely the way to go to fix those long term issues in stability and performance.
Sometimes you DO have to step back, and rewrite, but at least you have a better idea what you need based on the prior version, even in a separate language.
Sometimes you DO have to step back, and rewrite, but at least you have a better idea what you need based on the prior version, even in a separate language.
I cannot think of any problem being solved with a total rewrite in a completely new framework and in a completely new language. That too when the issue is graphical performance and UX design. Oh well..
Best I can luck and thanks for the browser.
Oh boy. This type of rhetoric is so annoying to people who work at big companies. Do you know what pcwalton works on?
He sort of implied that he works on the browser by what he quoted, and your post just sounds like you want to talk about yourself.
He "sort of" implied that he works on Servo which is not Firefox.
I brought up my own personal experience because I've also had people conflate where you work with what you do.
I brought up my own personal experience because I've also had people conflate where you work with what you do.
> Servo which is not Firefox.
Given that Servo is now vendored inside of Firefox's tree, and Project Quantum more generally, this is not exactly an either-or these days.
Given that Servo is now vendored inside of Firefox's tree, and Project Quantum more generally, this is not exactly an either-or these days.
https://testpilot.firefox.com/ Looks like they're focusing on UX to me.
Sweet, ty did not know about this. Will happily test pilot Containers and Tab Center! (and page shot prob useful too) Just enabled them instead of tab groups and tree style. So far fit my work flow very nicely, hope they make it into main.
(unless you like tab nesting which i guess it cant do? personally just looking for the more readable left-side layout)
(unless you like tab nesting which i guess it cant do? personally just looking for the more readable left-side layout)
((ok so containers isn't exactly a 1 to 1 comparison/replacement (yet( but i do like it))
They'll eventually add Web Extension APIs that allow new-style addons to put tabs on the side. So even if they don't enable tab nesting in this first-party feature, it will be reasonably easy to add that in a future extension.
Very nice, I didn't know about this.
> it will take years
Yes.
> if they start now
We started last year with e10s, testpilot and dozens of other improvements.
Yes.
> if they start now
We started last year with e10s, testpilot and dozens of other improvements.
What's "e10s"?
I don't know if it's just my computer, but Firefox lags horribly on my Mac after a certain amount of usage whereby I always have to restart the browser because the experience after a certain amount of usage, even if there's only one or two tabs open, is unbearable. And right now, the only add-on I have on it as uBlock Origin.
Whereas Chrome can have over a dozen tabs open and it still operates well. Which is a shame because I was hoping to switch over back to Firefox.
Whereas Chrome can have over a dozen tabs open and it still operates well. Which is a shame because I was hoping to switch over back to Firefox.
Interesting. I have the exact opposite experience on Mac. Chrome becomes unusable after about 15 tabs and especially after leaving it open for more than 2-3 days. Firefox is the only decent browser I can use on my Macbook.
What version are you running? My anecdotal counterpoint is that I find Firefox excellently performant these days despite heavy use.
The upcoming Quantum work will improve speed generally and specifically for multiple core setups. Electrolysis in the latest versions will also help prevent "bad tabs" from slowing the whole browser.
The upcoming Quantum work will improve speed generally and specifically for multiple core setups. Electrolysis in the latest versions will also help prevent "bad tabs" from slowing the whole browser.
Try a new profile or a refresh.
When the "majority of revenue" comes from Google or Yahoo, what purpose does this serve?
https://en.wikipedia.org/wiki/Mozilla_Foundation#Mozilla_Cor...
I don't see how going from Google to Google-by-proxy is going to "take back the web."
https://en.wikipedia.org/wiki/Mozilla_Foundation#Mozilla_Cor...
I don't see how going from Google to Google-by-proxy is going to "take back the web."
Worth pointing out; but since you can switch to another search engine it's not a deal-breaker, quite.
The problem isn't being for profit. The problem is deriving your income from advertising. Selling the customer.
It's both. Mozilla's advertising model isn't such a big deal. It's even easy to counter in their own product. Being for profit means Google must keep trying to drive its numbers up. A likely side effect is spying on its users in new ways whose results will be sold in new ways. That prediction certainly came to pass several times over.
"Xoogler" is a new one for me. I went looking for a company called "Xoogle.com" and wondering why anyone would pick a name so close to Google.
"Ex-Googler" seems to have been serving its purpose well.
"Ex-Googler" seems to have been serving its purpose well.
Just today I decided to ditch Firefox after using it daily since the first public release.
On my 15 inch Macbook Pro with Retina screen (and a 4k monitor attached) CSS animations are cooking the CPU. I've waited 2 years for improvements but with each new version the issue persisted.
It's clearly a Firefox issue, as Safari and Chrome don't have problems with it. But now almost every website uses CSS animations, the problems only have gotten worse. I was hoping the Servo project would somehow solve my problems, but I guess that will take a couple of years to fully develop.
https://github.com/servo/servo/wiki/Roadmap
On my 15 inch Macbook Pro with Retina screen (and a 4k monitor attached) CSS animations are cooking the CPU. I've waited 2 years for improvements but with each new version the issue persisted.
It's clearly a Firefox issue, as Safari and Chrome don't have problems with it. But now almost every website uses CSS animations, the problems only have gotten worse. I was hoping the Servo project would somehow solve my problems, but I guess that will take a couple of years to fully develop.
https://github.com/servo/servo/wiki/Roadmap
I was about to get a new machine as FF was getting unusable on my 2011 macbook pro. I installed Chrome and its like I have a new computer. Web browsers are hard but clearly FireFox is missing something chrome has.
I love this sentiment! Unfortunately, Mozilla absolutely disagrees this BSD/MIT license sentiment. Instead Mozilla's MPL 2.0 license makes a broad (hardware and software) patent-grab underlined with a "retaliation clause." So few companies will be "empowered" by this license, they'll have to avoid it like the plague. History shows they'll remake the software all over again, only under a truly permissive software license. An ungodly waste.