> Why not just run the 64-bit tools on a > 64-bit platform
There is no 64-bit version of the MSVC linker, and no plans for one according to public statements from Microsoft.
One can run the 32-bit linker on a 64-bit Windows, which gives you 4GB of address space to play with. That's the medium-term solution for Mozilla, but it does require updating the entire Windows build farm to 64-bit Windows and shaking out any compat issues that result. Doable, but will take a few weeks probably.
Firefox offers various security mitigation strategies (in terms of properly dealing with various memory-safety issues, say) that Chrome does not. As far as I can tell, this study just started off with a subset of the list of techniques that Chrome implements and then "studied" which other browsers also implement them, instead of studying what browsers actually do to ensure security and how difficult they are to actually exploit.
Your larger point that modern IE is a fairly secure browser (like any modern browser) is correct, of course.
> I have never seen any difference in google search or > gmail between Firefox and Chrome
Offline gmail only works in WebKit, by design.
> For example, Mozilla did essentially (3)
Mozilla has consistently tried to convince people to author pages that work in all browsers, in my experience. But I'm not omnicient, of course.;)
The extensions thing you raise is interesting. I have to admit I'd never thought of it that way before, if only because I'd never used any such extensions. But yes, the number of users of such extensions is minimal in general.
Let's hope that you're right and in a few years my worries here will in fact seem like paranoia...
Actually, I agree with the original point of the thread. What Google is doing with Chrome is:
1) Explicitly authoring its own web properties to specifically work better with Chrome, in hopes that its high market share for things like search and webmail will increase Chrome adoption.
2) Authoring its own web properties to only work with WebKit (e.g. a number of Google's sites sniff UAs and send WebKit-only content to any mobile UA).
3) Urging authors to create Chrome-only content (and more generally, together with Apple, encouraging WebKit-only content).
4) Paying other companies to bundle Chrome with their software so that people end up using it whether they want to or not. This part Microsoft didn't have to do because they just bundled with Windows, of course.
They're not _quite_ as bad as Microsoft was because a lot of this is in fact open source (though they don't exactly take contributions much or plan to share control, so most of what you can do with the source is fork). This does mean that some other project could import some part of Chrome or V8 if desired. But given that none of this code is static (security fixes, spec changes, etc), it's not like a one-time import is useful. You have to keep importing (and hoping that Google doesn't change something you care about) or as I said fork and take over maintenance. In practice, for most situations, both options suck to a good extent. http://3.bp.blogspot.com/-GBSGBbc9UtA/TmexwLaJN9I/AAAAAAAAAKs/06OaexCT5Ms/s1600/Lead%2Bdevice%2Bconcept.png is a good look into Google's general thinking on open in this context, for what it's worth.
One other important note is that Google is a bit better at participating in the standards process than Microsoft was in 2001 (though not than Microsoft in 1998, say). That's a net plus for them.
So no, comparing what Google is doing with what Microsoft was doing with IE6 is not at all ludicrous. There are some important differences, and Google doesn't have a browser monopoly, so they have to play a bit nicer so far, but the overall strategy looks similar to me. The key to frustrating it, of course, is for them to not end up with said browser monopoly.
> It seems more likely that other engines are just > better than V8 for some types of code
Well, yes. But some optimizations are just mutually exclusive because they require different design tradeoffs.
As for the rest... Google is definitely being evil in various ways, and is in fact much like Microsoft in the late 90s, in my opinion, but what they're doing with V8 is just self-interested, not so much evil.
Google's JavaScript libraries are purposefully written to run faster on V8 specifically, often at the expense of performance in other browsers. And at the same time, V8 is written to run the code patterns those libraries use faster, often at the expense of other code patterns used elsewhere.
The net result is that Google properties work better in Chrome and at the same time Chrome works better on Google properties.
> Microsoft has a history of abuse, while you are only > weary of the domination of Google.
More precisely, Microsoft has a history of abuse and has been moving away from that, while Google is actively adding new abuses.
Seriously, I think Google is a lot more evil than you think, and that today's Microsoft, at least from the outside, is a lot less evil than you think. They could revert to their old behavior, of course.
> Google has an inherent interest in the > development of a free Internet
No, it really doesn't. It _used_ to a few years ago, but at this point it has more of an interest in the development of an internet it controls and that is dependent on its services. You may be interested in https://news.ycombinator.com/item?id=2989550 starting at the part with "(c").
> I cited specific examples to defend my case
You cited specific instances of past bad behavior by Microsoft. I agree that they happened, absolutely. My point is that Microsoft's current behavior is very different from that. I care more about what these organizations will do tomorrow than what they did yesterday. Insofar as their actions yesterday allow me to predict their actions tomorrow, they're worth considering. But let's not make the mistake of fighting the last war all the time, ok?
> How would anyone be fooled by standardization > efforts just after the Open Document Format saga?
I think you fundamentally misunderstand the way large organizations like Microsoft and Google operate. The people responsible for web browsers and web stuff at Microsoft are a pretty distinct subset from the ones responsible for Word and ODF, with separate goals, etc. Chances are, they're opposed to each other in the internal political infighting.
Again, in the _web_ sphere Microsoft has actually been a pretty model citizen for the last two years or so. Before that they weren't present very much, because they were too busy playing catchup.
> Collectively, people are very well informed
Collectively, people have drunk the Google and Facebook cool-aid. It's starting to wear off for Facebook, with congressional probes and all. I hope it wears off for Google before it's too late for the web as we know it.
> who is still a very real threat to freedoms on the > Internet
Is it a bigger threat than Google? That's not clear to me.
> Again, the ethics of those two companies are very > different.
Agreed, but they're both terrible in slightly different ways.
> Linux (have you forgotten the SCO saga) and > Firefox are their targets.
Are, or were? At least in the web+browser space Microsoft is a very different beastie from 5-10 years ago. They're actually implementing standards, helping draft them, writing testcases, pushing for standardization and interoperability, etc. Much more so than some other browser vendors.
Now I agree that Microsoft has a pretty unpleasant history. But the point is that currently their behavior is improving while Google's is worsening. It's only a matter of time until the badness levels intersect, and not of all that much time.
Now you raise a good point which is that there are many misinformed folks who don't realize the above and think that Google can do no harm while Microsoft is pure evil, and that therefore the political fallout here can be bad. Some education would be in order, perhaps.
> As long as their browser remains standards- > compliant I won't complain.
You're going to need to define "standards-compliant".
Right now the modus operandi in many cases seems to be to add a feature to the browser, write up a draft standard (buggy, in that it doesn't match the implementation and can't actually be implemented by anyone else as-is) describing the feature, throw it over the wall into the W3C and then stop working on the standards end of it. c.f. CSS Transitions, CSS Animations, CSS Transforms for some perfect examples of this.
Yeah, current versions of Adblock Plus shouldn't have known crash issues. I did look at the extensions involved, and none of them look like obvious culprits to me, so I fully expect that you'll still hit the crashes.... but worth double-checking.
> I'm not sure how much money Google is spending > on Chrome
It's obviously not public information, but the following things are known:
1) The number of Google employees working on Chrome full time definitely measures in the hundreds. I'd be very surprised if it's less than 500-600 or so.
2) The amount Google spent on Chrome _marketing_ in the past year is somewhere between $400 million and $2 billion depending on the estimates you look at. Keep in mind that we're talking subway ads all over London, TV ads all over the US, bundling deals that they were paying for, prominent placement of ads on Google properties (forgoing other ad revenue) and so forth.
Google's budget for Chrome is much much bigger than what it paid to Mozilla.
There are certainly north of 500 people doing full-time work on Mozilla.
They're writing browser code, doing audits of CAs, maintaining the various infrastructure involved (build and test farms, addons.mozilla.org, various other Mozilla websites, the bug database, update servers, and so forth), writing documentation (see MDN), helping draft and edit W3C specifications, contributing to the W3C test suites, doing marketing, dealing with payroll and administrative issues, dealing with the legal issues that arise all the time with patents and whatnot, and so forth and so on. The same is true of any other browser vendor.
As one of the other commenters noted, a good estimate for how much an employee costs an employer is twice salary (if you include the employer half of payroll taxes in the US, things like unemployment insurance, noncash benefits provided to employees, 401(k) matches, office space, equipment, travel to things like W3C face to face meetings, etc).
So a spend of $100 million would about correspond to 500 people, yes.
The strategy of _Mozilla_ hinges on developing an alternative to the monopolistic behavior of whoever has such behavior. In 2000 that was Microsoft. If Google or Apple becomes a bigger problem in this sense than Microsoft (arguably already happened with Apple), then Mozilla would presumably want to counter them.
Remember, Mozilla's mission is not "You will be free of Microsoft" but "You will have choice in how you experience and contribute to the web". There's a distinction (and in particular, the former is not sufficient to achieve the latter).
> Chrome isn't bundled with anything other than > Chrome OS
You mean except Skype and the Flash plugin and Avast antivirus and the Adobe Reader and a few other things?
Google's been spending money like water on bundling deals for the last year or two.
None of which has anything to do with the search deal, really. The browsers are absolutely competing, equal footing or not. Whether that affects the search stuff is an interesting question that no one in this discussion is privy to the answer to, I suspect.
One other thing that may be worth trying is disabling the extensions you have installed one at a time and seeing whether any particular one of them is correlated with the crashes...
Sorry for the lag here... Those all look like memory corruption issues of some sort on the JS heap. I'll get the JS folks to take a look. Thank you for the links!
Actually, it does. Firefox market share has been flat or gone down by a percentage point or two (depending on which stats you look at) over the last 2 years. Over the same time, Chrome share is up 10-15%. Guess whose share those 10-15% came out of?
Would you mind looking at about:crashes in your browser and either sending me the links directly (bz at mit dot edu) or putting them in a comment here? Or are you running a version provided by your distribution that doesn't include the crash reporter?
Google and Microsoft started actually throwing money at the problem? That means both developers and marketing; Chrome's yearly marketing expenditure is estimated at between 4 and 20 times Mozilla's _total_ budget.
Oh, and the other thing that happened was that Firefox 4 took a long time to release. So people spent a year or so comparing new Chrome versions with a very old Firefox version (3.6). Some still are, if you read the press.
Add in the fact that some technical debt had to be paid off in Gecko and Spidermonkey (hence the timeline on Firefox 4), and here we are.
Now the good news, for Mozilla, is that the technical debt issue is much less now, and they have a lot more engineers than they did 2 years ago; almost to the point of having enough to cover the whole codebase.;)
Slashdot Mirror
> Why not just run the 64-bit tools on a
> 64-bit platform
There is no 64-bit version of the MSVC linker, and no plans for one according to public statements from Microsoft.
One can run the 32-bit linker on a 64-bit Windows, which gives you 4GB of address space to play with. That's the medium-term solution for Mozilla, but it does require updating the entire Windows build farm to 64-bit Windows and shaking out any compat issues that result. Doable, but will take a few weeks probably.
Usually that line is drawn at whatever the criteria are for getting a restraining order. What those are varies by jurisdiction, I bet.
Firefox offers various security mitigation strategies (in terms of properly dealing with various memory-safety issues, say) that Chrome does not. As far as I can tell, this study just started off with a subset of the list of techniques that Chrome implements and then "studied" which other browsers also implement them, instead of studying what browsers actually do to ensure security and how difficult they are to actually exploit.
Your larger point that modern IE is a fairly secure browser (like any modern browser) is correct, of course.
> I have never seen any difference in google search or
> gmail between Firefox and Chrome
Offline gmail only works in WebKit, by design.
> For example, Mozilla did essentially (3)
Mozilla has consistently tried to convince people to author pages that work in all browsers, in my experience. But I'm not omnicient, of course. ;)
The extensions thing you raise is interesting. I have to admit I'd never thought of it that way before, if only because I'd never used any such extensions. But yes, the number of users of such extensions is minimal in general.
Let's hope that you're right and in a few years my worries here will in fact seem like paranoia...
Actually, I agree with the original point of the thread. What Google is doing with Chrome is:
1) Explicitly authoring its own web properties to specifically work better with Chrome, in hopes that its high market share for things like search and webmail will increase Chrome adoption.
2) Authoring its own web properties to only work with WebKit (e.g. a number of Google's sites sniff UAs and send WebKit-only content to any mobile UA).
3) Urging authors to create Chrome-only content (and more generally, together with Apple, encouraging WebKit-only content).
4) Paying other companies to bundle Chrome with their software so that people end up using it whether they want to or not. This part Microsoft didn't have to do because they just bundled with Windows, of course.
They're not _quite_ as bad as Microsoft was because a lot of this is in fact open source (though they don't exactly take contributions much or plan to share control, so most of what you can do with the source is fork). This does mean that some other project could import some part of Chrome or V8 if desired. But given that none of this code is static (security fixes, spec changes, etc), it's not like a one-time import is useful. You have to keep importing (and hoping that Google doesn't change something you care about) or as I said fork and take over maintenance. In practice, for most situations, both options suck to a good extent. http://3.bp.blogspot.com/-GBSGBbc9UtA/TmexwLaJN9I/AAAAAAAAAKs/06OaexCT5Ms/s1600/Lead%2Bdevice%2Bconcept.png is a good look into Google's general thinking on open in this context, for what it's worth.
One other important note is that Google is a bit better at participating in the standards process than Microsoft was in 2001 (though not than Microsoft in 1998, say). That's a net plus for them.
So no, comparing what Google is doing with what Microsoft was doing with IE6 is not at all ludicrous. There are some important differences, and Google doesn't have a browser monopoly, so they have to play a bit nicer so far, but the overall strategy looks similar to me. The key to frustrating it, of course, is for them to not end up with said browser monopoly.
> It seems more likely that other engines are just
> better than V8 for some types of code
Well, yes. But some optimizations are just mutually exclusive because they require different design tradeoffs.
As for the rest... Google is definitely being evil in various ways, and is in fact much like Microsoft in the late 90s, in my opinion, but what they're doing with V8 is just self-interested, not so much evil.
> what's preventing anyone from implementing these
> optimizations in other javascript engines
They make other (non-Google) sites run slower?
> Is it reasonable to prevent Google from optimizing
> anything just because other browsers and libraries
> would then be slower?
I didn't say anything about preventing Google from doing whatever they want. I'm not sure where you got that.
I'm also not sure where you got your "evil" bit from. I made a factual statement, with no value judgments attached to it.
> V8 isn't a javascript library,
You misunderstood.
Google's JavaScript libraries are purposefully written to run faster on V8 specifically, often at the expense of performance in other browsers. And at the same time, V8 is written to run the code patterns those libraries use faster, often at the expense of other code patterns used elsewhere.
The net result is that Google properties work better in Chrome and at the same time Chrome works better on Google properties.
OK, so it's not the extensions. Thanks!
> Microsoft has a history of abuse, while you are only
> weary of the domination of Google.
More precisely, Microsoft has a history of abuse and has been moving away from that, while Google is actively adding new abuses.
Seriously, I think Google is a lot more evil than you think, and that today's Microsoft, at least from the outside, is a lot less evil than you think. They could revert to their old behavior, of course.
> Google has an inherent interest in the
> development of a free Internet
No, it really doesn't. It _used_ to a few years ago, but at this point it has more of an interest in the development of an internet it controls and that is dependent on its services. You may be interested in https://news.ycombinator.com/item?id=2989550 starting at the part with "(c").
> I cited specific examples to defend my case
You cited specific instances of past bad behavior by Microsoft. I agree that they happened, absolutely. My point is that Microsoft's current behavior is very different from that. I care more about what these organizations will do tomorrow than what they did yesterday. Insofar as their actions yesterday allow me to predict their actions tomorrow, they're worth considering. But let's not make the mistake of fighting the last war all the time, ok?
> How would anyone be fooled by standardization
> efforts just after the Open Document Format saga?
I think you fundamentally misunderstand the way large organizations like Microsoft and Google operate. The people responsible for web browsers and web stuff at Microsoft are a pretty distinct subset from the ones responsible for Word and ODF, with separate goals, etc. Chances are, they're opposed to each other in the internal political infighting.
Again, in the _web_ sphere Microsoft has actually been a pretty model citizen for the last two years or so. Before that they weren't present very much, because they were too busy playing catchup.
> Collectively, people are very well informed
Collectively, people have drunk the Google and Facebook cool-aid. It's starting to wear off for Facebook, with congressional probes and all. I hope it wears off for Google before it's too late for the web as we know it.
> who is still a very real threat to freedoms on the
> Internet
Is it a bigger threat than Google? That's not clear to me.
> Again, the ethics of those two companies are very
> different.
Agreed, but they're both terrible in slightly different ways.
> Linux (have you forgotten the SCO saga) and
> Firefox are their targets.
Are, or were? At least in the web+browser space Microsoft is a very different beastie from 5-10 years ago. They're actually implementing standards, helping draft them, writing testcases, pushing for standardization and interoperability, etc. Much more so than some other browser vendors.
Now I agree that Microsoft has a pretty unpleasant history. But the point is that currently their behavior is improving while Google's is worsening. It's only a matter of time until the badness levels intersect, and not of all that much time.
Now you raise a good point which is that there are many misinformed folks who don't realize the above and think that Google can do no harm while Microsoft is pure evil, and that therefore the political fallout here can be bad. Some education would be in order, perhaps.
> As long as their browser remains standards-
> compliant I won't complain.
You're going to need to define "standards-compliant".
Right now the modus operandi in many cases seems to be to add a feature to the browser, write up a draft standard (buggy, in that it doesn't match the implementation and can't actually be implemented by anyone else as-is) describing the feature, throw it over the wall into the W3C and then stop working on the standards end of it. c.f. CSS Transitions, CSS Animations, CSS Transforms for some perfect examples of this.
Yeah, current versions of Adblock Plus shouldn't have known crash issues. I did look at the extensions involved, and none of them look like obvious culprits to me, so I fully expect that you'll still hit the crashes.... but worth double-checking.
> I'm not sure how much money Google is spending
> on Chrome
It's obviously not public information, but the following things are known:
1) The number of Google employees working on Chrome full time definitely measures in the hundreds. I'd be very surprised if it's less than 500-600 or so.
2) The amount Google spent on Chrome _marketing_ in the past year is somewhere between $400 million and $2 billion depending on the estimates you look at. Keep in mind that we're talking subway ads all over London, TV ads all over the US, bundling deals that they were paying for, prominent placement of ads on Google properties (forgoing other ad revenue) and so forth.
Google's budget for Chrome is much much bigger than what it paid to Mozilla.
> why didn't the Mozilla Foundation start squirreling a
> lot of that money
They did. As you'd know if you read their publicly available financial statements.
There are certainly north of 500 people doing full-time work on Mozilla.
They're writing browser code, doing audits of CAs, maintaining the various infrastructure involved (build and test farms, addons.mozilla.org, various other Mozilla websites, the bug database, update servers, and so forth), writing documentation (see MDN), helping draft and edit W3C specifications, contributing to the W3C test suites, doing marketing, dealing with payroll and administrative issues, dealing with the legal issues that arise all the time with patents and whatnot, and so forth and so on. The same is true of any other browser vendor.
As one of the other commenters noted, a good estimate for how much an employee costs an employer is twice salary (if you include the employer half of payroll taxes in the US, things like unemployment insurance, noncash benefits provided to employees, 401(k) matches, office space, equipment, travel to things like W3C face to face meetings, etc).
So a spend of $100 million would about correspond to 500 people, yes.
The strategy of _Mozilla_ hinges on developing an alternative to the monopolistic behavior of whoever has such behavior. In 2000 that was Microsoft. If Google or Apple becomes a bigger problem in this sense than Microsoft (arguably already happened with Apple), then Mozilla would presumably want to counter them.
Remember, Mozilla's mission is not "You will be free of Microsoft" but "You will have choice in how you experience and contribute to the web". There's a distinction (and in particular, the former is not sufficient to achieve the latter).
> Chrome isn't bundled with anything other than
> Chrome OS
You mean except Skype and the Flash plugin and Avast antivirus and the Adobe Reader and a few other things?
Google's been spending money like water on bundling deals for the last year or two.
None of which has anything to do with the search deal, really. The browsers are absolutely competing, equal footing or not. Whether that affects the search stuff is an interesting question that no one in this discussion is privy to the answer to, I suspect.
Google is routinely authoring sites that only work in WebKit-based browsers.
Apple (but not Google, to their credit) routinely encourages web developers to create WebKit-specific website via their developer documentation.
So the exclusivity arrangements part is all effectively happening, just like in 1999 or so.
One other thing that may be worth trying is disabling the extensions you have installed one at a time and seeing whether any particular one of them is correlated with the crashes...
Sorry for the lag here... Those all look like memory corruption issues of some sort on the JS heap. I'll get the JS folks to take a look. Thank you for the links!
> Where are all those Chrome users coming from?
Try checking stats for South America....
Or if you want to be somewhat depressed about the whole thing, try China or South Korea. ;)
Actually, it does. Firefox market share has been flat or gone down by a percentage point or two (depending on which stats you look at) over the last 2 years. Over the same time, Chrome share is up 10-15%. Guess whose share those 10-15% came out of?
Would you mind looking at about:crashes in your browser and either sending me the links directly (bz at mit dot edu) or putting them in a comment here? Or are you running a version provided by your distribution that doesn't include the crash reporter?
> What happened?
Google and Microsoft started actually throwing money at the problem? That means both developers and marketing; Chrome's yearly marketing expenditure is estimated at between 4 and 20 times Mozilla's _total_ budget.
Oh, and the other thing that happened was that Firefox 4 took a long time to release. So people spent a year or so comparing new Chrome versions with a very old Firefox version (3.6). Some still are, if you read the press.
Add in the fact that some technical debt had to be paid off in Gecko and Spidermonkey (hence the timeline on Firefox 4), and here we are.
Now the good news, for Mozilla, is that the technical debt issue is much less now, and they have a lot more engineers than they did 2 years ago; almost to the point of having enough to cover the whole codebase. ;)