If you read the various NaCl marketing Google has been doing, that sends the same message: NaCl is for "cool new stuff".
Also, there is no indication that they plan to have NaCl in Chrome require any user opt-in. It needs opt-in for now while it's experimental, of course.
I'd really like to see a reference for your "IE6 compat" claim, if you have one, since it contradicts everything I've read coming out of Google about NaCl.
One other note. China's internet penetration is about 30%. Russia's is 43%. Neither one is really "wired out the wazoo" (for comparison, Iceland is at 97.6%, the US is at 77%, South Korea is at 81%; again if you believe the numbers).
So between the three of them, that's 26.4% of global users. What fraction of global searchers that is will depend on how search frequency varies. That's assuming that we trust the numbers above and "search engine market share" is fraction of number of searches done...
> The Chinese know full well the prize committee is > not an agent of the government
Uh... The peace prize committee is appointed by the Norwegian Parliament. There is now a restriction that sitting members of the Parliament can't be on the committee, but it's just made up of past members of the Parliament.
1) Chair was a member of the Parliament 1993-2009, president of the Parliament 2005-2009, Prime minister 1996-1997. On the committee since 2009.
2) Deputy chair was a member of the Parliament 1981-1997, on the committee since 2003.
3) Third member was a member of the Parliament 1977-1993, member of the committee since 1994.
4) Fourth member was a member of the Parliament 1989-1993, on the comittee since 2000.
5) Fifth member was a member of the Parliament 1997-2009, a member of the committee since 2009.
So as of 2010, two of the 5 committee members had been in the Parliament just the year before, and one more had been on the committee ever since she stopped being a member of the Parliament.
Given this setup, and the fact that the committee members have limited terms have to be reappointed by the Parliament, it'd have a pretty hard time not being an agent of the government. Now obviously the Parliament has no _official_ way to influence the decision... but all these people's friends very much "the government" and their standing in the social circles they frequent will depend on their committee's decision.
If they wanted a _really_ independent committee, the selection process for it and the composition of the committee would be radically different. But they want plausible deniability, not real independence.
Now what you say about China certainly seems true to me, and would be true even if the committee _were_ an official agent of the Norwegian government, instead of the unofficial one it actually is.
> So I should update from v2.0.0.20 sometime in the > next few years?
Unless you _like_ random websites being able to run arbitrary code on your computer, yes. You should.
I mostly agree with you about random UI changes for the sake of UI changes. But browsing with an out-of-date rendering engine, no matter what the UI, is just daft.
> If it was that good eight months ago, then why didn't they release it?
Because one of the goals of having a new release was to have the new JS engine, and that part was NOT ready eight months ago.
Hence the change in goal-setting. No more "must have for this release" features. Just releases, and features that are ready get released, while the rest keep getting worked on.
You're right that more frequent updates only work if they're less obtrusive. I think everyone involved realizes that.
> If it means the resulting product is bug-free > (read: well tested) and of higher quality--- so be it
Most people aren't like you, honestly. But more importantly, faster release schedule need not mean less testing. There are plenty of things that got fixed in Gecko in October 2009 and had lots of testing by May 2010 that are only going to ship now, in March 2011.
Now faster releases _could_ be done in a crappy way. We'll see whether Firefox manages to avoid that, but I think they will.
> I DON'T CARE if FF beats IE[0-9] or Chrome > by 3.2ms
Neither does anyone else. People do care about situations where one browser is 2x faster (or even worse, 200x faster) than another on some pages.
They also care if their favorite sites can't be accessed in their browser, or are just much slower in it due to it missing some feature that the sites work around in a crappy way.
The problem is, at this point, the experience of a 1-year-old browser is not good and snappy in many cases and the performance difference often _is_ too lopsided. I'm talking performance on things that matter to users, not crap like Sunspider or the V8 benchmark.
I agree that your concerns are legitimate ones; I don't think the Gecko and Firefox developers plan to compromise on leaks, privacy, or other important things to get the faster schedule. They _do_ plan to change how new features are added, though: features will no longer be able to land on the main development tree in a half-baked state with the assumption that the remaining problems can be sorted out before the scheduled beta or whatnot. If they land and there are major problems, they will get turned off or backed out until they're up to snuff. I suspect this will actually make things more like you want them, rather than less.
Nah, there aren't enough of them for that. Firefox 3.5, even while being supported, is under 3% overall usage (for comparison the unsupported Firefox 3.0 that no one cares about anymore is about 1.5%; all figures according to statcounter; I bet other sources will have different numbers).
Then again, IE6 is 5% according to the same source.... but was 13% just a year ago (when Firefox 3.5 was 27%, by the way; this was before 3.6 was released).
It makes the lag to shipping new web-facing features and performance improvements too long. As a result you end up with situations like the current one, where Firefox 3.6 is significantly worse than the already-shipping competition (except IE8) in various performance and standards-compliance metrics... while the builds as of June of 2010, say, were much better than 3.6.
This isn't about version numbers; it's about getting new features into the hands of users faster and not gating feature A, which is completely done, on feature B, which might get done sometime.
Out of curiousity, what do you plan to do once 3.5 stops getting security updates?
(This is a serious question; I'm trying to understand how users respond to that situation so we can take it into account when we decide how long to keep up security updates.)
Agreed re: maketing being just half of it; for most non-IE browsers getting users to try them has been the hard part, though.
> Lots of people try Opera
Lots of tech geeks try opera. I have yet to meet an "average" user who has. Not that I think they'd necessarily like it if they _did_ try it, but nevertheless.
b12 was released when all the hard blockers that needed beta exposure (as in, could result in web compat issues and such) were fixed.
The remaining hard blockers are ones where the fix is expected to be very safe and extremely unlikely to cause compat or user-facing problems. So it's OK to go directly to RC after fixing them, instead of having yet more beta testing.
> IE is not longer "seemingly invulnerable", but it's still > the dominant browser.
The point is at this point no one sane creating a new site creates it to work only in IE.
It doesn't matter what most people use, as long as all people have a choice of what browser to use and new browsers can be created by those who wish to do so. We're not there yet on this last (due to sites UA-sniffing, etc), but we're a lot closer to it than we were in 2002.
> Publishers WANT you to be able to view their content
More precisely, they want you to view their content in a particular way. Want to save that movie so you can watch it later on the plane? No, sorry, can't do that.
> much of the problems you mentioned can be > solved via the provided sdk.
How exactly do you solve "jump to arbitrary memory" with an SDK?
> all computer programming problems can be fixed > with adding another layer of indirection
At the cost of performance, as usual for layers of indirection.
> Memory leaks can be controlled through malloc > by just discarding the block after the user leaves > the page.
There are plenty of pages that users never leave nowadays (gmail, facebook, twitter, etc).
> JS is now both interpreted and compiled.
That depends on the implementation.
> It is first run in interpreted mode,
This is not true in V8, say. It has no interpreter (which is why it's not portable to architectures for which it has no jit backend, by the way). I can't speak to what Carakan and IE9's JS implementation do here.
> Thus compilation runs on a backend thread
Spidermonkey does not compile on a background thread. I can't speak to the others with certainty, but I'm pretty sure V8 doesn't either, nor JavaScriptCore. The closed-source implementation I have less infomation on.
Trying to do this is a worthy goal, but it's not happening yet, and there are some difficulties with doing it.
> and time is somewhat irrelevant
Not really, no. See above.
> why compile from source each time
There is talk of caching compiled forms, yes.
> A better solution is to include an intermediary > bytecode directly into the page.
The choice of the particular bytecode presupposes particular VM implementation strategies, typically. It also makes some optimizations easier while others are harder or impossible.
Witness the fact that the actual "bytecode" used by current browsers that use it (e.g. V8 does not) is different between different implementations.
I think you're assuming a uniformity of implementations strategies for JS that's just not there, not least because as you pointed out this is an area of active research.
Yes, but the original claim was that you have to do HTML5 games in canvas because you can't flip images otherwise. IE8 doesn't support canvas either, so if you're writing an HTML5 game with canvas you aren't targeting IE8 to start with.
The problem is that Core Animation is too high-level. It wants to handle the entire animation itself, which makes it suitable for implementing CSS Transitions and CSS Animations (heck, the initial specs for those were basically "do what Core Animation does, because that's how we implemented it"), but not great for handling painting of web pages where you don't know what the web page will do next
To put it in web terms, Core Animation is closer to being like SVG while Direct2D is closer to being like canvas. They both have their use cases, but do quite different things.
Slashdot Mirror
I'm not forgetting that at all. That would square quite nicely with China at about 25% of total internet users for the moment, in fact.
> NaCl is designed to let corporations stuck with
> old IE6 based apps
From http://code.google.com/p/nativeclient/ :
"We believe that Native Client technology will help web developers to create richer and more dynamic browser-based applications."
http://code.google.com/games/technology-nacl.html explicitly talks about using it for new games, not for IE6 compat.
If you read the various NaCl marketing Google has been doing, that sends the same message: NaCl is for "cool new stuff".
Also, there is no indication that they plan to have NaCl in Chrome require any user opt-in. It needs opt-in for now while it's experimental, of course.
I'd really like to see a reference for your "IE6 compat" claim, if you have one, since it contradicts everything I've read coming out of Google about NaCl.
The browser doesn't have to run on the original iPad.
If you're treating your iPad as mostly a netbook, then the performance improvement is quite real.
One other note. China's internet penetration is about 30%. Russia's is 43%. Neither one is really "wired out the wazoo" (for comparison, Iceland is at 97.6%, the US is at 77%, South Korea is at 81%; again if you believe the numbers).
If you can believe http://www.internetworldstats.com/stats.htm and http://www.internetworldstats.com/stats3.htm#asia and http://www.internetworldstats.com/stats4.htm#europe then there are 1.967 billion total internet users in the world, of whom .0597 billion are in Russia, 0.42 billion are in China, and 0.0394 billion in South Korea.
So between the three of them, that's 26.4% of global users. What fraction of global searchers that is will depend on how search frequency varies. That's assuming that we trust the numbers above and "search engine market share" is fraction of number of searches done...
> Because the interface to Google monoculture is just
> a standard compliant browser.
For now. Until the fail to push through standardization of NaCl, say, but decide to use it anyway.
> The Chinese know full well the prize committee is
> not an agent of the government
Uh... The peace prize committee is appointed by the Norwegian Parliament. There is now a restriction that sitting members of the Parliament can't be on the committee, but it's just made up of past members of the Parliament.
Looking at the makeup of the committee as of today ( http://nobelpeaceprize.org/en_GB/nomination_committee/members/ ) I see:
1) Chair was a member of the Parliament 1993-2009, president of the Parliament 2005-2009, Prime minister 1996-1997. On the committee since 2009.
2) Deputy chair was a member of the Parliament 1981-1997, on the committee since 2003.
3) Third member was a member of the Parliament 1977-1993, member of the committee since 1994.
4) Fourth member was a member of the Parliament 1989-1993, on the comittee since 2000.
5) Fifth member was a member of the Parliament 1997-2009, a member of the committee since 2009.
So as of 2010, two of the 5 committee members had been in the Parliament just the year before, and one more had been on the committee ever since she stopped being a member of the Parliament.
Given this setup, and the fact that the committee members have limited terms have to be reappointed by the Parliament, it'd have a pretty hard time not being an agent of the government. Now obviously the Parliament has no _official_ way to influence the decision... but all these people's friends very much "the government" and their standing in the social circles they frequent will depend on their committee's decision.
If they wanted a _really_ independent committee, the selection process for it and the composition of the committee would be radically different. But they want plausible deniability, not real independence.
Now what you say about China certainly seems true to me, and would be true even if the committee _were_ an official agent of the Norwegian government, instead of the unofficial one it actually is.
> So I should update from v2.0.0.20 sometime in the
> next few years?
Unless you _like_ random websites being able to run arbitrary code on your computer, yes. You should.
I mostly agree with you about random UI changes for the sake of UI changes. But browsing with an out-of-date rendering engine, no matter what the UI, is just daft.
> If it was that good eight months ago, then why didn't they release it?
Because one of the goals of having a new release was to have the new JS engine, and that part was NOT ready eight months ago.
Hence the change in goal-setting. No more "must have for this release" features. Just releases, and features that are ready get released, while the rest keep getting worked on.
You're right that more frequent updates only work if they're less obtrusive. I think everyone involved realizes that.
> If it means the resulting product is bug-free
> (read: well tested) and of higher quality--- so be it
Most people aren't like you, honestly. But more importantly, faster release schedule need not mean less testing. There are plenty of things that got fixed in Gecko in October 2009 and had lots of testing by May 2010 that are only going to ship now, in March 2011.
Now faster releases _could_ be done in a crappy way. We'll see whether Firefox manages to avoid that, but I think they will.
> I DON'T CARE if FF beats IE[0-9] or Chrome
> by 3.2ms
Neither does anyone else. People do care about situations where one browser is 2x faster (or even worse, 200x faster) than another on some pages.
They also care if their favorite sites can't be accessed in their browser, or are just much slower in it due to it missing some feature that the sites work around in a crappy way.
The problem is, at this point, the experience of a 1-year-old browser is not good and snappy in many cases and the performance difference often _is_ too lopsided. I'm talking performance on things that matter to users, not crap like Sunspider or the V8 benchmark.
I agree that your concerns are legitimate ones; I don't think the Gecko and Firefox developers plan to compromise on leaks, privacy, or other important things to get the faster schedule. They _do_ plan to change how new features are added, though: features will no longer be able to land on the main development tree in a half-baked state with the assumption that the remaining problems can be sorted out before the scheduled beta or whatnot. If they land and there are major problems, they will get turned off or backed out until they're up to snuff. I suspect this will actually make things more like you want them, rather than less.
Nah, there aren't enough of them for that. Firefox 3.5, even while being supported, is under 3% overall usage (for comparison the unsupported Firefox 3.0 that no one cares about anymore is about 1.5%; all figures according to statcounter; I bet other sources will have different numbers).
Then again, IE6 is 5% according to the same source.... but was 13% just a year ago (when Firefox 3.5 was 27%, by the way; this was before 3.6 was released).
> What's wrong with that?
It makes the lag to shipping new web-facing features and performance improvements too long. As a result you end up with situations like the current one, where Firefox 3.6 is significantly worse than the already-shipping competition (except IE8) in various performance and standards-compliance metrics... while the builds as of June of 2010, say, were much better than 3.6.
This isn't about version numbers; it's about getting new features into the hands of users faster and not gating feature A, which is completely done, on feature B, which might get done sometime.
Out of curiousity, what do you plan to do once 3.5 stops getting security updates?
(This is a serious question; I'm trying to understand how users respond to that situation so we can take it into account when we decide how long to keep up security updates.)
Agreed re: maketing being just half of it; for most non-IE browsers getting users to try them has been the hard part, though.
> Lots of people try Opera
Lots of tech geeks try opera. I have yet to meet an "average" user who has. Not that I think they'd necessarily like it if they _did_ try it, but nevertheless.
Chrome has a _much_ bigger marketing budget than Opera does.
So perhaps more precisely Chrome was marketed, period.
You're confusing "finding" and "fixing"...
b12 was released when all the hard blockers that needed beta exposure (as in, could result in web compat issues and such) were fixed.
The remaining hard blockers are ones where the fix is expected to be very safe and extremely unlikely to cause compat or user-facing problems. So it's OK to go directly to RC after fixing them, instead of having yet more beta testing.
> IE is not longer "seemingly invulnerable", but it's still
> the dominant browser.
The point is at this point no one sane creating a new site creates it to work only in IE.
It doesn't matter what most people use, as long as all people have a choice of what browser to use and new browsers can be created by those who wish to do so. We're not there yet on this last (due to sites UA-sniffing, etc), but we're a lot closer to it than we were in 2002.
> Publishers WANT you to be able to view their content
More precisely, they want you to view their content in a particular way. Want to save that movie so you can watch it later on the plane? No, sorry, can't do that.
There are H.264 patents in France. They're just method patents, not software patents.
See http://weblogs.mozillazine.org/bz/archives/020400.html for a direct link to the patent list for H.264 and a summary of some of the countries those patents were granted in.
> much of the problems you mentioned can be
> solved via the provided sdk.
How exactly do you solve "jump to arbitrary memory" with an SDK?
> all computer programming problems can be fixed
> with adding another layer of indirection
At the cost of performance, as usual for layers of indirection.
> Memory leaks can be controlled through malloc
> by just discarding the block after the user leaves
> the page.
There are plenty of pages that users never leave nowadays (gmail, facebook, twitter, etc).
> JS is now both interpreted and compiled.
That depends on the implementation.
> It is first run in interpreted mode,
This is not true in V8, say. It has no interpreter (which is why it's not portable to architectures for which it has no jit backend, by the way). I can't speak to what Carakan and IE9's JS implementation do here.
> Thus compilation runs on a backend thread
Spidermonkey does not compile on a background thread. I can't speak to the others with certainty, but I'm pretty sure V8 doesn't either, nor JavaScriptCore. The closed-source implementation I have less infomation on.
Trying to do this is a worthy goal, but it's not happening yet, and there are some difficulties with doing it.
> and time is somewhat irrelevant
Not really, no. See above.
> why compile from source each time
There is talk of caching compiled forms, yes.
> A better solution is to include an intermediary
> bytecode directly into the page.
The choice of the particular bytecode presupposes particular VM implementation strategies, typically. It also makes some optimizations easier while others are harder or impossible.
Witness the fact that the actual "bytecode" used by current browsers that use it (e.g. V8 does not) is different between different implementations.
I think you're assuming a uniformity of implementations strategies for JS that's just not there, not least because as you pointed out this is an area of active research.
Yes, but the original claim was that you have to do HTML5 games in canvas because you can't flip images otherwise. IE8 doesn't support canvas either, so if you're writing an HTML5 game with canvas you aren't targeting IE8 to start with.
> Which browser can horizontally flip an image in a div?
Any browser implementing CSS 2D Transforms (Firefox 3.5 or newer, Safari 3.2 or newer, Chrome at 7 and maybe even older, Opera 10.5 or newer, IE 9).
So pretty much anything on the market that supports canvas supports transforms too.
The problem is that Core Animation is too high-level. It wants to handle the entire animation itself, which makes it suitable for implementing CSS Transitions and CSS Animations (heck, the initial specs for those were basically "do what Core Animation does, because that's how we implemented it"), but not great for handling painting of web pages where you don't know what the web page will do next
To put it in web terms, Core Animation is closer to being like SVG while Direct2D is closer to being like canvas. They both have their use cases, but do quite different things.
> JS is a much higher language than C
Higher _level_ language. I have got to stop commenting on /. at 2am....