Sure. Getting all your history and bookmarks wiped out is worse than the browser crashing. Getting your hard drive wiped is worse than the browser crashing.
The former is not all that far-fetched with nightly builds. The latter is not guaranteed to not happen.
Sure thing. See the original paper at http://base.google.com/base_media?q=hand1017890191470242229&size=8 (sadly it's a zipped-up PDF instead of just having the compression in the PDF itself). The author does some pretty explicit comparisons against existing Java JITs and discusses the tradeoffs involved in tracing instead of doing whole-method JIT.
The problem with SVG is often not the CPU but the graphics card... Most of what's needed to do SVG fast can be offloaded to the GPU, if the GPU is up to it. Of course with netbooks you lose on that front too.
> Remember when browsers were considered I/O-bound apps, anyway?
That was before people were trying to do fancy layout and even more importantly fancy graphics in browsers, right?
Try out http://www.mozbox.org/jdll/video.xhtml in Firefox 3.1 beta 1, for example. Realtime edge-detection filtering on a running video isn't so much I/O bound...
> Unlike the US, they have a massive manufacturing base
Actually, the two have approximately equally sized manufacturing bases in terms of production. The US has much higher productivity, so many fewer manufacturing _workers_, but the total production is pretty similar. Further, China's total exports are about the same as those of the US. It's just that the US imports so much more than China does that causes our current trade imbalance.
There's also the fact that China's manufacturing base has been growing recently while ours has been shrinking, so if current trends continue then eventually what you say will be true.
> They have the rest of the world (Europe, Asia, Africa, Russia, the list goes on).
The real problem is that they employ people by producing all sorts of stuff that their own people don't (can't, largely) buy. So they HAVE to export to keep the economy going. They've been managing it so far by keeping a currency peg against the dollar so that their production is cheap in the US. This works because the US doesn't impose tariffs much on manufactured goods, even in the face of blatant currency manipulation.
The situation with Europe, Asia, Africa, Russia is quite different. No qualms about tariffs there, especially if it will protect domestic industries. So attempts by China to shift their exports elsewhere might be met with strong protectionist measures, making the US rather hard to replace.
Of course all this is speculation. And really, China should be working on creating domestic demand for its products. The problem is that doing too much of that threatens the political stability of the current setup, so it's been a pretty slow process.
> Epic fail. How do you measure IQ? Unless they have some sort of new magic way of doing it, > IQ is measured by knowledge and skill, ergo, IQ IS knowledge and skill.... > Even if IQ is constant, there's no way to scientifically prove it
I can think of one off the top of my head, even for your "knowledge and skill" definition.
Say we have a knowledge and skill test administered to adults every 10 years (20, 30, 40, 50, say).
Say we have another such test administered to 2-year-olds (clearly quite different; administering the same test to both makes no sense).
Now the big question is whether the five points we get for every person are correlated, and if so how much. If they're not correlated at all, then there's not much to talk about with these two tests. But if they're highly correlated, then what we have is a way to predict, at age 2, how well a person will score on knowledge and skill tests through the rest of their life. Which would imply that there might be, in fact, some underlying quality affecting performance on the tests involved.
Environmental effects could be controlled for in the standard ways with adoption studies and so forth, or simply by controlling for the environment, though this does reduce the number of people who can be studied.
I'm pretty sure the above is pretty much what's been done with IQ tests (or at least I would hope so!).
Now as far as calling this underlying factor "intelligence", that's a different kettle of fish. As you point out, intelligence has a wide range of denotations, not to mention connotations, and it's not clear that all of these would be covered by whatever factor it is our testing battery isolates.
A further interesting twist to the experiment would be attempting to establish whether there is a correlation between our test results and the ability to acquire new knowledge and skills. I would strongly suspect that _if_ all the test results correlate, then there will be some areas where skill-acquisition would also strongly correlate with the test results.
A good question is whether a test can be designed that would strongly correlate with skill-acquisition in a wide range of areas. I honestly don't know the answer to that, but I have a hard time believing it hasn't been studied.
The artist has the right to pursue any model he sees fit, but that does not imply that some models should therefore receive government protection.
In other words, if the model works, great. But you are in fact advocating that the government support and encourage what would otherwise be an unsuccessful business model (mostly due to the fact that it's a poor business model nowadays).
> It seems like Mozilla is now intent on adding features rather than having a great stable > product.
That's a false dichotomy, really.
But there's one more issue here. All the browsers are in a position where they'll become totally irrelevant in a few years if they don't have feature parity with Flash and Silverlight... so they're working on said feature parity (except maybe IE, which doesn't mind being irrelevant if Silverlight wins).
> Would you be happy with a plugin that stored, and filled in your bank account login > details when your bank's site asked for them?
If it had the option to never fill them in, and otherwise asked me every time? I might not mind, though I'd prefer those not be stored on my hard drive because my computer might be stolen. That part of the analogy doesn't apply here, since no additional data is being stored on your computer, though.
> It wouldn't be more than a few days until there was an exploit that could access that > information without asking you.
If your opinion of browser security is that low, you should have much bigger worries than this API. You should be arguing for the removal of file inputs, for example. And of XMLHttpRequest. If those are subverted you lose a lot more than just your location.
> I even run Firefox as a separate user on my home system
OK, that does mitigate potential file input issues a good bit....;)
That's a really good question. For now, the W3C is publishing a spec that will require it, but there've been rumblings to the effect that not all W3C specs are worth implementing...
Realistically, this feature is no more dangerous than many other things (like file inputs, etc) if properly secured. You're right that it increases attack area, but so does SVG. See, for example, http://lists.w3.org/Archives/Public/www-svg/2008Sep/0112.html
So the only real question is whether the benefits outweigh the risks. On the desktop, I'm not sure. On mobile devices, almost certainly (especially since those often have an existing way to pinpoint their location anyway).
You know what's funny? I can't figure out whether the "political enemies" you're talking about are Republicans or Democrats. They're both doing every single thing on your list (except possibly for the "give peace a chance" shirts; you'd have to have a different shirt to get hassled by Democrats).
Point is, both of the major parties are trying to expand the power of all levels of government (and hence their own) as fast as they can, at the expense of, first of all, citizens of the US, and secondly the lower levels of government (the ones more likely to actually respond to their constituents) as more and more power is concentrated upward.
That second link gives some examples of the use cases geolocation can address that are not addressed by existing technology.
I agree that on the desktop there's not as much use for this, but on mobile devices (think smartphones), it's a heck of a lot more useful.
If you feel strongly about how this should work and the current spec draft at http://dev.w3.org/geo/api/spec-source.html doesn't match that, you may want to send e-mail to public-geolocation@w3.org with your comments.
So what, exactly, is a web browser? Heck, what's "web"? The problem is that the definition keeps changing. Used to be, no image support was fine in a web browser. And no script support. And no CSS support. And no DOM support.
Note that the W3C is standardizing a geolocation API; this is just an implementation of that W3C spec.
I should also point out that Firefox 3 is faster and more memory-efficient than Firefox 2 or Firefox 1.5 or Firefox 1.0 (though when comparing to 1.0 you probably need to make sure to be comparing on a computer with a reasonably modern graphics card). It's certainly more secure than the older versions: a number of bugs have been fixed since the Firefox 1.0 days.
I'm curious: what is the exact problem you have with the geolocation stuff? No data is given to the server unless you explicitly allow it for that particular site. If you _do_ allow it, it enables a class of very interesting web applications, especially on mobile devices.
Could you explain what your objections are, past an instinctive reaction to the word "track"?
Lawsuits? The loss of goodwill would be a lot more important. Even more important would be the fact that a default of anything other than "nothing at all" is fundamentally the wrong thing to do.
I don't think you need to worry much on this score. The people working on this are neither stupid nor malicious.
Slashdot Mirror
Sure. Getting all your history and bookmarks wiped out is worse than the browser crashing. Getting your hard drive wiped is worse than the browser crashing.
The former is not all that far-fetched with nightly builds. The latter is not guaranteed to not happen.
This would make a kinda useful bug report, you know. ;)
I can't answer the "what gives" question yet, but I filed a bug at https://bugzilla.mozilla.org/show_bug.cgi?id=460964 if you want to keep track of the progress on figuring out what is giving.
Sure thing. See the original paper at http://base.google.com/base_media?q=hand1017890191470242229&size=8 (sadly it's a zipped-up PDF instead of just having the compression in the PDF itself). The author does some pretty explicit comparisons against existing Java JITs and discusses the tradeoffs involved in tracing instead of doing whole-method JIT.
The problem with SVG is often not the CPU but the graphics card... Most of what's needed to do SVG fast can be offloaded to the GPU, if the GPU is up to it. Of course with netbooks you lose on that front too.
So to be precise, a lot of stuff browsers are asked to do nowadays is more and more graphics-subsystem bound.
> Remember when browsers were considered I/O-bound apps, anyway?
That was before people were trying to do fancy layout and even more importantly fancy graphics in browsers, right?
Try out http://www.mozbox.org/jdll/video.xhtml in Firefox 3.1 beta 1, for example. Realtime edge-detection filtering on a running video isn't so much I/O bound...
> Firefox 3 even runs javascript JIT, just like java.
Except it's completely different in the way it operates from all the Java jits....
> but the President submits the first cut
Only because Congress let him. This didn't use to be the case, and need not be in the future. I'd argue even that it shouldn't be, in the future.
> By 2000, Clinton (and his VP Gore) had cut the deficit all the way back to a "surplus"
You do realize that budgets (and hence deficits) are controlled by Congress (a Republican congress, in this case), not the president?
The president can make suggestions, but Congress ultimately decides on the whole thing (modulo veto, of course).
> Disclaimer: I don't know if the OP is a professor of economics
Doesn't matter, since the quote he pasted in _was_ in fact said by a professor of economics.
> Unlike the US, they have a massive manufacturing base
Actually, the two have approximately equally sized manufacturing bases in terms of production. The US has much higher productivity, so many fewer manufacturing _workers_, but the total production is pretty similar. Further, China's total exports are about the same as those of the US. It's just that the US imports so much more than China does that causes our current trade imbalance.
There's also the fact that China's manufacturing base has been growing recently while ours has been shrinking, so if current trends continue then eventually what you say will be true.
> They have the rest of the world (Europe, Asia, Africa, Russia, the list goes on).
The real problem is that they employ people by producing all sorts of stuff that their own people don't (can't, largely) buy. So they HAVE to export to keep the economy going. They've been managing it so far by keeping a currency peg against the dollar so that their production is cheap in the US. This works because the US doesn't impose tariffs much on manufactured goods, even in the face of blatant currency manipulation.
The situation with Europe, Asia, Africa, Russia is quite different. No qualms about tariffs there, especially if it will protect domestic industries. So attempts by China to shift their exports elsewhere might be met with strong protectionist measures, making the US rather hard to replace.
Of course all this is speculation. And really, China should be working on creating domestic demand for its products. The problem is that doing too much of that threatens the political stability of the current setup, so it's been a pretty slow process.
And just to clarify, that's just the argument that copyright is not a "natural right" in the way that freedoms of speech, association, etc. are.
It might still be a useful tool, of course. My personal view is that it is, in a more limited form than what we have now.
> Epic fail. How do you measure IQ? Unless they have some sort of new magic way of doing it, > IQ is measured by knowledge and skill, ergo, IQ IS knowledge and skill. ...
> Even if IQ is constant, there's no way to scientifically prove it
I can think of one off the top of my head, even for your "knowledge and skill" definition.
Say we have a knowledge and skill test administered to adults every 10 years (20, 30, 40, 50, say).
Say we have another such test administered to 2-year-olds (clearly quite different; administering the same test to both makes no sense).
Now the big question is whether the five points we get for every person are correlated, and if so how much. If they're not correlated at all, then there's not much to talk about with these two tests. But if they're highly correlated, then what we have is a way to predict, at age 2, how well a person will score on knowledge and skill tests through the rest of their life. Which would imply that there might be, in fact, some underlying quality affecting performance on the tests involved.
Environmental effects could be controlled for in the standard ways with adoption studies and so forth, or simply by controlling for the environment, though this does reduce the number of people who can be studied.
I'm pretty sure the above is pretty much what's been done with IQ tests (or at least I would hope so!).
Now as far as calling this underlying factor "intelligence", that's a different kettle of fish. As you point out, intelligence has a wide range of denotations, not to mention connotations, and it's not clear that all of these would be covered by whatever factor it is our testing battery isolates.
A further interesting twist to the experiment would be attempting to establish whether there is a correlation between our test results and the ability to acquire new knowledge and skills. I would strongly suspect that _if_ all the test results correlate, then there will be some areas where skill-acquisition would also strongly correlate with the test results.
A good question is whether a test can be designed that would strongly correlate with skill-acquisition in a wide range of areas. I honestly don't know the answer to that, but I have a hard time believing it hasn't been studied.
Smart people also tend to have fewer kids, on average, than less smart people. At least today.
The artist has the right to pursue any model he sees fit, but that does not imply that some models should therefore receive government protection.
In other words, if the model works, great. But you are in fact advocating that the government support and encourage what would otherwise be an unsuccessful business model (mostly due to the fact that it's a poor business model nowadays).
> I'm just quite puzzled that Mozilla sees the need to build this into the browser.
I though you just agreed that this was important on mobile devices...
> If a tool can be abused, it will be abused.
That's an argument for having no tools at all, which isn't a good idea. The goal should be to strike a balance.
> It seems like Mozilla is now intent on adding features rather than having a great stable
> product.
That's a false dichotomy, really.
But there's one more issue here. All the browsers are in a position where they'll become totally irrelevant in a few years if they don't have feature parity with Flash and Silverlight... so they're working on said feature parity (except maybe IE, which doesn't mind being irrelevant if Silverlight wins).
> Would you be happy with a plugin that stored, and filled in your bank account login
> details when your bank's site asked for them?
If it had the option to never fill them in, and otherwise asked me every time? I might not mind, though I'd prefer those not be stored on my hard drive because my computer might be stolen. That part of the analogy doesn't apply here, since no additional data is being stored on your computer, though.
> It wouldn't be more than a few days until there was an exploit that could access that
> information without asking you.
If your opinion of browser security is that low, you should have much bigger worries than this API. You should be arguing for the removal of file inputs, for example. And of XMLHttpRequest. If those are subverted you lose a lot more than just your location.
> I even run Firefox as a separate user on my home system
OK, that does mitigate potential file input issues a good bit.... ;)
Yep, exactly.
> And why should it be in there by default?
That's a really good question. For now, the W3C is publishing a spec that will require it, but there've been rumblings to the effect that not all W3C specs are worth implementing...
Realistically, this feature is no more dangerous than many other things (like file inputs, etc) if properly secured. You're right that it increases attack area, but so does SVG. See, for example, http://lists.w3.org/Archives/Public/www-svg/2008Sep/0112.html
So the only real question is whether the benefits outweigh the risks. On the desktop, I'm not sure. On mobile devices, almost certainly (especially since those often have an existing way to pinpoint their location anyway).
You know what's funny? I can't figure out whether the "political enemies" you're talking about are Republicans or Democrats. They're both doing every single thing on your list (except possibly for the "give peace a chance" shirts; you'd have to have a different shirt to get hassled by Democrats).
Point is, both of the major parties are trying to expand the power of all levels of government (and hence their own) as fast as they can, at the expense of, first of all, citizens of the US, and secondly the lower levels of government (the ones more likely to actually respond to their constituents) as more and more power is concentrated upward.
> Why not enter it manually?
More effort on the user's part.
> In fact, if it is at the websites discretion, why even have this at all?
The website can ask for the information. Whether the browser tells it is at the user's discretion.
> that is the only rational reason to include I can think of; ...
> Further, I don't understand what is going through the heads
See http://www.w3.org/2008/geolocation/ and especially http://www.w3.org/2008/geolocation/charter/
That second link gives some examples of the use cases geolocation can address that are not addressed by existing technology.
I agree that on the desktop there's not as much use for this, but on mobile devices (think smartphones), it's a heck of a lot more useful.
If you feel strongly about how this should work and the current spec draft at http://dev.w3.org/geo/api/spec-source.html doesn't match that, you may want to send e-mail to public-geolocation@w3.org with your comments.
So what, exactly, is a web browser? Heck, what's "web"? The problem is that the definition keeps changing. Used to be, no image support was fine in a web browser. And no script support. And no CSS support. And no DOM support.
Note that the W3C is standardizing a geolocation API; this is just an implementation of that W3C spec.
I should also point out that Firefox 3 is faster and more memory-efficient than Firefox 2 or Firefox 1.5 or Firefox 1.0 (though when comparing to 1.0 you probably need to make sure to be comparing on a computer with a reasonably modern graphics card). It's certainly more secure than the older versions: a number of bugs have been fixed since the Firefox 1.0 days.
So what's with the "used to be" meme?
I'm curious: what is the exact problem you have with the geolocation stuff? No data is given to the server unless you explicitly allow it for that particular site. If you _do_ allow it, it enables a class of very interesting web applications, especially on mobile devices.
Could you explain what your objections are, past an instinctive reaction to the word "track"?
Lawsuits? The loss of goodwill would be a lot more important. Even more important would be the fact that a default of anything other than "nothing at all" is fundamentally the wrong thing to do.
I don't think you need to worry much on this score. The people working on this are neither stupid nor malicious.