Clearly Rust isn't giving them a big boost in productivity, because Servo is making so little progress.
So Gecko is the result of thousands of man-years in development. On the other hand, servo probably had a few dozens of man-years of development time. How does that lead to the conclusion that Rust sucks exactly? I'm not qualified to say how successful Servo really is, but it was never expected to produce a direct Gecko replacement with 1/100 of the resources. Keep in mind that the stack of W3C specs browsers have to implement is *huge*.
Now we see Mozilla squandering more resources on dumb projects like Rust and Servo. Servo is, in my opinion, fucking atrocious. Try it for yourself. Really! See how goddamn awful it is. I tried it recently and I couldn't believe how bad it was. It makes Firefox look like a damn fine browser in comparison, that's how bad Servo is. Rust is just a hype-ridden joke in my experience.
(disclaimer: I work for Mozilla, but on codecs, not browsers) At this point, Servo is merely a proof-of-concept to experiment with new ways of doing rendering. The reason it sucks for you is that it's far from being feature-complete, and that's not even the point (yet). The point is to see if it's possible to write an engine that's both faster (because it runs in parallel) and safer (because of Rust) than current technology. Given the small team, the focus was on implementing things that were expected to be hard first (to show they were still possible), not implementing all the features. I've not been following the project too closely, but for the features it supports, it's already much faster than other browsers. And this is done by a rather tiny team (compared to Gecko). Turning it into a feature-complete would take a *lot* of people. I don't know if/when/how that decision will be made.
I think it all comes down to how "essential" a service is. You can be OK without seeing a particular artist or visiting a particular website. But an organization is often not OK if it can't use Visa and MasterCard. Look at the opposite case: if my state votes a law to protect LGBT people and some artist/company decides to boycott, well I didn't want to see them or do business with them anyway! Again, unless it's some kind of "essential" service.
The assumption is that the laser would point at a specific star system. Otherwise, you would not only need to operate the laser 365 days a year, but its power would need to be equal to all the power the Earth receives from the sun. Oh and hiding the Earth is only useful if you hide all the other planets too, otherwise the aliens could just assume there's life on Venus and go check it out.
The relentless advance of computing tech has far, far outpaced the computing needs of most people for at least the past 10-15 years.
I tend to see it the other way. Recent (last 5 years) advances haven't been enough to cause new, more compute-intensive applications to emerge and make the older hardware obsolete.
Maybe we read different books. Sure, the O-rings would partially burn, and then the engine blades would also tend to crack, and some of the tiles had to be replaced after every flight, and the SRBs had to be repaired after every flight, and so on. Your conclusion is that the entire engineering team there were totally incompetent since so many parts had issues. My conclusion is that given the design goals, having almost no safety factor on many parts became inevitable. The shuttle failed twice in slightly less than 200 mission, which is pretty close to the 1 failure in 100 figure Feynman got when talking to engineers and very far from the 1 in 10,000 he got from managers. That being said, I'd say the shuttle was incredibly safe compared to Apollo's one fatal accident and one near-fatal in less than 20 missions... which was also expected because they were pushing the limits even more.
Since you obviously know a lot about aerospace engineering, care to enlighten us on what the correct thing to do would have been for the engineers?
Risk 0 does not exist with rockets, much less in the late 1970s when the shuttle was designed, especially when you consider all the constraints that were imposed on the design (e.g. higher payload than the original idea). The O-ring design may or may not have been a good idea to begin with (I'm unqualified to judge here), but the thing is that they worked fine in all but the one flight where the engineers themselves warned they wouldn't hold. If some manager decides to launch during a hurricane, who's fault is it if it blows up?
If you design something and you tell people "it's going to kill someone if you use it in condition X", you may or may not be a good engineer depending on what condition X is, but the person responsible for the death is the one who carries on despite X.
I'd really like to see the Armenian version of that software. In France, it's a criminal offense to deny the Armenian genocide, while in Turkey it's a criminal offense to say there was an Armenian genocide. Isn't it nice?
There a lot more variability between individuals then between ethnic origins. Sure, the Olympics have demonstrated that at very high levels, black people win much more often than white people at a 100m race. But pick one black guy and one white guy completely at random on this planet. What's the likelihood of the black guy winning a 100m race? I'd say about 50%. The variability across individuals is just too much. Even if you had a precise way of measuring intelligence and managed to prove with 99.9% confidence that people of origin X are on average more intelligent than people of origin Y, it would mean absolutely nothing at the individual level.
Farmers (primary) have moved to become manufacturers (secondary). Then manufacturers are (mostly) moved to services (tertiary). Some of us have then moved to R&D (quaternary), but at this point it's far from clear that everyone can move there. So in the longer term, we'll really have to figure out how to make sure everyone can earn a living somehow.
I suggest you read the Mozilla post. Basically, the output *will* be free, with open-source software under the Apache 2.0 license, and the patents being licensed according to W3C rules. So it doesn't get any more free than that. As for DRM, it's not a property of the codec. There certainly won't be any in the codec itself, but people can put DRM on top of anything they want (including ASCII art).
Keep in mind that: 1) Apple can still join this effort. All the reasons why Microsoft, Cisco, Mozilla, and Google want a free codec also apply to Apple. 2) Even if Apple doesn't join the Alliance for Open Media, they would still benefit from using the codec 3) Even if they don't join and don't ship the codec, there's more than enough players already involved to make the resulting codec successful. That being said, 1) would still be the best outcome.
The problem with DRM on "rental" content isn't so much that it goes away (that part is the same for a book I borrow). The problem is that the only way to actually *implement* DRM is to have your machine is now obeying the content owner rather than you. To me this is like renting a DVD and leaving the key to your house at the store so that the clerk can enter your home when it's time to get the DVD back. The problem isn't that the DVD's going away, it's letting someone sneak into your house.
What I was essentially pointing out is there there's no clear binary decision between testable and untestable. There's stuff that's very hard to test, stuff we may be able to test in 1000 years, stuff we don't know if we'll ever be able to test,... And then when you have two theories that are "correct" wrt all the tests so far, then you have to use Occam's razor and pick the simplest. When you have hundreds of theory that all agree with experiments, then all the debate shifts to "Occam's razor-type arguments over which is most elegant/simplest". It's kinda unavoidable.
But what happens when you have multiple theories that agree with all the experiments we can make, and the only areas where they differ are for experiments we cannot make? Any of them would still be better than the current theories.
I googled "chinese cheating": got 22.6M results, top results are about exam cheating. I also googled "americans cheating", got 14.8M results, top results are about marital cheating.
So, China, with 4.2 times the US population has 1.5 times more cheaters. I guess the irrefutable conclusion from your data is that Americans cheat 2.75 times more than the Chinese, right?
Audio detection isn't nearly as broken as the article pretends. Sure if all you have is a single mic, then you have no hope. OTOH, with multiple mics, you can *localize* sounds, which means you don't need to recognize the sounds of a drone, just realize that there's some noise coming from something in the air where there shouldn't be anything. With a microphone array, you can actually pinpoint sound sources much weaker than ambient noise. It's certainly not trivial, but within the realm of what's realistic (assuming there aren't simpler solutions).
I do not know a single politician outside the US who would think that even remotely considering pushing an agenda as harebrained as creationism is anything but political suicide....well, except in Canada these days. A couple years ago, our minister of *science* was refusing to answer questions about whether he believed in evolution. More recently, Alberta also had a creationist minister of education.. So unfortunately, some of the madness has escaped North of the US.
Having lived in Australia a few years, I've been amazed at how good the voting system is (mandatory, with ranking)... and how bad the outcome has been (Howard at the time) despite the good system.
Similar here. One day the connection went out and I called tech support. I told them it was probably related to the technician I had just seen in the neighborhood. They couldn't even track that there was a technician around, so they couldn't help at all. Eventually (with tech support on the phone), I just opened the door and yelled "are you the one that took down my connection?" to the technician outside and he shouted back "yes". Cause identified.
Slashdot Mirror
No Silver Bullet
Clearly Rust isn't giving them a big boost in productivity, because Servo is making so little progress.
So Gecko is the result of thousands of man-years in development. On the other hand, servo probably had a few dozens of man-years of development time. How does that lead to the conclusion that Rust sucks exactly? I'm not qualified to say how successful Servo really is, but it was never expected to produce a direct Gecko replacement with 1/100 of the resources. Keep in mind that the stack of W3C specs browsers have to implement is *huge*.
Now we see Mozilla squandering more resources on dumb projects like Rust and Servo. Servo is, in my opinion, fucking atrocious. Try it for yourself. Really! See how goddamn awful it is. I tried it recently and I couldn't believe how bad it was. It makes Firefox look like a damn fine browser in comparison, that's how bad Servo is. Rust is just a hype-ridden joke in my experience.
(disclaimer: I work for Mozilla, but on codecs, not browsers)
At this point, Servo is merely a proof-of-concept to experiment with new ways of doing rendering. The reason it sucks for you is that it's far from being feature-complete, and that's not even the point (yet). The point is to see if it's possible to write an engine that's both faster (because it runs in parallel) and safer (because of Rust) than current technology. Given the small team, the focus was on implementing things that were expected to be hard first (to show they were still possible), not implementing all the features. I've not been following the project too closely, but for the features it supports, it's already much faster than other browsers. And this is done by a rather tiny team (compared to Gecko). Turning it into a feature-complete would take a *lot* of people. I don't know if/when/how that decision will be made.
I think it all comes down to how "essential" a service is. You can be OK without seeing a particular artist or visiting a particular website. But an organization is often not OK if it can't use Visa and MasterCard. Look at the opposite case: if my state votes a law to protect LGBT people and some artist/company decides to boycott, well I didn't want to see them or do business with them anyway! Again, unless it's some kind of "essential" service.
The assumption is that the laser would point at a specific star system. Otherwise, you would not only need to operate the laser 365 days a year, but its power would need to be equal to all the power the Earth receives from the sun. Oh and hiding the Earth is only useful if you hide all the other planets too, otherwise the aliens could just assume there's life on Venus and go check it out.
The relentless advance of computing tech has far, far outpaced the computing needs of most people for at least the past 10-15 years.
I tend to see it the other way. Recent (last 5 years) advances haven't been enough to cause new, more compute-intensive applications to emerge and make the older hardware obsolete.
Maybe we read different books. Sure, the O-rings would partially burn, and then the engine blades would also tend to crack, and some of the tiles had to be replaced after every flight, and the SRBs had to be repaired after every flight, and so on. Your conclusion is that the entire engineering team there were totally incompetent since so many parts had issues. My conclusion is that given the design goals, having almost no safety factor on many parts became inevitable. The shuttle failed twice in slightly less than 200 mission, which is pretty close to the 1 failure in 100 figure Feynman got when talking to engineers and very far from the 1 in 10,000 he got from managers. That being said, I'd say the shuttle was incredibly safe compared to Apollo's one fatal accident and one near-fatal in less than 20 missions... which was also expected because they were pushing the limits even more.
Since you obviously know a lot about aerospace engineering, care to enlighten us on what the correct thing to do would have been for the engineers?
Risk 0 does not exist with rockets, much less in the late 1970s when the shuttle was designed, especially when you consider all the constraints that were imposed on the design (e.g. higher payload than the original idea). The O-ring design may or may not have been a good idea to begin with (I'm unqualified to judge here), but the thing is that they worked fine in all but the one flight where the engineers themselves warned they wouldn't hold. If some manager decides to launch during a hurricane, who's fault is it if it blows up?
If you design something and you tell people "it's going to kill someone if you use it in condition X", you may or may not be a good engineer depending on what condition X is, but the person responsible for the death is the one who carries on despite X.
I certainly hope I'm not in an airplane relying on GPS when they discover such black hole. Imagine, so many planes with no navigation.
I'd really like to see the Armenian version of that software. In France, it's a criminal offense to deny the Armenian genocide, while in Turkey it's a criminal offense to say there was an Armenian genocide. Isn't it nice?
Settings are for advanced users, so we're getting rid of them. Using Gnome is now easier than ever!
There a lot more variability between individuals then between ethnic origins. Sure, the Olympics have demonstrated that at very high levels, black people win much more often than white people at a 100m race. But pick one black guy and one white guy completely at random on this planet. What's the likelihood of the black guy winning a 100m race? I'd say about 50%. The variability across individuals is just too much. Even if you had a precise way of measuring intelligence and managed to prove with 99.9% confidence that people of origin X are on average more intelligent than people of origin Y, it would mean absolutely nothing at the individual level.
Farmers (primary) have moved to become manufacturers (secondary). Then manufacturers are (mostly) moved to services (tertiary). Some of us have then moved to R&D (quaternary), but at this point it's far from clear that everyone can move there. So in the longer term, we'll really have to figure out how to make sure everyone can earn a living somehow.
AFAIK chlorine isn't used for disinfecting the water, but only to keep it clean as it travels in the pipes.
I suggest you read the Mozilla post. Basically, the output *will* be free, with open-source software under the Apache 2.0 license, and the patents being licensed according to W3C rules. So it doesn't get any more free than that. As for DRM, it's not a property of the codec. There certainly won't be any in the codec itself, but people can put DRM on top of anything they want (including ASCII art).
Keep in mind that:
1) Apple can still join this effort. All the reasons why Microsoft, Cisco, Mozilla, and Google want a free codec also apply to Apple.
2) Even if Apple doesn't join the Alliance for Open Media, they would still benefit from using the codec
3) Even if they don't join and don't ship the codec, there's more than enough players already involved to make the resulting codec successful.
That being said, 1) would still be the best outcome.
The problem with DRM on "rental" content isn't so much that it goes away (that part is the same for a book I borrow). The problem is that the only way to actually *implement* DRM is to have your machine is now obeying the content owner rather than you. To me this is like renting a DVD and leaving the key to your house at the store so that the clerk can enter your home when it's time to get the DVD back. The problem isn't that the DVD's going away, it's letting someone sneak into your house.
What I was essentially pointing out is there there's no clear binary decision between testable and untestable. There's stuff that's very hard to test, stuff we may be able to test in 1000 years, stuff we don't know if we'll ever be able to test, ... And then when you have two theories that are "correct" wrt all the tests so far, then you have to use Occam's razor and pick the simplest. When you have hundreds of theory that all agree with experiments, then all the debate shifts to "Occam's razor-type arguments over which is most elegant/simplest". It's kinda unavoidable.
But what happens when you have multiple theories that agree with all the experiments we can make, and the only areas where they differ are for experiments we cannot make? Any of them would still be better than the current theories.
I googled "chinese cheating": got 22.6M results, top results are about exam cheating.
I also googled "americans cheating", got 14.8M results, top results are about marital cheating.
So, China, with 4.2 times the US population has 1.5 times more cheaters. I guess the irrefutable conclusion from your data is that Americans cheat 2.75 times more than the Chinese, right?
Audio detection isn't nearly as broken as the article pretends. Sure if all you have is a single mic, then you have no hope. OTOH, with multiple mics, you can *localize* sounds, which means you don't need to recognize the sounds of a drone, just realize that there's some noise coming from something in the air where there shouldn't be anything. With a microphone array, you can actually pinpoint sound sources much weaker than ambient noise. It's certainly not trivial, but within the realm of what's realistic (assuming there aren't simpler solutions).
I do not know a single politician outside the US who would think that even remotely considering pushing an agenda as harebrained as creationism is anything but political suicide. ...well, except in Canada these days. A couple years ago, our minister of *science* was refusing to answer questions about whether he believed in evolution. More recently, Alberta also had a creationist minister of education.. So unfortunately, some of the madness has escaped North of the US.
Having lived in Australia a few years, I've been amazed at how good the voting system is (mandatory, with ranking)... and how bad the outcome has been (Howard at the time) despite the good system.
Similar here. One day the connection went out and I called tech support. I told them it was probably related to the technician I had just seen in the neighborhood. They couldn't even track that there was a technician around, so they couldn't help at all. Eventually (with tech support on the phone), I just opened the door and yelled "are you the one that took down my connection?" to the technician outside and he shouted back "yes". Cause identified.
Just what we needed to convince websites to switch to https!