1) How is someone auditing the code for the updater more closely going to change anything?
Updates are already digitally signed. What further advantage does this system offer to offset the disadvantage that a service can silently alter at least one previously trusted binary on my system?
What, am I supposed to pray to god that (A) the browser cant be fooled into downloading a bogus update, and (B) that the mozilla boys remain vigilant about protecting their private key, and that (C) nobodies botnet with more computing power than the #1 supercomputer on the planet cracks the key anyways?
I dont think that you understand real security methods, or what the real dangers are.
How is it SAFER, or BETTER? This is no better than The Same as Now, only with dangerous silent updates added.
One can take the word 'directed' and infer that there must be an intelligent Director
One can also take the word "directed" and infer that there must be a "producer" too (and "investors", and a "studio")
Some people might infer crazy whacked out shit, but in general you dont defend against the dumbest of the dumb.
The point of "Directed Random Search" is that it does not have the romance bullshit like "evolutionary", "genetic", "population" and other assorted crap that *really* muddies the waters.
I don't know how Windows ACLs work in depth, but if the updater runs as a user that can't write outside/Program Files/Mozilla Firefox, that's another way to limit the damage it can do.
It can't if it runs as a regular user. Thats sort of the point.
The article summary claims silent updating, so the service can't run as the logged in user.
I do remember the days when they said x86 couldn't compete with the RISC chips (alpha/mips/etc..)
The reality is that (A) all processors offer instruction forms that are rarely used, and (B) complex instruction sets increase code density as long as the common ones are short. (C) extensive amounts of registers arent all that useful in most circumstances because most long term data has to be written back to memory anyways.
The read/modify/write instructions are often the best instructions to use these days on x86 because of code density. Instruction scheduling isnt nearly as important as it once was, so RISCifying your x86 assembler code is often a mistake.
I also thought about this almost immediately. You cant silently do anything under/Program Files/ or/Program Files (x86)/ without administrator rights.
Is it their intention to install the binaries/etc some place that doesnt require admin rights to modify them? How could that possibly be safer/better?
Maybe instead they intend to install a service set up with admin privileges. How is an extra service, with admin and network access rights and intent on modifying/program files/, safer/better?
Updates should be noisy anyways, regardless of the increased risks this sort of thing creates.
Yes, economies of scale are a valid argument, despite whatever you're smoking. Bigger companies can buy in bigger quantities, and get better discounts.
Companies can enter the market as a big player. This can be a new startup with a lot of investors, or an existing company entering a new market. If you dont address that point then you are just repeating the refuted argument.
As far as Standard Oil, it was broken up and rightly so. But you are forgetting that Standard Oil's main leverage was railroad discrimination, and the existing railroads had a government protected monopoly. The government created Standard Oil's advantage.
Unfortunately, the Rand-worshiping free-market fans almost always forget about the role government has in ensuring the marketplace remains a level playing field.
The rand-worshipers that I have heard have pointed out the role that government has in enabling and creating monopolies.
Patents, Licenses, Grants, etc.. barriers to entry.. brought to you by Uncle Sam.
Economies of scale isnt a valid argument. Even in mature markets with only 2 or 3 competitors, its not unheard of for a new competitor to enter the market and be successful. If there is enough money to be made, investors will follow, and start-ups can be huge from the outset because of it.
When I was growing up, there was no Walmart. There was K-Mart, Caldors, and Bradlees. Caldor is notable because it was the largest chain in the country. Now its gone.
Right now, that advantage is tilted so far in Intel's favor that 4 Intel Nehalem cores are generally going to be better than 6 AMD cores at the same clock.
The evidence doesnt support your assertion. Yes, Core2 is more efficient, but you are imagining that its 50% better clock-for-clock (that 4 Intel cores beats 6 AMD cores) You might be able to find some obscure task where thats the case, but in general its just not true.
They had expensive high end models back when they had the aces (when it was Athlon 64 / Opteron versus Inte'ls Pentium 4 line). Now they don't.
Um, now you are talking about server chips (what the hell do you think an opteron is) Intel had server chips too.. but you want to compare AMD server prices with Pentium 4 prices? Please. You arent even being honest with yourself here.
Lets talk about server chips, where AMD is beating the snot out of Intel with 12-core chips vs Intel's 8.
The mark of a liar is not being honest with others. The mark of a fanboy is not even being honest with yourself. Stop being a fanboy. Deep down you know that you are arguing without facts to support your assertions, so why do you do it?
I totally agree. A bear is much better than even several dogs. Even in the case of several intruders, while the dogs could tackle them both simultaneously, the bear makes short work of even the largest thug.
You said 4.2ghz, now you want to talk about 4.5ghz.
What gives in your moving goalpost? This is just more evidence that you dont know what you are talking about. 1090T's have been overclocked to 4.5ghz too. Before you were talking about "easy" overclocking, now you moved to goalpost to the extremes.
The fact is that the *6* core 1090T is going to outperform that 4 core 930 in multi-threaded tasks when clocked at the same rate. You have not presented evidence that the 930 is more overclockable than the 1090T, but have waved your hands claiming that its true. Meanwhile you moved the goalpost from 4.2ghz to 4.5ghz because you previously didnt know that 1090T's are easily overclockable to 4ghz and beyond, which you claimed wasn't true before. You made it up.
Now apparently you didn't know that 1090T's have been overclocked to 4.5ghz on air as well.
Let me clue you in on why machines are so overclockable today. The clock rates on these CPU's are set based on a portion of the market boxing up poor cooling conditions (pick any cramped Dell/HP box) together with consumers maybe not having an air conditioned environment (41C ambient temperatures are not surprising.) This pushes both AMD and Intel, who offer warranties on their products, to clock their chips far below how well they can perform under more ideal conditions. A segment of the market has atrocious cooling conditions and thats why we can reliably push these chips so far above spec.
Even the 1055T, a $200 2.8ghz chip, easily overclocks to 4ghz and beyond with the stock heatsink and fan. Thats 68% of the price of the i7 930 chip, with 50% more cores.
And since you want to drag the e-peen into this, 1090T overclocked to 6.5ghz. Yeah.. we can brag about large overclocks but only "easy" overclocks are relevant. All these chips go over 4ghz on air. You didnt know that. Now you do. Leave your e-peen at the door.
For any sort of multi-threaded performance, AMD easily wins at every price point.
Of course a $1000 CPU (6-core i7 980 extreme edition) beats a $200 CPU (6-core AMD 1055T). But when you start looking at AMD's server chips, which are in the same price range as that i7 980 extreme ripoff, you are looking at getting a 12 core chip that easily destroys Intel's 6-core chip in multi-threaded applications, and has far greater i/o bandwidth too.
I'm sure that Intel will eventually start taking many-core seriously enough to be price competitive, but right now it just looks like they got caught with their pants down. You cannot buy more than 4 cores from Intel for less than a grand, while for less than a grand you can get a 12-core chip from AMD.
We expected better from the years of Larrabee research. We were wrong.
The 1090T overclocks to over 4ghz on air cooling and this was done by many people the first week the chips were out so how could you possibly not know this yet speak so authoritatively?
Why did you speak at all, knowing yourself that you dont know what you are fucking talking about? Did you think that nobody on slashdot would call you out on your obviously made up bullshit?
Intel does not have any i5 that is even close in performance with the higher end 1090T, which is what the poster you were replying to said he was talking about. Read that? Not Even Close.
The lower end 1055T (which you are talking about) also beats the best performing i5, the 760, and it is cheaper than Intels chip too.
On top of that, the OEM special-edition 1035T, even cheaper than the 1055T, also outperforms all the i5's.
The only thing the i5 does better than the AMD 6 core offerings is better single threaded integer performance (and thats only the best most expensive i5), but is worse at single threaded floating point. For multi-threaded tasks it gets literally destroyed by AMD's 6-core offerings.
But in that case, didnt the camera itself already stagger the time that each photo element was collecting data? That effectively it was already taking what amounted to 4 successive images with staggered pixels?
Making people pay for costs that they've been getting away with externalizing is not "artificially increasing" the price.
We arent talking about paying external costs. We are talking about charging for external costs. There has been no proposal for the proceeds to be used to pay for them. Instead, the proposals are to subsidize alternatives, and I think we all know thats bullshit too. The end result will be that governments will spend the money on whatever pork they can.
Yes. Artificially increasing the price of energy will harm the poorest of the poor, and increase poverty and misery throughout the world.
This. Seriously.
There are 250 million people below the poverty line in India alone, and Americans and Europeans really dont understand what poverty means elsewhere in the world. People that can only get antibiotics (which are dirt cheap!) if someone donates them. People that dont even have access to clean water, let alone running clean water. People that have no idea where their very next meal will come from. People that have literally nothing but some tattered rags to wear.
Compared to poverty, global warming nothing. Climate change is code for control over industry.
You do realize that the rate of warming at the end of the last glaciation is about two orders of magnitude than the rate of warming we're experiencing today, right?
You know this, how?
Proxies that dont have yearly resolution cannot be used to make judgments about yearly temperature variations. We are talking about proxies that have centennial resolution here.
Even if we suppose that the proxies are good proxies, we cant know what you are claiming we know.
I see articles everyday that satisfaction is low among Facebook users. They are hanging around, in part, because there aren't any worthy alternatives from their perspective.
"These are not the satisfactions that you are looking for."
People are bored with Facebook. A replacement Facebook is just as non-compelling as Facebook is now. They use Facebook because they are expected to: "Did you see Johns Facebook status?" The novelty has worn off and now people are there because people are there. They arent looking for an alternative.
The only people that want an alternative are people that havent grown tired of it because they've been apprehensive from the start about privacy.
Slashdot Mirror
1) How is someone auditing the code for the updater more closely going to change anything?
Updates are already digitally signed. What further advantage does this system offer to offset the disadvantage that a service can silently alter at least one previously trusted binary on my system?
What, am I supposed to pray to god that (A) the browser cant be fooled into downloading a bogus update, and (B) that the mozilla boys remain vigilant about protecting their private key, and that (C) nobodies botnet with more computing power than the #1 supercomputer on the planet cracks the key anyways?
I dont think that you understand real security methods, or what the real dangers are.
How is it SAFER, or BETTER? This is no better than The Same as Now, only with dangerous silent updates added.
That's why it runs as a different user with appropriate permissions.
Go back to the first post of mine that you replied to. How is having such a service safer/better?
One can take the word 'directed' and infer that there must be an intelligent Director
One can also take the word "directed" and infer that there must be a "producer" too (and "investors", and a "studio")
Some people might infer crazy whacked out shit, but in general you dont defend against the dumbest of the dumb.
The point of "Directed Random Search" is that it does not have the romance bullshit like "evolutionary", "genetic", "population" and other assorted crap that *really* muddies the waters.
You think network administrators own the networks they administer? So that was Terry Childs network?
Genetic/Evolutionary Algorithms is best (but not often) summed up as A Directed Random Search.
In this case, the laws are pretty reasonable: Sysadmins can't lock everyone else out of a system they don't own.
Wasn't his job to lock others out?
Exactly what do you think a network administrator does?
I don't know how Windows ACLs work in depth, but if the updater runs as a user that can't write outside /Program Files/Mozilla Firefox, that's another way to limit the damage it can do.
It can't if it runs as a regular user. Thats sort of the point.
The article summary claims silent updating, so the service can't run as the logged in user.
How big is that overclock?
PCPro's $1000 6-core i7 980 Extreme (which is stock 3.33ghz) took 73 seconds.
My $200 6-core Phenom II 1055T (overclocked to 3.34ghz) took 188 seconds.
Your $200 6-core Phenom II 1055T took 44 seconds.
Thats a big assed difference. I was using the same executable as PCPro (the Win32 version of smallpt)
The pcpro smallpt distro is a 32-bit application and uses the 32-bit version of OpenMP.
I do remember the days when they said x86 couldn't compete with the RISC chips (alpha/mips/etc..)
The reality is that (A) all processors offer instruction forms that are rarely used, and (B) complex instruction sets increase code density as long as the common ones are short. (C) extensive amounts of registers arent all that useful in most circumstances because most long term data has to be written back to memory anyways.
The read/modify/write instructions are often the best instructions to use these days on x86 because of code density. Instruction scheduling isnt nearly as important as it once was, so RISCifying your x86 assembler code is often a mistake.
Both CISC and RISC have their pro's and con's.
I also thought about this almost immediately. You cant silently do anything under /Program Files/ or /Program Files (x86)/ without administrator rights.
/program files/, safer/better?
Is it their intention to install the binaries/etc some place that doesnt require admin rights to modify them? How could that possibly be safer/better?
Maybe instead they intend to install a service set up with admin privileges. How is an extra service, with admin and network access rights and intent on modifying
Updates should be noisy anyways, regardless of the increased risks this sort of thing creates.
Yes, economies of scale are a valid argument, despite whatever you're smoking. Bigger companies can buy in bigger quantities, and get better discounts.
Companies can enter the market as a big player. This can be a new startup with a lot of investors, or an existing company entering a new market. If you dont address that point then you are just repeating the refuted argument.
As far as Standard Oil, it was broken up and rightly so. But you are forgetting that Standard Oil's main leverage was railroad discrimination, and the existing railroads had a government protected monopoly. The government created Standard Oil's advantage.
Unfortunately, the Rand-worshiping free-market fans almost always forget about the role government has in ensuring the marketplace remains a level playing field.
The rand-worshipers that I have heard have pointed out the role that government has in enabling and creating monopolies.
Patents, Licenses, Grants, etc.. barriers to entry.. brought to you by Uncle Sam.
Economies of scale isnt a valid argument. Even in mature markets with only 2 or 3 competitors, its not unheard of for a new competitor to enter the market and be successful. If there is enough money to be made, investors will follow, and start-ups can be huge from the outset because of it.
When I was growing up, there was no Walmart. There was K-Mart, Caldors, and Bradlees. Caldor is notable because it was the largest chain in the country. Now its gone.
Right now, that advantage is tilted so far in Intel's favor that 4 Intel Nehalem cores are generally going to be better than 6 AMD cores at the same clock.
The evidence doesnt support your assertion. Yes, Core2 is more efficient, but you are imagining that its 50% better clock-for-clock (that 4 Intel cores beats 6 AMD cores) You might be able to find some obscure task where thats the case, but in general its just not true.
They had expensive high end models back when they had the aces (when it was Athlon 64 / Opteron versus Inte'ls Pentium 4 line). Now they don't.
Um, now you are talking about server chips (what the hell do you think an opteron is) Intel had server chips too.. but you want to compare AMD server prices with Pentium 4 prices? Please. You arent even being honest with yourself here.
Lets talk about server chips, where AMD is beating the snot out of Intel with 12-core chips vs Intel's 8.
Intel isn't winning at the high end of servers because you simply cannot build a 48-core server built on Intel. The best you can do is 32 cores.
The mark of a liar is not being honest with others. The mark of a fanboy is not even being honest with yourself. Stop being a fanboy. Deep down you know that you are arguing without facts to support your assertions, so why do you do it?
Dogs are useful and all, but bear ..
I totally agree. A bear is much better than even several dogs. Even in the case of several intruders, while the dogs could tackle them both simultaneously, the bear makes short work of even the largest thug.
You said 4.2ghz, now you want to talk about 4.5ghz.
What gives in your moving goalpost? This is just more evidence that you dont know what you are talking about. 1090T's have been overclocked to 4.5ghz too. Before you were talking about "easy" overclocking, now you moved to goalpost to the extremes.
The fact is that the *6* core 1090T is going to outperform that 4 core 930 in multi-threaded tasks when clocked at the same rate. You have not presented evidence that the 930 is more overclockable than the 1090T, but have waved your hands claiming that its true. Meanwhile you moved the goalpost from 4.2ghz to 4.5ghz because you previously didnt know that 1090T's are easily overclockable to 4ghz and beyond, which you claimed wasn't true before. You made it up.
Now apparently you didn't know that 1090T's have been overclocked to 4.5ghz on air as well.
Let me clue you in on why machines are so overclockable today. The clock rates on these CPU's are set based on a portion of the market boxing up poor cooling conditions (pick any cramped Dell/HP box) together with consumers maybe not having an air conditioned environment (41C ambient temperatures are not surprising.) This pushes both AMD and Intel, who offer warranties on their products, to clock their chips far below how well they can perform under more ideal conditions. A segment of the market has atrocious cooling conditions and thats why we can reliably push these chips so far above spec.
Even the 1055T, a $200 2.8ghz chip, easily overclocks to 4ghz and beyond with the stock heatsink and fan. Thats 68% of the price of the i7 930 chip, with 50% more cores.
And since you want to drag the e-peen into this, 1090T overclocked to 6.5ghz. Yeah.. we can brag about large overclocks but only "easy" overclocks are relevant. All these chips go over 4ghz on air. You didnt know that. Now you do. Leave your e-peen at the door.
This is exactly it.
For any sort of multi-threaded performance, AMD easily wins at every price point.
Of course a $1000 CPU (6-core i7 980 extreme edition) beats a $200 CPU (6-core AMD 1055T). But when you start looking at AMD's server chips, which are in the same price range as that i7 980 extreme ripoff, you are looking at getting a 12 core chip that easily destroys Intel's 6-core chip in multi-threaded applications, and has far greater i/o bandwidth too.
I'm sure that Intel will eventually start taking many-core seriously enough to be price competitive, but right now it just looks like they got caught with their pants down. You cannot buy more than 4 cores from Intel for less than a grand, while for less than a grand you can get a 12-core chip from AMD.
We expected better from the years of Larrabee research. We were wrong.
The 1090T overclocks to over 4ghz on air cooling and this was done by many people the first week the chips were out so how could you possibly not know this yet speak so authoritatively?
Why did you speak at all, knowing yourself that you dont know what you are fucking talking about? Did you think that nobody on slashdot would call you out on your obviously made up bullshit?
The benchmarks I've seen show even an i5 being competitive with a Phenom II X6
I am backing up my assertions.
Intel does not have any i5 that is even close in performance with the higher end 1090T, which is what the poster you were replying to said he was talking about. Read that? Not Even Close.
The lower end 1055T (which you are talking about) also beats the best performing i5, the 760, and it is cheaper than Intels chip too.
On top of that, the OEM special-edition 1035T, even cheaper than the 1055T, also outperforms all the i5's.
The only thing the i5 does better than the AMD 6 core offerings is better single threaded integer performance (and thats only the best most expensive i5), but is worse at single threaded floating point. For multi-threaded tasks it gets literally destroyed by AMD's 6-core offerings.
Its that Europeans (on average) dont upgrade their hardware as often. This leads to increased specializations.
Americans jumped off the C64 bandwagon long before the Europeans did, and it happened again with Amiga and IIgs.
But in that case, didnt the camera itself already stagger the time that each photo element was collecting data? That effectively it was already taking what amounted to 4 successive images with staggered pixels?
Making people pay for costs that they've been getting away with externalizing is not "artificially increasing" the price.
We arent talking about paying external costs. We are talking about charging for external costs. There has been no proposal for the proceeds to be used to pay for them. Instead, the proposals are to subsidize alternatives, and I think we all know thats bullshit too. The end result will be that governments will spend the money on whatever pork they can.
Yes. Artificially increasing the price of energy will harm the poorest of the poor, and increase poverty and misery throughout the world.
This. Seriously.
There are 250 million people below the poverty line in India alone, and Americans and Europeans really dont understand what poverty means elsewhere in the world. People that can only get antibiotics (which are dirt cheap!) if someone donates them. People that dont even have access to clean water, let alone running clean water. People that have no idea where their very next meal will come from. People that have literally nothing but some tattered rags to wear.
Compared to poverty, global warming nothing. Climate change is code for control over industry.
You do realize that the rate of warming at the end of the last glaciation is about two orders of magnitude than the rate of warming we're experiencing today, right?
You know this, how?
Proxies that dont have yearly resolution cannot be used to make judgments about yearly temperature variations. We are talking about proxies that have centennial resolution here.
Even if we suppose that the proxies are good proxies, we cant know what you are claiming we know.
I see articles everyday that satisfaction is low among Facebook users. They are hanging around, in part, because there aren't any worthy alternatives from their perspective.
"These are not the satisfactions that you are looking for."
People are bored with Facebook. A replacement Facebook is just as non-compelling as Facebook is now. They use Facebook because they are expected to: "Did you see Johns Facebook status?" The novelty has worn off and now people are there because people are there. They arent looking for an alternative.
The only people that want an alternative are people that havent grown tired of it because they've been apprehensive from the start about privacy.