Divisive issue - Microsoft does not design things with compatibility in mind.
Irrelevant response - It is not important why Microsoft does what it does. An end user considering this Linux disto is the one concerned with performance connecting to Exchange. You may not personally like the reason why, but thats moot.
Visual basic = not so great Microsoft code. Why the hell are people sending around large spreadsheets with shitty code?
Visual Basic is the most used programming language on the planet and your opinion of it is moot... and maybe this person happens to have many large spreadsheets with shitty code and doesnt want to pay a person for a year refactoring them.
WTF? Stop trolling. Active-X applications are the bane of open source...
Whos trolling? Your opinion of Active-X is moot and your open source leanings are also.
There are tons of applications out there for Linux, and as Linux gains market share, the quantity will only increase.
So he should castrate himself and then hope that in the future his nuts grow back?
As a programmer, I do not care about open file formats.
I care about SIMPLE file formats.
All these "open" formats end up being overly complicated for what they are predominantly used for. Obviously if CSV is good enough then it isnt wise to use ODS, but not so obvious is that if CSV is almost-good-enough then it STILL isnt wise to use ODS.
I want to write code to gather the data with minimal effort, without relying on a 3rd library specific to the format.
What really broils my sack are all these fields in these formats for attribution to who "owns" the data, what software created it, and so forth. I load that shit into a dummy variable and discard it when reading, so why is it there?
I am estimating that over half of my XP64 boot time is loading up crap that has nothing to do with the OS. As already mentioned, iTunes has its bootup horseshit. My webcam has its bootup horseshit. Games have DRM bootup horseshit. A good chunk of the software I have installed have their phone-home bootup horseshit. My motherboard driver package even installs several bootup horseshits whenever I grab an update. Video card drivers? Ditto.
The sad thing is that I have given up on monitoring it. I no longer care enough to police it because it is a never ending battle. Some programs, even when you remove their bootup horseshit, simply put it back in the next time you use that program.
An alternative OS isnt the answer because I like backward compatability. An OS that makes this horseshit more visible is. The startup process should be protected, yet apparently it isnt even on Linux or OSX.
Here in windows land, if I look at whats loaded during startup (msconfig) I get uninformative information for programs and services, such as "NvCpl: RUNDLL32.EXE NvCpl.DLL/starup".. I have to search the web to find out what this is, and invariably the site telling me what it is happens to be a site dedicated to virus protection of some kind. "This is nVidia blah blah blah, but a virus can be named anything!" they say, for everything. Gee thanks!
Lets not even mention the "services" listed with msconfig, where I can't even get a program path or filename of any kind.
Even something very simple could be extremely usefull, such as whenever a program sticks something in startup/services I get a prompt which allows to be place my own note about what I was doing when this happened, attached to the horseshit that got placed there, so that I can look at my own notes such as "nvidia display driver installation" so that when I run msconfig I can make rational decisions about the item in question.
Do you honestly think OpenGL would catch up if MS stopped developing DX? OpenGL isnt even up to DX9 standards yet, and probably will never be because the people who run that show are too diversified in areas that don't compare to DX (Microsoft doesnt decide if a feature stays, goes, or gets added based on what a small market share CAD programs need, but the OpenGL folks do)
Almost all RNG's are 32-bit. This isnt to say that they dont use more than 32-bits of state (Mersenne Twister uses a hell of a lot more state) but they still mostly only spit out 32-bit values or less.
Think of the birthday paradox here. Lets say that you are selecting 1 of 4294967296 different problems, that I already have 65536 of them solved, and that I will have my bot try to post rapidly 1024 times before giving up..
The odds are very good (much better than 50/50) that in one of these 1024 posting that it was challenged with one of the 65536 problems I already have solved..
The problem isnt as easy to solve as it sounds and will eventualy boil down to how well your PRNG gets seeded. Even if you use a true 64-bit or 128 bit generator, how do you go about seeding a generator with more than 32 unpredictable bits? 32 is already a tough problem and you can completely forget about open sourcing with these techniques.
Re:Sorry, Loebner Has Done Nothing for AI
on
Loebner Talks AI
·
· Score: 1
The bottom-up approach doesn't show any more promise.
The idea that since bottom-up gives good results in regards to Knowledge that it is evidence that bottom-up is making strides into A.I. is wrong. Its making strides in knowledge representation and that, my friend, is not A.I. Yes, some researchers and practicioners like to kid themselves into thinking they are dealing with A.I. but they are not. They certainly leverage the dead-ends of the A.I. field, but that doesnt make it A.I.
A.I. is machine-based problem solving, pure and simple. There are some very successfull renditions of this, such as game tree searching techniques (AlphaBeta, MID(f), NegaScout, and so on) as well as an entire subfield of A.I. called Machine Learning.
Neural Nets are not A.I., although the training techniques used might be. Neural Nets themselves are simply function approximators that have no more problem solving ability than their training sets had.
The rise of "Wintel" to its monopoly state is the victory of long term backward compatability. Their competition tried to do things the way you suggest, and the end result is that their competition is gone.
I don't suggest "forgiving" Microsoft because I don't think that they have anything to appologize for. Its their OS and the market speaks.
You clearly don't know the issue. The issue is that its users want to run application that do things which might break securuty, and this goes to the clear advantages of backward compatability that its users want. The vista method is to allow programs to break security, but only after prompting the user beforehand.
The widespread complaints about UAC is clear proof that backward compatability is of concern to its users because they are running programs which require it.
Mandatory insurance is not meant as a restriction on drivers who want to under/uninsure their vehicles, but to protect everyone else from them. If I get hit by a driver without insurance, *my* rates go up regardless if I was at fault or not.
You argue that I need to have auto insurance because you choose to have auto insurance.
You've spent your entire life being brainwashed by a bunch of bookies into thinking that you should place a bet with them that is statistically in their favor, and now you rationalize mandating this wager upon everyone because you engage in the gambling yourself.
Why do you suppose it is that during prime-time T.V. in america, there is always 4 or 5 commercials on each station pitching you an insurance policy of one kind or another?
Apart from the "fact" that is legalized gambling, I am interested to hear why the industry is insane.
It is not legalized gambling. It is state mandated gambling. Legalized gambling is those scratch-off tickets that you probably don't buy (because you do sometimes know what a bad bet looks like), or that casino on the indian reservation. This isnt anything like that at all. This isn't poker, either.
This is the State demanding that its population all place bets with bookies, and there is no chance that this bet is favorable towards the population (the insurance company cannot pay out more than it has recieved.)
Insane is the fact that the only way to turn this bet in your favor is to stack the deck by driving like a maniac, taking every risk in the book. Tired of your current car? Smash it into someone elses! If accidents make you nervous, report it stolen, rip apart the stearing column, and then push it off a cliff!
Your rates are higher because people like me who know how to turn a bad bet into a good one are forced to wager, my friend.
It may be illegal, but at least its not immoral (like mandatory insurance.)
This isnt like health insurance where you can opt out. How pleased would the local bookies be if they found out that you had a bit of a gambling problem and couldn't stop yourself from betting every single game?
Mandatory auto insurance is similar to a gambling problem, except society doesnt think that you are "sick" and "need help," instead they think "you better place a wager motherfucker!"
..and the sickest thing of them all, like most insurance, is that you are betting against your own safety and welfare.
Driver: "I'd like to bet $600 on my driving skills for one year please"
Bookie: "Do you want the over or the under?"
Driver: "I'll take the under, and dont expect me to drive any safer cause I'm trying to win!"
The insurance industry is insane.. and mandatory insurance is so far from being a reasonable idea that its hard to figure out if you should laugh or cry.
So you recommend bastardizing the term 'Shareware' instead of your pet favorite term 'Open Source', essentialy screwing the shareware guys because you have a hard-on for the open source guys? The shareware guys have been aroud a lot longer than you open source zealots, and arent fanatical about what they do, so I am here to be fanatical for them. Don't piss on someone elses patch of territory.
Certainly there are high end SSD's. But that is completely moot. You cannot relate HD to SSD performance so naively.
The point being that even a 30mb/sec SSD will be ready immediately after power-on, while a convention HD takes many *seconds* to spin up to full RPM, and only then lets the OS continue POST.
POST times on systems with HD's is often 10 or more seconds. If he says he can fully boot in 10 seconds on a HD then he has selected a HD that can fully spin up in under a second, which is anything but normal. Such a drive would be 3600 RPM (or a rare slower drive) with an abnormally large power consumption profile.
Note that 3600 RPM drives do not read any faster than these low end SSD's. Also, I have my doubts that any consumer level 3600 RPM drive exists that can spin up in under a second.
I dont doubt the original authors abilities to optimize the boot process, I just doubt his knowledge of the technical limitations of HD's.
..and by "slow consumer grade SSD" you mean one that seeks in less than 1ms (probably 0.1ms) *cough*
Those high end 15,000 RPM HD's, the fastest in the business, take an eternity to spin up compared to the instant readyness of low end SSD's.
SSD's are ideal for bragging about boot times, so I am quite suprised that some people have the delusion that HD's will offer comperable boot performance. No way in hell. The laws of physics and all that.
The 250GB decision wasnt made based on how many people it will effect.. it was based on how much profit the provider will make.
The cost/GB that they pay is and has been going down. Simultaneously, the GB/user has been going up. Its akin to a race.
Over time these caps will cut their bottom line significantly, removing the race from the equation.
This is the same ISP that is notorious for traffic shaping. As they say, "its the money, stupid!" Evidence that tey are just being greedy is that many other large ISP's (Metrocast for example) have not dived into either traffic shaping or capping.
If the RIAA grants MediaSentry the rights to distribute, and they do so to get on the "good side" of a pirates sharing quota, then pirates no longer need to pirate whatever it is being distributed.. MediaSentry will send it to you for free!!!!111one!!11one
I am curious as to why you think the spending habits of this government effects the economy so much. I see this sort of idea thrown around such that its simply taken for granted that its true, yet I havent seen any evidence that it actualy is.
The economy grew very rapidly during Reagans period of unprecedented deficits yet the economy also grew very rapidly during the Clinton period of reduced deficits. These two things dont jive well with the idea that the governments budget has much of an impact on the economy.
It is certainly true that during good times the administration that is in power at the time claims credit for it, but like I said, I havent seen any actual evidence of that... just a bunch of claims
For an economy to grow there must be unemployment, so when you admire Bill for leaving office with low unemployment numbers, you are also admiring Bill for setting up a period of low growth.
The university should spend money on hiring some admins with better computer skills and teaching skills rather than paying lawyers.
These are criminal charges. The university doesnt have to pay lawyers for this. The tax payers of Carlton County get to pay a D.A. to prosecute this kid.
The SSD's will take over without reaching competitive prices because they will offer significant performance advantages.
The "battleship mtron" is an example of what is comming. An array of SSD's packed into a high bandwidth raid-0 configuration already blows away physical disks in performance, approaching 1 gigabyte per second sustained transfer rates in moms-basement hack jobs.
SSD's will eventualy be by-default raid-0, where a SS-drive will be composed of many flash modules working in tandem to deliver arbitrarily high transfer rates. What is really lacking, currently, is a standardized I/O interface for high bandwidths. IDE is grossly inadequate, SATA's sights were set way too low, and the USB/Firewire interfaces are also significantly out of place for the comming bandwidths.
What we currently have are direct PCI-Express solutions in the form of raid-0 controller cards, but this puts the raid-0 controller on the wrong end of things. SSD manufacturers will be tailoring the controllers specifically for their devices in order to provide significant cost-benefits by packaging the entire configuration, and that means finding a bare-bones I/O path rather than piggybacking on someone elses raid-0 controller. It will be the return of the hard-cards, SSD style.
Indeed.. but if we are to truely remain true to topic... who in their right mind is going to play Ghost Busters 100,000 times? There is no "typical usage pattern" for Movies on USB Keys, and when there finally (if ever) is such a thing, its going to be way less than 100,000 times.
You said "Flash memory has a finite lifespan for the number of writes it can take, which these days works out to be several years."
Do you realize that the number of writes is not a measure of time, right?
Its a measure of the number of writes and cannot be translated into time. You would have to include a scaler value with a time dimension in order to translate the number of writes into time. One such scaler value might be writes per second (commonly denoted writes/second.) With both the mean maximum number of sustainable writes and the number of writes per second, we could then calculate some measure of device lifetime.
My 10 year old 64 meg compact flash card that originally came with my digital camera still works fine. Modern flash memory has a relative lifetime several orders of magnitude larger. Here, I get to evoke the measure of time without considering writes/second because I am simply comparing relative values both with a time dimension.
..it didnt say anything about how much memory each browser *requires*.. nothing at all. It went on and on about how much each browser used in the specific case of a computer with more than enough memory, and quite honestly, it didnt even do a good job of measuring resources since he just yanked a single value from a.net performance monitor, which, ironically, isnt recommended for what he is trying to measure.
But hey.. you made up your mind before ever reading the article, and certainly made up your mind before reading what I had to say, and completely ignored certain details such as the correct usage of the word 'required', and why the correct usage might be important in the context of comparing software.
I am happy to let a browser use up 4 gigs of free memory, as long as it still runs when only 64 megs are free. Firefox + all the extensions that bring it up to par with Opera will *NEVER* run in 64 megs of memory.. but Opera does.
Slashdot Mirror
Irrelevant response - It is not important why Microsoft does what it does. An end user considering this Linux disto is the one concerned with performance connecting to Exchange. You may not personally like the reason why, but thats moot.
Visual Basic is the most used programming language on the planet and your opinion of it is moot... and maybe this person happens to have many large spreadsheets with shitty code and doesnt want to pay a person for a year refactoring them.
Whos trolling? Your opinion of Active-X is moot and your open source leanings are also.
So he should castrate himself and then hope that in the future his nuts grow back?
As a programmer, I do not care about open file formats.
I care about SIMPLE file formats.
All these "open" formats end up being overly complicated for what they are predominantly used for. Obviously if CSV is good enough then it isnt wise to use ODS, but not so obvious is that if CSV is almost-good-enough then it STILL isnt wise to use ODS.
I want to write code to gather the data with minimal effort, without relying on a 3rd library specific to the format.
What really broils my sack are all these fields in these formats for attribution to who "owns" the data, what software created it, and so forth. I load that shit into a dummy variable and discard it when reading, so why is it there?
If DX development was dropped, those companies would still use DX. They arent going to downgrade to OpenGL.
Anecdotal Evidence
/starup" .. I have to search the web to find out what this is, and invariably the site telling me what it is happens to be a site dedicated to virus protection of some kind. "This is nVidia blah blah blah, but a virus can be named anything!" they say, for everything. Gee thanks!
I am estimating that over half of my XP64 boot time is loading up crap that has nothing to do with the OS. As already mentioned, iTunes has its bootup horseshit. My webcam has its bootup horseshit. Games have DRM bootup horseshit. A good chunk of the software I have installed have their phone-home bootup horseshit. My motherboard driver package even installs several bootup horseshits whenever I grab an update. Video card drivers? Ditto.
The sad thing is that I have given up on monitoring it. I no longer care enough to police it because it is a never ending battle. Some programs, even when you remove their bootup horseshit, simply put it back in the next time you use that program.
An alternative OS isnt the answer because I like backward compatability. An OS that makes this horseshit more visible is. The startup process should be protected, yet apparently it isnt even on Linux or OSX.
Here in windows land, if I look at whats loaded during startup (msconfig) I get uninformative information for programs and services, such as "NvCpl: RUNDLL32.EXE NvCpl.DLL
Lets not even mention the "services" listed with msconfig, where I can't even get a program path or filename of any kind.
Even something very simple could be extremely usefull, such as whenever a program sticks something in startup/services I get a prompt which allows to be place my own note about what I was doing when this happened, attached to the horseshit that got placed there, so that I can look at my own notes such as "nvidia display driver installation" so that when I run msconfig I can make rational decisions about the item in question.
Do you honestly think OpenGL would catch up if MS stopped developing DX? OpenGL isnt even up to DX9 standards yet, and probably will never be because the people who run that show are too diversified in areas that don't compare to DX (Microsoft doesnt decide if a feature stays, goes, or gets added based on what a small market share CAD programs need, but the OpenGL folks do)
I forsee bad implementations.
Almost all RNG's are 32-bit. This isnt to say that they dont use more than 32-bits of state (Mersenne Twister uses a hell of a lot more state) but they still mostly only spit out 32-bit values or less.
Think of the birthday paradox here. Lets say that you are selecting 1 of 4294967296 different problems, that I already have 65536 of them solved, and that I will have my bot try to post rapidly 1024 times before giving up..
The odds are very good (much better than 50/50) that in one of these 1024 posting that it was challenged with one of the 65536 problems I already have solved..
The problem isnt as easy to solve as it sounds and will eventualy boil down to how well your PRNG gets seeded. Even if you use a true 64-bit or 128 bit generator, how do you go about seeding a generator with more than 32 unpredictable bits? 32 is already a tough problem and you can completely forget about open sourcing with these techniques.
The bottom-up approach doesn't show any more promise.
The idea that since bottom-up gives good results in regards to Knowledge that it is evidence that bottom-up is making strides into A.I. is wrong. Its making strides in knowledge representation and that, my friend, is not A.I. Yes, some researchers and practicioners like to kid themselves into thinking they are dealing with A.I. but they are not. They certainly leverage the dead-ends of the A.I. field, but that doesnt make it A.I.
A.I. is machine-based problem solving, pure and simple. There are some very successfull renditions of this, such as game tree searching techniques (AlphaBeta, MID(f), NegaScout, and so on) as well as an entire subfield of A.I. called Machine Learning.
Neural Nets are not A.I., although the training techniques used might be. Neural Nets themselves are simply function approximators that have no more problem solving ability than their training sets had.
The rise of "Wintel" to its monopoly state is the victory of long term backward compatability. Their competition tried to do things the way you suggest, and the end result is that their competition is gone.
I don't suggest "forgiving" Microsoft because I don't think that they have anything to appologize for. Its their OS and the market speaks.
You clearly don't know the issue. The issue is that its users want to run application that do things which might break securuty, and this goes to the clear advantages of backward compatability that its users want. The vista method is to allow programs to break security, but only after prompting the user beforehand.
The widespread complaints about UAC is clear proof that backward compatability is of concern to its users because they are running programs which require it.
Mandatory insurance is not meant as a restriction on drivers who want to under/uninsure their vehicles, but to protect everyone else from them. If I get hit by a driver without insurance, *my* rates go up regardless if I was at fault or not.
You argue that I need to have auto insurance because you choose to have auto insurance.
You've spent your entire life being brainwashed by a bunch of bookies into thinking that you should place a bet with them that is statistically in their favor, and now you rationalize mandating this wager upon everyone because you engage in the gambling yourself.
Why do you suppose it is that during prime-time T.V. in america, there is always 4 or 5 commercials on each station pitching you an insurance policy of one kind or another?
Apart from the "fact" that is legalized gambling, I am interested to hear why the industry is insane.
It is not legalized gambling. It is state mandated gambling. Legalized gambling is those scratch-off tickets that you probably don't buy (because you do sometimes know what a bad bet looks like), or that casino on the indian reservation. This isnt anything like that at all. This isn't poker, either.
This is the State demanding that its population all place bets with bookies, and there is no chance that this bet is favorable towards the population (the insurance company cannot pay out more than it has recieved.)
Insane is the fact that the only way to turn this bet in your favor is to stack the deck by driving like a maniac, taking every risk in the book. Tired of your current car? Smash it into someone elses! If accidents make you nervous, report it stolen, rip apart the stearing column, and then push it off a cliff!
Your rates are higher because people like me who know how to turn a bad bet into a good one are forced to wager, my friend.
It may be illegal, but at least its not immoral (like mandatory insurance.)
This isnt like health insurance where you can opt out. How pleased would the local bookies be if they found out that you had a bit of a gambling problem and couldn't stop yourself from betting every single game?
..and the sickest thing of them all, like most insurance, is that you are betting against your own safety and welfare.
Mandatory auto insurance is similar to a gambling problem, except society doesnt think that you are "sick" and "need help," instead they think "you better place a wager motherfucker!"
Driver: "I'd like to bet $600 on my driving skills for one year please"
Bookie: "Do you want the over or the under?"
Driver: "I'll take the under, and dont expect me to drive any safer cause I'm trying to win!"
The insurance industry is insane.. and mandatory insurance is so far from being a reasonable idea that its hard to figure out if you should laugh or cry.
So you recommend bastardizing the term 'Shareware' instead of your pet favorite term 'Open Source', essentialy screwing the shareware guys because you have a hard-on for the open source guys? The shareware guys have been aroud a lot longer than you open source zealots, and arent fanatical about what they do, so I am here to be fanatical for them. Don't piss on someone elses patch of territory.
"...invokes the 'open source' label"
Who owns this label?
Certainly there are high end SSD's. But that is completely moot. You cannot relate HD to SSD performance so naively.
The point being that even a 30mb/sec SSD will be ready immediately after power-on, while a convention HD takes many *seconds* to spin up to full RPM, and only then lets the OS continue POST.
POST times on systems with HD's is often 10 or more seconds. If he says he can fully boot in 10 seconds on a HD then he has selected a HD that can fully spin up in under a second, which is anything but normal. Such a drive would be 3600 RPM (or a rare slower drive) with an abnormally large power consumption profile.
Note that 3600 RPM drives do not read any faster than these low end SSD's. Also, I have my doubts that any consumer level 3600 RPM drive exists that can spin up in under a second.
I dont doubt the original authors abilities to optimize the boot process, I just doubt his knowledge of the technical limitations of HD's.
..and by "slow consumer grade SSD" you mean one that seeks in less than 1ms (probably 0.1ms) *cough*
Those high end 15,000 RPM HD's, the fastest in the business, take an eternity to spin up compared to the instant readyness of low end SSD's.
SSD's are ideal for bragging about boot times, so I am quite suprised that some people have the delusion that HD's will offer comperable boot performance. No way in hell. The laws of physics and all that.
The 250GB decision wasnt made based on how many people it will effect.. it was based on how much profit the provider will make.
The cost/GB that they pay is and has been going down. Simultaneously, the GB/user has been going up. Its akin to a race.
Over time these caps will cut their bottom line significantly, removing the race from the equation.
This is the same ISP that is notorious for traffic shaping. As they say, "its the money, stupid!" Evidence that tey are just being greedy is that many other large ISP's (Metrocast for example) have not dived into either traffic shaping or capping.
If the RIAA grants MediaSentry the rights to distribute, and they do so to get on the "good side" of a pirates sharing quota, then pirates no longer need to pirate whatever it is being distributed.. MediaSentry will send it to you for free!!!!111one!!11one
I am curious as to why you think the spending habits of this government effects the economy so much. I see this sort of idea thrown around such that its simply taken for granted that its true, yet I havent seen any evidence that it actualy is.
The economy grew very rapidly during Reagans period of unprecedented deficits yet the economy also grew very rapidly during the Clinton period of reduced deficits. These two things dont jive well with the idea that the governments budget has much of an impact on the economy.
It is certainly true that during good times the administration that is in power at the time claims credit for it, but like I said, I havent seen any actual evidence of that... just a bunch of claims
For an economy to grow there must be unemployment, so when you admire Bill for leaving office with low unemployment numbers, you are also admiring Bill for setting up a period of low growth.
..and thats exactly what happened!
The university should spend money on hiring some admins with better computer skills and teaching skills rather than paying lawyers.
These are criminal charges. The university doesnt have to pay lawyers for this. The tax payers of Carlton County get to pay a D.A. to prosecute this kid.
3. Accept the challenge and lose, thereby destroying your niche market business.
Challenges such as this are not accepted by commercial enterprises because there isnt much to gain, but a lot to lose.
The SSD's will take over without reaching competitive prices because they will offer significant performance advantages.
The "battleship mtron" is an example of what is comming. An array of SSD's packed into a high bandwidth raid-0 configuration already blows away physical disks in performance, approaching 1 gigabyte per second sustained transfer rates in moms-basement hack jobs.
SSD's will eventualy be by-default raid-0, where a SS-drive will be composed of many flash modules working in tandem to deliver arbitrarily high transfer rates. What is really lacking, currently, is a standardized I/O interface for high bandwidths. IDE is grossly inadequate, SATA's sights were set way too low, and the USB/Firewire interfaces are also significantly out of place for the comming bandwidths.
What we currently have are direct PCI-Express solutions in the form of raid-0 controller cards, but this puts the raid-0 controller on the wrong end of things. SSD manufacturers will be tailoring the controllers specifically for their devices in order to provide significant cost-benefits by packaging the entire configuration, and that means finding a bare-bones I/O path rather than piggybacking on someone elses raid-0 controller. It will be the return of the hard-cards, SSD style.
Indeed.. but if we are to truely remain true to topic... who in their right mind is going to play Ghost Busters 100,000 times? There is no "typical usage pattern" for Movies on USB Keys, and when there finally (if ever) is such a thing, its going to be way less than 100,000 times.
You said "Flash memory has a finite lifespan for the number of writes it can take, which these days works out to be several years."
Do you realize that the number of writes is not a measure of time, right?
Its a measure of the number of writes and cannot be translated into time. You would have to include a scaler value with a time dimension in order to translate the number of writes into time. One such scaler value might be writes per second (commonly denoted writes/second.) With both the mean maximum number of sustainable writes and the number of writes per second, we could then calculate some measure of device lifetime.
My 10 year old 64 meg compact flash card that originally came with my digital camera still works fine. Modern flash memory has a relative lifetime several orders of magnitude larger. Here, I get to evoke the measure of time without considering writes/second because I am simply comparing relative values both with a time dimension.
Have a nice day.
I read the article...
..it didnt say anything about how much memory each browser *requires* .. nothing at all. It went on and on about how much each browser used in the specific case of a computer with more than enough memory, and quite honestly, it didnt even do a good job of measuring resources since he just yanked a single value from a .net performance monitor, which, ironically, isnt recommended for what he is trying to measure.
But hey.. you made up your mind before ever reading the article, and certainly made up your mind before reading what I had to say, and completely ignored certain details such as the correct usage of the word 'required', and why the correct usage might be important in the context of comparing software.
I am happy to let a browser use up 4 gigs of free memory, as long as it still runs when only 64 megs are free. Firefox + all the extensions that bring it up to par with Opera will *NEVER* run in 64 megs of memory.. but Opera does.