Organic molecules involve many carbon atoms. One way of removing carbon is to grow forests, cut them down and turn the wood into buildings.
Rather than spending megabucks to dump carbon into a hole, why not ship wood to the third world, so long as they don't burn it?
Wood is a lot safer than carbon dioxide stored somewhere, which can suddenly escape.
One of causes of unaffordable housing is the lack of it. Who wants to work for 30 years just to pay off a house? Meanwhile we hear that the rich are getting richer and everyone else isn't. Simultaneously people are driving around emitting carbon dioxide, driving the environment to ruin, trying tomake ends meet. Raising real-estate quality in the third world helps the poor and reduces the environmental impact of our efforts just to live. Furthermore, industries that have been hurting from subprime can be rejuvenated. It just gets better and better.
Just wondering if this will lead to an alternative energy source, like building a dwarf star under the hood. That would cost about as much as a tank of gas anyways.
There are a good many places in my city advertising free Wi-Fi, but that's not bringing any customers with computers. However, many people go to the library with their laptops. Lots of reasons for the discrepancy: the lack of electrical outlets at the businessplaces, maybe people feel more private at the library, and the library's Internet is really quick.
If the Wi-Fi isn't supposed to be free, but is available for a fee, then there must be a fairly good demand! What makes people pay for their Internet outside the home, perhaps aside from not having Internet at home? I guess there are some places where the population density is so high, there's a signal even in the home from a pay-to-use hotspot.
But I agree, most of it is repetitive. Every time I start a new project, I think, "Oh crap, I have to do this AGAIN!?!?"
Well, that's work. If there wasn't work to do, how do you earn money?
Actually, it's good that it's so repetitive because it means you don't have to invent something for every little cotton picking detail. So much structure and so many paradigms have been recognized and instilled that it becomes easier to convince people to pay for creating software that takes things to the next level.
Software for business needs to make a quantum leap: natural language processing, audio and visual input, and then virtual reality and artificial intelligence. Someone working as an IT grunt won't be inventing these funky toys. An IT developer just brings technology together.
If you want to work as an IT developer on programs that have more variety, perhaps find jobs programming small tools that don't have many end users. These programs tend to be cut to the bone and the user might not want enough features to provide full time employment, but there is a bit of urgency to have the computer do something exotic. Only thing is, these days many young people know how to write a teensy program to crank on small chunks of data, even if it's O(n^2) more inefficient than it could be and takes O(n^2) more debugging time.
HTML form input straight into a SQL query is so... fucking... bad
Some time ago I wrote a debugging routine to output some info to a browser, and unwittingly output HTML code embedded in the info. Little experiences like that can teach one about vulnerabilities and bizarreness of computers following your every command. FTFA, "little Bobby Tables" is quite illuminating:) and I'm happy to say that all my program-generated SQL queries have always been constructed to treat text as text by escaping single quotes so I didn't become vulnerable to such antics.
I prefer deep nesting, a nasty code to read as it is when that's what it takes to do the job. As I see it, if you outsource:) the inner loops you may have declared a number of local variables many of which may be relevant at a feature change's notice to an inner loop. If you have to call a function of a function, which looking good as it may be if done right, the function becomes a function of many parameters.
A habit of mine is to start writing code into the same function if it is not immediately known whether there will be a separate function caller of portions of the function. However, coupling through local variables is kept to a minimum between obvious segments that look separable. After debugging everything, which is easier in a single function than skipping all over in a trace, separation into different functions is painless. The separation could have been done to start with, but the coupling through local variables and big parameterization give me pause.
BTW, I see unit tests as good things for lengthy runs or runs with lots of particular inputs. However, I've avoided unit tests lately to keep costs down. Instead I pretty much know what I do is going to work, run a couple of tests, and things are fine, except lately I've seen some odd bugs due to Microsoft Windows and Office--that's where the costs really occur, making unit tests seem more a requirement except how do you unit test visual output from Office without writing almost as much code as was written?
So the thread all goes:
A better way to handle that is to turn the loop body into a function or group of functions. makes the code easier to read and a good compiler will inline the function so their's no performance loss.
I somewhat disagree with what you and... *sigh* Monkeybaister posted. Yes, there are many times when long stretches of code should be broken out into functions. But I tend to do that mostly when the same bit of code is used in several different cases. The reason being is that when you start modularizing off all your while loops that are more than a dozen lines long, you create a whole new type of spaghetti code. I'm going to coin a term and call it "spaghetti-O code." You try to track down a bug and what would have been a straightforward couple pages of code now has all kinds of functions at different places in the code. As such, it can often make debugging or forming a mental map of the code much harder
When you consider the sheer amount of information required for a person to cope with reality, that's still a costly amount for computing. There are anecdotes of people under hypnosis being able to recall minute details of what they've seen even though these details would hardly rate any attention. The mind may well be storing high definition video of everything you see--how much storage would that require?
If an AI were to build an idea of its environment from video, audio, and tactile inputs, Seagate's first hard disks would have been fairly prohibitive, but hard disks for PCs might be on the thresholds of affordability.
If you want real performance and aren't afraid of having to do a complete rebuild on a regular basis then the best bet is to purely use a huge amount of RAM, not Flash or other solid state disks but real genuine RAM.
Okay so its insanely expensive and a power cut and UPS failure means you lose everything.... but the SPEED is fantastic
Talking about speed, this is an effective design. Multiple UPS and a separate computer that maintains the RAM will give reliability. Not sure if it's worth it just to boot faster, and I admit being scared by the volatility. Hard disk speedups have not been so easy to come by. It took 7 minutes to copy a 3 Gb file and everything felt slow when that was happening.
Having plenty of RAM does speed things up because files are cached in RAM. I do not miss the days when I had a lot more swapping even on a 1 Gb RAM computer.
Does anyone have any ideas about multiple heads? If the heads are swinging independently, the mechanics are quite complex, but what if the heads all swing in unison - all together at the same direction and speed? Then the heads can spend less time per cylinder. Heads can also be given a positional offset in order to be on different cylinders at the same time. Complex mechanics, but in today's level of technology, par for the course as the saying goes. Even a slower RPM drive can still have the performance of a fast RPM drive. If the drive isn't too busy, some heads can be kept parked. RPM Speedstep may also contribute to power savings.
Exactly. You can get a 500 GB hard disk for $100. Why not just use those for backup. At the cost of this holographic storage, you could buy 2 or 3 500 GB hard drives, and keep multiple copies just in case one died. Since you'd only be using them for backup, and they would actually get very little wear and tear, I would guess that it would be easy to have a hard drive last for 50 years
But let's think of the possibilities. Backing up is the tip of the iceberg. Maybe you would be nuts to buy the product, but it still proves a concept, and that can be enough to think of taking the business to the next level.
Because lasers are being used, the size of the storage medium might be quite large. Too large to fit in a computer? That's ok. Think about the holodeck in Star Trek. Hi-res display device. Walking about a room full of lasers does give one pause though.
But it makes planning an exploit much harder. Before they might have been able to say they had 12 minutes (say) between sweeps, giving them that amount of time to get through a door, set a bomb, whatever. Now they might have an AVERAGE of 12 minutes, and possibly just 2 minutes Much more risky, and if they have to pass through more than one such point, they're almost sure to get caught
But now the left hand might not know what the right hand is doing so an impostor can show up at a door and say "I'm running a little late," which would happen a lot if there are short gaps between arrivals.
If the robot looked for a target without ever being commanded that'd be a pretty horrendous software error.
There are two sides on this coin. Heads: robots are more expendable than people, and intimidating, trigger-happy, seemingly out-of-control robots can scare enough bejesus out of militant insurgents to turn the tide and keep terrorists to themselves. Tails: a robot can be captured by the enemy and leads to the scenario, unlikely as it may be, that it is sent back with enough sneakiness to gun down commanding officers.
Now we leave you with a few words from Weird Al Yankovic: Trigger happy! Trigger happy every day!
At my alma mater, engineering students in the Society of Automotive Engineers club of University of Saskatchewan, entered these contests back in the 80's before budget cutbacks. The year I graduated from high school, 1986, they had a record breaking gasoline powered car that went 5,691 miles per gallon. I think they, or someone else, went over 7000 miles per gallon a few years later. Even in those days 2000-3000 miles per gallon was being achieved by many teams. University of British Columbia achieved 3,145 miles per gallon in 2006.
All the same, I'd like to see the test done with the consumption of an entire gallon of fuel, especially record-breakers. Over a 1000+ mile journey, a lot can happen to decrease fuel efficiency, such as engine contamination, varying atmospheric conditions, etc. Of course, these cars are so tiring to drive, with the driver lying down horizontally and peering around the toes, the fatigue factor will demolish any extrapolated record.
It's amazing how much quality improves when you have to deal with your own mistakes.
Who else would be dealing with them? A programmer or designer might be able to do only so much damage control, but worst case consequences for mistakes should always have a planned fix, particularly for implementations that did not have to be hacked together in a rush. There has to be backups and undos, if possible. Testing under extreme conditions is needed if failure is hard to accept. Criteria for ready-to-use need to be determined, rather than a fuzzy you-have-incentive-to-not-screw-up.
I wouldn't mind it in the slightest if it were limited to non-voice uses. What's to object to? But conversations would be justification for homicide.
Well, isn't it more a problem with the phone rather than with the talker?
If you were with someone, you two can talk quietly and unobtrusively, preferring to maintain your privacy, but something is wrong with the phone.
Either the phone isn't clear enough to hear when you don't speak up, or people don't care if their one-sided conversation is overheard. The correction is to either make phones louder or pick up sound better, or to have phone speakers hearable from a distance when the remote talker is loud.
Well, look at it this way. The serial bus on your computer can transmit either a 0 or a 1 followed by more of the same so it takes N transmissions to transmit N bits. However, if you can transmit any number between 0 and 4^16-1 then you can transmit 32 bits in one shot, and N bits in N/32 shots. Now, how many shots per second can be made depends on how fast the encoding can occur after a lasering. At the molecular size, it could be fast, but near absolute zero maybe it's slow. But to transport the molecules over a channel, it might not be like voltage either, but rather a like a freight train, who knows? If the processing units are close together, it could be hard to keep the temperature down, and if the processing units are farther apart, this might be an easier way to transport large chunks of data all at once.
Another application may be parallel computing. Suppose you can encode 32 bits at once with a molecular construct. Then if you make an row of 32 or 64 of these encoders, you can encode the partial computations of 32 threads that an x86 machine might do. A chain of rows can form a pipeline. The details are left as an exercise requiring venture capital.
Cell was a fairly radical design departure. If IBM continues to refine Cell, and as more experience is gained, the challenge will likely diminish.
For one thing, IBM will likely add double precision floating point support.
The reason why x86 never died the thousand deaths predicted by the RISC camp is that heat never much mattered.
The Cell does sound pretty good, but for now I'll stick to Intel. You see, if you were to tell me about your heterosexual experiences with the Cell, I'd buy into it in a New York minute. That's where Intel is winning hands down.
I think this comment shows a failure to recognize the value of basic utilities.
Sure, basic utilities are recognized when they don't work, but things like plumbing have been around for millenia. They ought to be quietly working in the background so that people don't get distracted about committees for upgrading their plumbing or phones.
IT, for the moment, is being relegated to the status of a mere utility, a quiet tool that should not detract from the main effort. Of course, IT has more potential than a utility, only none of this potential is that easy to realize with current technology.
Like a craftsman in shop full of tools, the user of IT is thankful for all the different tools, but is typically clueless about building a better tool. Also, the user is thankful that the tool doesn't render him/her obsolete, but rather provides an opportunity while not having to do the work even more manually.
When more opportunities arise than can be tackled with mere staffing, or competition causes a need to increase output or reduce costs, the tools become a bottleneck, and investment in technology is considered. IT spending is driven by IT distracting from the core business, and who wants to be distracted?
The difference between IT and a dumb tool is that computers are potentially smart while information alone is more noise than signal. Currently IT mindset is to serve up dumb information and provide well understood calculations on the information. The problem is that the calculations don't go far enough because there are still more thoughts to think after the information is output. No one expects the ultimate goals to be obtained by the time the computer is done.
Incomplete information is one of the big stumbling blocks of computers. They have enough power to calculate everything well before the cows get home, but they don't have the information inputs and the storage to calculate deeper reliable results.
Another problem is the lack of computer understanding for what is going on. A computer is told to follow a sequence of steps but does not handle new situations. A person has to reprogram or reconfigure, and by the time this work is done, the computer might not be applicable. Even having a huge storehouse of ready-made programs for numerous scenarios doesn't guarantee the right program would be found and operated in time, as software is hard to use in just the right way to get the right results.
Gradually, these problems will be reduced and computer results will be more interesting. Then, computers will be able to run larger and larger segments of a business. The role of computers in the economy increases quite rapidly, but IT changes per business are made in quantum leaps rather than continuously. Perhaps we can blame Windows for being practically hateful for business upgrades, but the real problem is the results of the Top 500 - the fastest supercomputers have yet to prove to the public how the world can get out of its current dilemmas. Skyrocketing oil prices, foreclosures, terrorism, global warming, disease, aging, etc. are problems that have persisted without any solution from computers, so people have to do the thinking. It's not unlikely that computers will be the ones that provide many of the answers because the problems require so much processing, geographically diverse information, and rapidly changing data. Such computers are a few quantum jumps from where we are, but people will make each jump when the technology shows a hint of promise, or when an opportunity or disaster large enough is at hand.
Slashdot Mirror
A flying car is great in cheesy novels and movies, but horrible in reality
For now, but once everyone knows how to travel between the planets, flying around a city should be nothing.
Wood.
Organic molecules involve many carbon atoms. One way of removing carbon is to grow forests, cut them down and turn the wood into buildings.
Rather than spending megabucks to dump carbon into a hole, why not ship wood to the third world, so long as they don't burn it?
Wood is a lot safer than carbon dioxide stored somewhere, which can suddenly escape.
One of causes of unaffordable housing is the lack of it. Who wants to work for 30 years just to pay off a house? Meanwhile we hear that the rich are getting richer and everyone else isn't. Simultaneously people are driving around emitting carbon dioxide, driving the environment to ruin, trying to make ends meet. Raising real-estate quality in the third world helps the poor and reduces the environmental impact of our efforts just to live. Furthermore, industries that have been hurting from subprime can be rejuvenated. It just gets better and better.
Just wondering if this will lead to an alternative energy source, like building a dwarf star under the hood. That would cost about as much as a tank of gas anyways.
There are a good many places in my city advertising free Wi-Fi, but that's not bringing any customers with computers. However, many people go to the library with their laptops. Lots of reasons for the discrepancy: the lack of electrical outlets at the businessplaces, maybe people feel more private at the library, and the library's Internet is really quick.
If the Wi-Fi isn't supposed to be free, but is available for a fee, then there must be a fairly good demand! What makes people pay for their Internet outside the home, perhaps aside from not having Internet at home? I guess there are some places where the population density is so high, there's a signal even in the home from a pay-to-use hotspot.
But I agree, most of it is repetitive. Every time I start a new project, I think, "Oh crap, I have to do this AGAIN!?!?"
Well, that's work. If there wasn't work to do, how do you earn money?
Actually, it's good that it's so repetitive because it means you don't have to invent something for every little cotton picking detail. So much structure and so many paradigms have been recognized and instilled that it becomes easier to convince people to pay for creating software that takes things to the next level.
Software for business needs to make a quantum leap: natural language processing, audio and visual input, and then virtual reality and artificial intelligence. Someone working as an IT grunt won't be inventing these funky toys. An IT developer just brings technology together.
If you want to work as an IT developer on programs that have more variety, perhaps find jobs programming small tools that don't have many end users. These programs tend to be cut to the bone and the user might not want enough features to provide full time employment, but there is a bit of urgency to have the computer do something exotic. Only thing is, these days many young people know how to write a teensy program to crank on small chunks of data, even if it's O(n^2) more inefficient than it could be and takes O(n^2) more debugging time.
HTML form input straight into a SQL query is so... fucking... bad
:) and I'm happy to say that all my program-generated SQL queries have always been constructed to treat text as text by escaping single quotes so I didn't become vulnerable to such antics.
Some time ago I wrote a debugging routine to output some info to a browser, and unwittingly output HTML code embedded in the info. Little experiences like that can teach one about vulnerabilities and bizarreness of computers following your every command. FTFA, "little Bobby Tables" is quite illuminating
The sad truth is he is 'weird' and should never have been permitted, by his attorney, to take the stand
It seems that his behavior was weird as in different from usual self after his wife was gone.
I can only wonder what pleading the fifth would have achieved.
I prefer deep nesting, a nasty code to read as it is when that's what it takes to do the job. As I see it, if you outsource :) the inner loops you may have declared a number of local variables many of which may be relevant at a feature change's notice to an inner loop. If you have to call a function of a function, which looking good as it may be if done right, the function becomes a function of many parameters.
A habit of mine is to start writing code into the same function if it is not immediately known whether there will be a separate function caller of portions of the function. However, coupling through local variables is kept to a minimum between obvious segments that look separable. After debugging everything, which is easier in a single function than skipping all over in a trace, separation into different functions is painless. The separation could have been done to start with, but the coupling through local variables and big parameterization give me pause.
BTW, I see unit tests as good things for lengthy runs or runs with lots of particular inputs. However, I've avoided unit tests lately to keep costs down. Instead I pretty much know what I do is going to work, run a couple of tests, and things are fine, except lately I've seen some odd bugs due to Microsoft Windows and Office--that's where the costs really occur, making unit tests seem more a requirement except how do you unit test visual output from Office without writing almost as much code as was written?
So the thread all goes:
A better way to handle that is to turn the loop body into a function or group of functions. makes the code easier to read and a good compiler will inline the function so their's no performance loss.
I somewhat disagree with what you and... *sigh* Monkeybaister posted. Yes, there are many times when long stretches of code should be broken out into functions. But I tend to do that mostly when the same bit of code is used in several different cases. The reason being is that when you start modularizing off all your while loops that are more than a dozen lines long, you create a whole new type of spaghetti code. I'm going to coin a term and call it "spaghetti-O code." You try to track down a bug and what would have been a straightforward couple pages of code now has all kinds of functions at different places in the code. As such, it can often make debugging or forming a mental map of the code much harder
Putting a comment above the line telling people why you're doing it isn't.
I don't have to tell people why I do it. I do it for money. Well, sometimes for love and world peace, but that's rare.
1/50 cents/Mb
When you consider the sheer amount of information required for a person to cope with reality, that's still a costly amount for computing. There are anecdotes of people under hypnosis being able to recall minute details of what they've seen even though these details would hardly rate any attention. The mind may well be storing high definition video of everything you see--how much storage would that require?
If an AI were to build an idea of its environment from video, audio, and tactile inputs, Seagate's first hard disks would have been fairly prohibitive, but hard disks for PCs might be on the thresholds of affordability.
Now at last I have an inkling of how werewolves come to be.
If you want real performance and aren't afraid of having to do a complete rebuild on a regular basis then the best bet is to purely use a huge amount of RAM, not Flash or other solid state disks but real genuine RAM.
Okay so its insanely expensive and a power cut and UPS failure means you lose everything.... but the SPEED is fantastic
Talking about speed, this is an effective design. Multiple UPS and a separate computer that maintains the RAM will give reliability. Not sure if it's worth it just to boot faster, and I admit being scared by the volatility. Hard disk speedups have not been so easy to come by. It took 7 minutes to copy a 3 Gb file and everything felt slow when that was happening.
Having plenty of RAM does speed things up because files are cached in RAM. I do not miss the days when I had a lot more swapping even on a 1 Gb RAM computer.
Does anyone have any ideas about multiple heads? If the heads are swinging independently, the mechanics are quite complex, but what if the heads all swing in unison - all together at the same direction and speed? Then the heads can spend less time per cylinder. Heads can also be given a positional offset in order to be on different cylinders at the same time. Complex mechanics, but in today's level of technology, par for the course as the saying goes. Even a slower RPM drive can still have the performance of a fast RPM drive. If the drive isn't too busy, some heads can be kept parked. RPM Speedstep may also contribute to power savings.
Exactly. You can get a 500 GB hard disk for $100. Why not just use those for backup. At the cost of this holographic storage, you could buy 2 or 3 500 GB hard drives, and keep multiple copies just in case one died. Since you'd only be using them for backup, and they would actually get very little wear and tear, I would guess that it would be easy to have a hard drive last for 50 years
But let's think of the possibilities. Backing up is the tip of the iceberg. Maybe you would be nuts to buy the product, but it still proves a concept, and that can be enough to think of taking the business to the next level.
Because lasers are being used, the size of the storage medium might be quite large. Too large to fit in a computer? That's ok. Think about the holodeck in Star Trek. Hi-res display device. Walking about a room full of lasers does give one pause though.
Multi-Touch is kind of unimaginative, when you consider the alternative: Fondle.
Now consider a computer that responds to touch all over. The intent of the user tends to be a bit vague however.
But it makes planning an exploit much harder. Before they might have been able to say they had 12 minutes (say) between sweeps, giving them that amount of time to get through a door, set a bomb, whatever. Now they might have an AVERAGE of 12 minutes, and possibly just 2 minutes Much more risky, and if they have to pass through more than one such point, they're almost sure to get caught
But now the left hand might not know what the right hand is doing so an impostor can show up at a door and say "I'm running a little late," which would happen a lot if there are short gaps between arrivals.
If the robot looked for a target without ever being commanded that'd be a pretty horrendous software error.
There are two sides on this coin. Heads: robots are more expendable than people, and intimidating, trigger-happy, seemingly out-of-control robots can scare enough bejesus out of militant insurgents to turn the tide and keep terrorists to themselves. Tails: a robot can be captured by the enemy and leads to the scenario, unlikely as it may be, that it is sent back with enough sneakiness to gun down commanding officers.
Now we leave you with a few words from Weird Al Yankovic: Trigger happy! Trigger happy every day!
At my alma mater, engineering students in the Society of Automotive Engineers club of University of Saskatchewan, entered these contests back in the 80's before budget cutbacks. The year I graduated from high school, 1986, they had a record breaking gasoline powered car that went 5,691 miles per gallon. I think they, or someone else, went over 7000 miles per gallon a few years later. Even in those days 2000-3000 miles per gallon was being achieved by many teams. University of British Columbia achieved 3,145 miles per gallon in 2006.
All the same, I'd like to see the test done with the consumption of an entire gallon of fuel, especially record-breakers. Over a 1000+ mile journey, a lot can happen to decrease fuel efficiency, such as engine contamination, varying atmospheric conditions, etc. Of course, these cars are so tiring to drive, with the driver lying down horizontally and peering around the toes, the fatigue factor will demolish any extrapolated record.
Just imagine when one can WALK, no, sa-shay, into your cube as YOU, for you...
So start a consulting business by training homeless people to do your profession.
As robots redefine work, who will be the homeless?
It's amazing how much quality improves when you have to deal with your own mistakes.
Who else would be dealing with them? A programmer or designer might be able to do only so much damage control, but worst case consequences for mistakes should always have a planned fix, particularly for implementations that did not have to be hacked together in a rush. There has to be backups and undos, if possible. Testing under extreme conditions is needed if failure is hard to accept. Criteria for ready-to-use need to be determined, rather than a fuzzy you-have-incentive-to-not-screw-up.
2.36 inches per instruction at the speed of light. He should have said "width of your hand"
Where's your knuckle? Mine is about 2.5" from my fingertip, and yes, I did clip my nails.
"Fortunately the problem is easy to fix: encrypt TPMS data the way keyless entry systems do."
Tracking me? Encrypt the license plate too!
I wouldn't mind it in the slightest if it were limited to non-voice uses. What's to object to? But conversations would be justification for homicide.
Well, isn't it more a problem with the phone rather than with the talker?
If you were with someone, you two can talk quietly and unobtrusively, preferring to maintain your privacy, but something is wrong with the phone.
Either the phone isn't clear enough to hear when you don't speak up, or people don't care if their one-sided conversation is overheard. The correction is to either make phones louder or pick up sound better, or to have phone speakers hearable from a distance when the remote talker is loud.
Doesn't sound like impressive computing
Well, look at it this way. The serial bus on your computer can transmit either a 0 or a 1 followed by more of the same so it takes N transmissions to transmit N bits. However, if you can transmit any number between 0 and 4^16-1 then you can transmit 32 bits in one shot, and N bits in N/32 shots. Now, how many shots per second can be made depends on how fast the encoding can occur after a lasering. At the molecular size, it could be fast, but near absolute zero maybe it's slow. But to transport the molecules over a channel, it might not be like voltage either, but rather a like a freight train, who knows? If the processing units are close together, it could be hard to keep the temperature down, and if the processing units are farther apart, this might be an easier way to transport large chunks of data all at once.
Another application may be parallel computing. Suppose you can encode 32 bits at once with a molecular construct. Then if you make an row of 32 or 64 of these encoders, you can encode the partial computations of 32 threads that an x86 machine might do. A chain of rows can form a pipeline. The details are left as an exercise requiring venture capital.
Cell was a fairly radical design departure. If IBM continues to refine Cell, and as more experience is gained, the challenge will likely diminish.
For one thing, IBM will likely add double precision floating point support.
The reason why x86 never died the thousand deaths predicted by the RISC camp is that heat never much mattered.
The Cell does sound pretty good, but for now I'll stick to Intel. You see, if you were to tell me about your heterosexual experiences with the Cell, I'd buy into it in a New York minute. That's where Intel is winning hands down.
I think this comment shows a failure to recognize the value of basic utilities.
Sure, basic utilities are recognized when they don't work, but things like plumbing have been around for millenia. They ought to be quietly working in the background so that people don't get distracted about committees for upgrading their plumbing or phones.
IT, for the moment, is being relegated to the status of a mere utility, a quiet tool that should not detract from the main effort. Of course, IT has more potential than a utility, only none of this potential is that easy to realize with current technology.
Like a craftsman in shop full of tools, the user of IT is thankful for all the different tools, but is typically clueless about building a better tool. Also, the user is thankful that the tool doesn't render him/her obsolete, but rather provides an opportunity while not having to do the work even more manually.
When more opportunities arise than can be tackled with mere staffing, or competition causes a need to increase output or reduce costs, the tools become a bottleneck, and investment in technology is considered. IT spending is driven by IT distracting from the core business, and who wants to be distracted?
The difference between IT and a dumb tool is that computers are potentially smart while information alone is more noise than signal. Currently IT mindset is to serve up dumb information and provide well understood calculations on the information. The problem is that the calculations don't go far enough because there are still more thoughts to think after the information is output. No one expects the ultimate goals to be obtained by the time the computer is done.
Incomplete information is one of the big stumbling blocks of computers. They have enough power to calculate everything well before the cows get home, but they don't have the information inputs and the storage to calculate deeper reliable results.
Another problem is the lack of computer understanding for what is going on. A computer is told to follow a sequence of steps but does not handle new situations. A person has to reprogram or reconfigure, and by the time this work is done, the computer might not be applicable. Even having a huge storehouse of ready-made programs for numerous scenarios doesn't guarantee the right program would be found and operated in time, as software is hard to use in just the right way to get the right results.
Gradually, these problems will be reduced and computer results will be more interesting. Then, computers will be able to run larger and larger segments of a business. The role of computers in the economy increases quite rapidly, but IT changes per business are made in quantum leaps rather than continuously. Perhaps we can blame Windows for being practically hateful for business upgrades, but the real problem is the results of the Top 500 - the fastest supercomputers have yet to prove to the public how the world can get out of its current dilemmas. Skyrocketing oil prices, foreclosures, terrorism, global warming, disease, aging, etc. are problems that have persisted without any solution from computers, so people have to do the thinking. It's not unlikely that computers will be the ones that provide many of the answers because the problems require so much processing, geographically diverse information, and rapidly changing data. Such computers are a few quantum jumps from where we are, but people will make each jump when the technology shows a hint of promise, or when an opportunity or disaster large enough is at hand.