Shouldn't a software engineer, aside from write code, be able to accurately assess and estimate the time necessary to implement new features?
And the answer by now should be obvious to anyone - no. Look at all the games that ship late, and with bugs. Look at the OSes that run years over. Look at the apps which drop half the features from first ads. Software engineers can make rough guesses, but until the coding is well underway, there is no way to make an accurate assessment on features.
It's not the engineer's fault either. They are pressured by management and marketing to make a guess about whether something can be done and in what timeframe, when there is simply NO WAY to know. Management and marketing don't want to hear "I don't know, and neither will anyone else." They want a yes answer and a (short) time. Period.
The poor engineer does his best to guess on something a fortune teller couldn't handle, then works his -ss off to make it come about. Much of the time, they actually manage to pull it off. When they don't, I think the blame rightly goes where it should - management and marketing.
FF3 did well AT THAT TIME. Try to release it today and it'll bomb. No one other than a few old-timers will give a game on the level of FF3 more than a glance before dismissing it.
They've been making tiny switches for decades. I've got an issue of MIT Technology Review from 1990 that covers nanometer scale switches. The one I liked the best was the planar resonant tunneling field effect transistor (PRESFET).
The PRESFET is fabricated like normal FETs on ICs, so it's very compatible with the modern fabrication technology, but scaled down. They've just about reached the sizes needed to make PRESFETs (30 to 60 nm total size, 5 nm feature size). PRESFETs are also very low power, using less than 10 nanoamps with voltages of less than 0.2 volts.
One word - electromigration. The bane of integrated circuit designers everywhere. An electric field can cause atoms to be moved through a solid. The smaller the feature, the higher the electric field across the feature.
That is one BIG reason why the voltage is dropped every time chip fabs move to a smaller feature size. If they didn't, electromigration would kill the chips toot-sweet.
I imagine electromigration will be a HUGE factor in this technology.
I think you're interpreting that incorrectly. The author was pointing out that many classes could be taught faster. They contain a lot of wasted time. Eliminate the wasted time, and you've got a faster class learning the same amount of material.
For example, most math classes I took in high-school were taught at the speed of the slowest student in the class. You know - where the teacher writes "2+2" on one line, then "4" on the next and invariably ONE idiot raises his hand and asks "How'd you get that?"
Eliminate that student from the class and you could actually cover the same material in half the time or better! But that is an apt analogy for CPUs... you have to get rid of that idiot before you can raise the clockrate.
Correct on both counts. ALL benchmarks run by EVERYONE so far show AMD dual-core kicking the sh-t out of Intel dual-core. Clearly, integrated cores on the same die is superior to hacked together separate cores.
You can order and get AMD dual-cores TODAY. You can pre-order Intel's stuff, but good luck on a ship date any time in the next few months. Intel was the desparate party, looking for any way to divert attention from AMDs triumph. Idiots think Intel won this round, but we all know better.
Hmm, the first thing I do when I take custody of a place is kill all the bugs, vermin, and weeds. Makes the place much better to live in. Seems to me God didn't want us taking care of the 'poor sabertooth tiger'.
Will somebody PLEASE think of the sabertooth tigers!
Sorry bud, I ain't caring for no tiger... I'm calling the exterminator. Makes it easier for the kids to play in the yard.
So you're saying that a handful of hunters in North America hunted mammoths and other megafauna to extinction while the masses of humans in Africa and Asia didn't hunt elephants and other megafauna at all? Yeah, that makes REAL sense.
Because in other parts of the world, the megafauna survived. For intance, in Africa: Elephants, Lions, Giraffes, Rhinos, Hippos, and Gazelles; in Europe: Cows, Deer, and Reindeer; in Asia: Pigs, Sheep, Yaks, and Water Buffalo.
Yeah! Because there were no humans in Africa, Europe , and Asia to hunt them, they survived while the others where hunted to extinction. Uh... wait a minute...
Slashdot Mirror
Shouldn't a software engineer, aside from write code, be able to accurately assess and estimate the time necessary to implement new features?
And the answer by now should be obvious to anyone - no. Look at all the games that ship late, and with bugs. Look at the OSes that run years over. Look at the apps which drop half the features from first ads. Software engineers can make rough guesses, but until the coding is well underway, there is no way to make an accurate assessment on features.
It's not the engineer's fault either. They are pressured by management and marketing to make a guess about whether something can be done and in what timeframe, when there is simply NO WAY to know. Management and marketing don't want to hear "I don't know, and neither will anyone else." They want a yes answer and a (short) time. Period.
The poor engineer does his best to guess on something a fortune teller couldn't handle, then works his -ss off to make it come about. Much of the time, they actually manage to pull it off. When they don't, I think the blame rightly goes where it should - management and marketing.
Good at killing mammoths, but stunk at killing elephants... yeah, that makes a whole lotta sense.
:D
Must be all the hair... yeah... that's it. The hair caused the mammoths to get stuck in shrubbery while the hairless elephants escaped.
FF3 did well AT THAT TIME. Try to release it today and it'll bomb. No one other than a few old-timers will give a game on the level of FF3 more than a glance before dismissing it.
They've been making tiny switches for decades. I've got an issue of MIT Technology Review from 1990 that covers nanometer scale switches. The one I liked the best was the planar resonant tunneling field effect transistor (PRESFET).
The PRESFET is fabricated like normal FETs on ICs, so it's very compatible with the modern fabrication technology, but scaled down. They've just about reached the sizes needed to make PRESFETs (30 to 60 nm total size, 5 nm feature size). PRESFETs are also very low power, using less than 10 nanoamps with voltages of less than 0.2 volts.
One word - electromigration. The bane of integrated circuit designers everywhere. An electric field can cause atoms to be moved through a solid. The smaller the feature, the higher the electric field across the feature.
That is one BIG reason why the voltage is dropped every time chip fabs move to a smaller feature size. If they didn't, electromigration would kill the chips toot-sweet.
I imagine electromigration will be a HUGE factor in this technology.
I think you're interpreting that incorrectly. The author was pointing out that many classes could be taught faster. They contain a lot of wasted time. Eliminate the wasted time, and you've got a faster class learning the same amount of material.
For example, most math classes I took in high-school were taught at the speed of the slowest student in the class. You know - where the teacher writes "2+2" on one line, then "4" on the next and invariably ONE idiot raises his hand and asks "How'd you get that?"
Eliminate that student from the class and you could actually cover the same material in half the time or better! But that is an apt analogy for CPUs... you have to get rid of that idiot before you can raise the clockrate.
Correct on both counts. ALL benchmarks run by EVERYONE so far show AMD dual-core kicking the sh-t out of Intel dual-core. Clearly, integrated cores on the same die is superior to hacked together separate cores. You can order and get AMD dual-cores TODAY. You can pre-order Intel's stuff, but good luck on a ship date any time in the next few months. Intel was the desparate party, looking for any way to divert attention from AMDs triumph. Idiots think Intel won this round, but we all know better.
Hmm, the first thing I do when I take custody of a place is kill all the bugs, vermin, and weeds. Makes the place much better to live in. Seems to me God didn't want us taking care of the 'poor sabertooth tiger'. Will somebody PLEASE think of the sabertooth tigers! Sorry bud, I ain't caring for no tiger... I'm calling the exterminator. Makes it easier for the kids to play in the yard.
So you're saying that a handful of hunters in North America hunted mammoths and other megafauna to extinction while the masses of humans in Africa and Asia didn't hunt elephants and other megafauna at all? Yeah, that makes REAL sense.
Because in other parts of the world, the megafauna survived. For intance, in Africa: Elephants, Lions, Giraffes, Rhinos, Hippos, and Gazelles; in Europe: Cows, Deer, and Reindeer; in Asia: Pigs, Sheep, Yaks, and Water Buffalo.
Yeah! Because there were no humans in Africa, Europe , and Asia to hunt them, they survived while the others where hunted to extinction. Uh... wait a minute...