Bertrand Meyer has argued that the increasing performance of hardware makes algorithm design even more important. If you choose an algorithm with exponential complexity, then its performance will increase only linearly with time as hardware follows Moore's law, while the performance of a linear algorithm will improve exponentially.
It's best to "think big and code small". That is, make your code's internal interfaces flexible enough to accommodate any implementations you can brainstorm; then, with the interfaces in place, choose the most straightforward implementations that are most likely to be correct. Once it's working, if the program is too slow, then your brilliant interface design should allow you to re-implement the slow parts with ease.
After the 1.0 release, you can continue to tune for performance, and even re-architect some of the interfaces if your designs were too shortsighed to allow for the improvements you want to make.
Sounds of similar frequencies create beat frequencies. It requires no nonlinearity; only alternating constructive and destructive interference of waves.
Often used functions that are easy to find may take several mouse clicks to use when a keyboard command, while not intuitive, would make it much easier to do the same thing.
You're talking about synchronous circuits. With asynchronous circuits, there's no reason the whole chip needs to be reachable every clock cycle because there's not necessarily any clock at all.
Imagine have a porous vertical silicon cylinder 30cm in diameter and 50cm high running clockless reconfigurable logic (essentially implementing your algorithms in hardware) infused with alcohol that boils to cool it. Imagine tiny speed-of-light-reachable islands of optical logic gates connected to each other through a massive wireless laser crossbar.
Perhaps, but I think they have a long way to go before they hit that limit. For instance, remember that integrated circuits are still essentially two-dimensional. Also, the die area could continue to grow, increasing the number of transistors without actually shrinking the transistors themselves.
I think it's entirely possible that it's a continuum. Still, to claim that consciousness "emerges" from non-consciousness and leave it at that is still pseudoscience.
Moore's Law is not a physical theory, it is the observation of a common phenomenon, namely the curve that technology goes through as it becomes cheaper and eventually free.
First of all, anyone that drags out this old, tired "Moore's law is NOT A LAW!!!111!!" argument is an instant Foe in my books.
Second, what you state is not Moore's law. Moore's Law states that the number of transistors that fit on a chip doubles every N months, where N is usually taken to be 18, but varies from 12 to 24. It has nothing to do with cost, and certainly has no "eventually free" clause.
Third, in this particular context, your point is not even relevant. The article only mentions Moore's law/curve/observation/whatever in passing, stating that the trend cannot continue for more than 600 years. It makes no claim that Moore's law is a "physical theory".
I don't buy that argument. "Emergent byproduct" in this context is entirely equivalent to "magic". To say that non-conscious computational processes become conscious simply because there are "enough" of them to cross some threshold, without explaining or understanding how this occurs, is a vacuous, undisprovable argument, and hence it is pseudoscientific at best.
That's exactly the kind of crappy code that makes web sites insecure. If $var includes the right apostrophes, then an attacker can execute pretty much arbitrary SQL code.
If not, take a long exposure photograph with the lens cap on. This "dark" frame will be noisy because of hot CCD cells. In Photoshop/whatever, subtract this image from your photograph to subtract the noise.
Uh, how does that work? If it's really noise, then it will be different each time. Thus, subtracting the image is as likely to make it worse as better.
Re:Why are you only using even-numbered releases?
on
GCC 3.4.0 Released
·
· Score: 1
He never said he was only using those releases. Just that those releases broke binary compatibility.
Slashdot Mirror
Yes sir, that's the lamest comment I have read in a while.
Is there actually something new here?
Bertrand Meyer has argued that the increasing performance of hardware makes algorithm design even more important. If you choose an algorithm with exponential complexity, then its performance will increase only linearly with time as hardware follows Moore's law, while the performance of a linear algorithm will improve exponentially.
After the 1.0 release, you can continue to tune for performance, and even re-architect some of the interfaces if your designs were too shortsighed to allow for the improvements you want to make.
We already have about 10 of them.
What a stupid question. What does Linux have to do with America?
The word is "ridiculous". Root word: "ridicule".
Except clones don't share fingerprints.
Sounds of similar frequencies create beat frequencies. It requires no nonlinearity; only alternating constructive and destructive interference of waves.
Wow, it doesn't take much to earn your trust.
Imagine have a porous vertical silicon cylinder 30cm in diameter and 50cm high running clockless reconfigurable logic (essentially implementing your algorithms in hardware) infused with alcohol that boils to cool it. Imagine tiny speed-of-light-reachable islands of optical logic gates connected to each other through a massive wireless laser crossbar.
Have some imagination!! :-)
Perhaps, but I think they have a long way to go before they hit that limit. For instance, remember that integrated circuits are still essentially two-dimensional. Also, the die area could continue to grow, increasing the number of transistors without actually shrinking the transistors themselves.
I think it's entirely possible that it's a continuum. Still, to claim that consciousness "emerges" from non-consciousness and leave it at that is still pseudoscience.
Second, what you state is not Moore's law. Moore's Law states that the number of transistors that fit on a chip doubles every N months, where N is usually taken to be 18, but varies from 12 to 24. It has nothing to do with cost, and certainly has no "eventually free" clause.
Third, in this particular context, your point is not even relevant. The article only mentions Moore's law/curve/observation/whatever in passing, stating that the trend cannot continue for more than 600 years. It makes no claim that Moore's law is a "physical theory".
I don't buy that argument. "Emergent byproduct" in this context is entirely equivalent to "magic". To say that non-conscious computational processes become conscious simply because there are "enough" of them to cross some threshold, without explaining or understanding how this occurs, is a vacuous, undisprovable argument, and hence it is pseudoscientific at best.
Oh man I think I got trolled.
That's exactly the kind of crappy code that makes web sites insecure. If $var includes the right apostrophes, then an attacker can execute pretty much arbitrary SQL code.
Yikes. I have no idea what your point was, but I have the strong impression that you could have said it in a paragraph.
Dang, man, if you're going to nitpick, do it right.
He never said he was only using those releases. Just that those releases broke binary compatibility.
In the 1950s I had a customer who wanted the Internet on punch cards.
Try this version.