On the whole, this isn't so bad. A few years back a Chineese newspaper reported one of their stories as fact. IIRC it was about rebuilding the capital buildings.
On the contrary, your analogy is more harmful than helpful, due to the assumptions you make, and the conclusions you invite.
For example, you assume that work() is not already well-optimised, or even a no-op. You also assume that the procedure call overhead is negligible. In many modern languages, that overhead is far from insignificant.
It would make far more sense to optimise foo(), find an algorithym that doesn't involve 1*10^27 procedure calls to work()!
You say that the space portion of the missions are the safest, therefore they should fix the intra-atmospheric portions. Your assertions fail to profile the missions in their entirety. The space flight is a continous event; one of the reasons take-off is so hazardous is all the extra fuel, tools and materials needed to make the space-flight portions so safe!
Using your logic, the best way to increase the safty of space-flight missions is to reduce the EM shielding on the spacecraft to reduce the weight of the shuttle, and the forces that it suffers during take-off. Congratulations, you have just given the crew cancer.
I would take this discussion further, but my point is close to being made. Let's hope you can work the rest out for yourself.
Looking through all the comments posted so far, all I see is uninformed people who didn't read any of the fucking articles, have no idea what a hacker is, or what a website is.
Here is an analogy which is significanty less flawed than the others people have been throwing about:
At each of these universities, there is a big room marked 'Public Access'. It is filled with hundreds of cupboards, with memos, letters, advertising information, application forms and all kinds of documents. People were invited to come, open the cupboards and read the documents they found inside. Some of the documents also told people which cupboards they could look in for more information.
Now somebody pointed out that if you open up certain ones of less well-marked cupboards, you could see if they had processed the forms you had submitted yet. And so, people went and opened those cupboards.
The universities then said, "oopsie, we didn't mean for you to see that!" and said that these people are EVIL for seeing things that they shouldn't.
Not to mention that the stardate is totally out of bounds, but he has had that.sig for a while. He keeps it as a kind of permanent troll. Maybe he should be modded as such?
Except, of course, to run a file called README.EXE under a MS OS, the user would just type 'readme'. The new unix user would sit for 10 minutes wondering why that wasn't working, and then somebody would walk past and say "in unix, you need to type out the full name."
The user would then sit for half an hour, trying to work out why typing 'readme.exe' doesn't start 'README.EXE'.
Once he is told that the new shell is case sensitive, three days will pass before he finds out that files in the current directory need to be prefixed with "./" to me executed.
So, in summary, 'readme' != './README.EXE'
(Note: yeah, the./ would only apply in his home directory. But that is where his shell starts.)
This is an idea that has been bandied about for quite a while. The first time I recall hearing it was from a friend in the second half of the 1990s, between the time of the demise of Cyrix and the rise of AMD, where Intel essentially had no competition.
At that time, the idea was possible, if implausible. Recently I did reconsider this conjecture, with regards to the current market. At this time I can dismiss it out of hand.
There is indeed a lot of money to be made, and that that offsets the high cost of entry. Thus, if better processors are possible with modern technology, and the start-up costs are finite, competition will enter the market. If, on the other hand the start-up cost are insurmountable, then an oligopoly is possible. (if, for example, they are greater than can be raised with a safe investor base, allow either of the market leaders to buy the start-up out before their product hits the market.)
[note: as you can see, I am not the most eloquent writer around, but please bear with me.]
Untimately, if Moore's Law is dying, as the article states, then now would be the time they would be releasing their 'buffered' technology. It is pretty obvious that this isn't the case. The near future chip improvements all seem to revolve around multiple-core processors, an innovation that look like it will have little effect on overall performance.
Thus, I dismiss your idea of 'buffered' technology.
There is a far more likely scenario, that allows for the 'throttling' of new chip release performance and the existance of effective competition. These companies have huge resources, but they are still finite resources. They keep multiple lines of research open, and when they discover their competitors's advances in a certain area, they funnel their development budget into that technology, to keep up with the competition. The throttling is a result of research costs: They spend as much money as they need to keep up with their competitors, and perhaps to keep in check with Moore's Law.
Perhaps the recent 64-bit desktop chips were an example of this. AMD had a particularly successful (cost/benefit) line of research with this technology, and Intel had to spend extra money (increase costs) to catch up.
In an industry with ample competition, this idea is preposterous. If you want your company to SEE the next ten years, you must have your best product out.
I don't know why everyone is bitching. With all the piracy going on today, extreme measures like this are necessary to protect today's top games.
After all, the "don't authenticate our customer's copies of the game" form of copy protection seems to be even more successful than their controversial "don't release the game" copy protection.
Slashdot Mirror
Wrong, rather than picky. Electrical != digital. The electrical signals tranmitted in the nervous system are analogue.
".. the machine would basically, putting it in Windows terms, core dump or blue screen at random."
Heard inside Crest Electronics:
"I told you it was a bad idea to install that BSOD screensaver! Now what will we tell the PHB?"
I agree. 'Kills per death' will.
Punitive damages? Idiot.
On the whole, this isn't so bad. A few years back a Chineese newspaper reported one of their stories as fact. IIRC it was about rebuilding the capital buildings.
On the contrary, your analogy is more harmful than helpful, due to the assumptions you make, and the conclusions you invite.
For example, you assume that work() is not already well-optimised, or even a no-op. You also assume that the procedure call overhead is negligible. In many modern languages, that overhead is far from insignificant.
It would make far more sense to optimise foo(), find an algorithym that doesn't involve 1*10^27 procedure calls to work()!
You say that the space portion of the missions are the safest, therefore they should fix the intra-atmospheric portions. Your assertions fail to profile the missions in their entirety. The space flight is a continous event; one of the reasons take-off is so hazardous is all the extra fuel, tools and materials needed to make the space-flight portions so safe!
Using your logic, the best way to increase the safty of space-flight missions is to reduce the EM shielding on the spacecraft to reduce the weight of the shuttle, and the forces that it suffers during take-off. Congratulations, you have just given the crew cancer.
I would take this discussion further, but my point is close to being made. Let's hope you can work the rest out for yourself.
Well, then it's the cardiac arrest and resultant lack of blood supply to the brain that kills you, not the fall.
Imagine a beowulf cluster of these... An orgy?
Looking through all the comments posted so far, all I see is uninformed people who didn't read any of the fucking articles, have no idea what a hacker is, or what a website is.
Here is an analogy which is significanty less flawed than the others people have been throwing about:
At each of these universities, there is a big room marked 'Public Access'. It is filled with hundreds of cupboards, with memos, letters, advertising information, application forms and all kinds of documents. People were invited to come, open the cupboards and read the documents they found inside. Some of the documents also told people which cupboards they could look in for more information.
Now somebody pointed out that if you open up certain ones of less well-marked cupboards, you could see if they had processed the forms you had submitted yet. And so, people went and opened those cupboards.
The universities then said, "oopsie, we didn't mean for you to see that!" and said that these people are EVIL for seeing things that they shouldn't.
And that is the long an short of it.
From the review:
If you buy one "Revenge of the Sith" tie-in, this is one that won't disappoint.
If you were not a fan, how many product tie-ins would you buy?
Not to mention that the stardate is totally out of bounds, but he has had that .sig for a while. He keeps it as a kind of permanent troll. Maybe he should be modded as such?
Except, of course, to run a file called README.EXE under a MS OS, the user would just type 'readme'. The new unix user would sit for 10 minutes wondering why that wasn't working, and then somebody would walk past and say "in unix, you need to type out the full name."
./ would only apply in his home directory. But that is where his shell starts.)
The user would then sit for half an hour, trying to work out why typing 'readme.exe' doesn't start 'README.EXE'.
Once he is told that the new shell is case sensitive, three days will pass before he finds out that files in the current directory need to be prefixed with "./" to me executed.
So, in summary, 'readme' != './README.EXE'
(Note: yeah, the
Yeah, you forgot that the d10 isn't a Platonic solid.
What the hell is Slashdot coming to? A Monty Python joke has been up for almost half an hour and nobody has moderated it!
Thus, in summary, buffered technology is very unlikely, but throttled development spending could produce the same result.
This is an idea that has been bandied about for quite a while. The first time I recall hearing it was from a friend in the second half of the 1990s, between the time of the demise of Cyrix and the rise of AMD, where Intel essentially had no competition.
At that time, the idea was possible, if implausible. Recently I did reconsider this conjecture, with regards to the current market. At this time I can dismiss it out of hand.
There is indeed a lot of money to be made, and that that offsets the high cost of entry. Thus, if better processors are possible with modern technology, and the start-up costs are finite, competition will enter the market. If, on the other hand the start-up cost are insurmountable, then an oligopoly is possible. (if, for example, they are greater than can be raised with a safe investor base, allow either of the market leaders to buy the start-up out before their product hits the market.)
[note: as you can see, I am not the most eloquent writer around, but please bear with me.]
Untimately, if Moore's Law is dying, as the article states, then now would be the time they would be releasing their 'buffered' technology. It is pretty obvious that this isn't the case. The near future chip improvements all seem to revolve around multiple-core processors, an innovation that look like it will have little effect on overall performance.
Thus, I dismiss your idea of 'buffered' technology.
There is a far more likely scenario, that allows for the 'throttling' of new chip release performance and the existance of effective competition. These companies have huge resources, but they are still finite resources. They keep multiple lines of research open, and when they discover their competitors's advances in a certain area, they funnel their development budget into that technology, to keep up with the competition. The throttling is a result of research costs: They spend as much money as they need to keep up with their competitors, and perhaps to keep in check with Moore's Law.
Perhaps the recent 64-bit desktop chips were an example of this. AMD had a particularly successful (cost/benefit) line of research with this technology, and Intel had to spend extra money (increase costs) to catch up.
In an industry with ample competition, this idea is preposterous. If you want your company to SEE the next ten years, you must have your best product out.
People should realise that (shock and horror!) if they put something in their sig, it gets attached to every comment they post.
Once again for the hard-of-thinking: You are posting it with every comment you make.
It is part of every comment you make.
***And what do people do to other people's comments? They REPLY.***
If Microsoft release shoddy versions of these things in Windows XP, won't it slow the adoption of these technologies?
I might be wrong. I still use FAT32 in Window for the performance advantage.
Plus: Windows Mobile Pocket PC, Windows Mobile Pocket PC Smartphone Edition, Windows Mobile Smartphone, and Windows Mobile Media Center.
And don't forget XP comes in SP1, SP2 and vanilla flavours.
Post as AC, and you get what you deserve.
Yeah, I agree. I mod this article -1, flamebait.
You are not familiar with the constant 'st'
st = 1
1st = 1
So in fact, you are an idiot.
How is a "text MMO game" different to a MUD, MUSH, MOO, or whatever M** you prefer?
I don't know why everyone is bitching. With all the piracy going on today, extreme measures like this are necessary to protect today's top games.
After all, the "don't authenticate our customer's copies of the game" form of copy protection seems to be even more successful than their controversial "don't release the game" copy protection.