The ~50 engineers at my workplace switched to NX 2 years ago. There have been many complaints about the instability of NX ever since. Sometimes engineers lose a handle on a simulation that has been running for more than 24 hours and have to start over. I stayed with VNC and lately people are asking me how to switch back.
For me, (Ultra)VNC has been rock-solid. If Google can make NX reliable, maybe I'll make the move, too.
I admit no knowledge of Blackberry phones or their apps. Your detailed account of development was super-interesting and I wish you the best of luck. Just one comment:
You've been trying to contact Audible but they're not returning your calls. Reminds me of this:
Two years ago one of my hard drives started making strange noises like a grinding motor. When trying to read some of the bits on the disk, the hard drive would fail for a minute or two before giving up.
I used Handy Recovery. It scans the drive and gives a file explorer similar to the one in WindowsXP. Recovery is easy. I've also been able to recover files from an old file system even after formating a disk and putting files on the new filesystem.
I realize that many tools exist. Are most able to recover files from disks with PHYSICAL malfunctions? That seems important.
Also agree. VHDL is the dinosaur. Verilog is the way to go. SystemVerilog is also a bonus. It's now a standard and while not as powerful as Vera or SpecMan, it's getting better and still much less expensive. Many companies are switching to SystemVerilog for cost. Recent graduates will probably write more SystemVerilog than Verilog in their first years at work.
When you reward a company with money per traffic violation, obviously it will be in their interest for there to be more traffic violations. And the traffic laws are there to protect lives. Basically, governments are rewarding companies for killing people.
It seems like it would be trivial to implement a system to make gift cards. Having thought about it for just five minutes:
On each gift card you print a serial number and then the HMAC of thatserial number using a super-secret key. (Maybe as simple as concatenating the super-secret key and the serial number then hashing with SHA-1 or whatever is strongest today.) The iTunes store takes serial number and hash result as input.
That's it. To break the algorithm you either have to discover the secret key, which means that you've broken SHA-1. Good luck breaking SHA-1.
The point is, if hacking Apple's algorithm involved anything less than breaking a cryptographic algorithm that is believed unbreakable then Apple screwed up.
This sounds familiar. Like when there are too many insects in some farmer's field so he brings in some frog from a far-away country to eat all the insects. But then the frog becomes a nuisance and has side effects of his own so they bring in some cats that eat this frog, but the cats...
I wouldn't want to be the first to try out this new bacterium.
I think that he and other futurists need to be judged on two scales: Accuracy and relevance. Accuracy is easy. Wait 10 years and see what has come true and what hasn't. For relevance, the idea is to measure the implication of the prediction. To somehow give a grade to how radical it is. That should be done as soon as the prediction is written down. Hell, I can predict the time of day and season 10-years from now but who cares?
So he was wrong about the voice-recognition but he was right about the notebooks. You might say that the notebooks seem obvious. *Now* they seem obvious. Did they seem obvious back then?
In all, I'm unimpressed. I definitely wouldn't pay him $25,000 for dinner.
"Your usage of the infrastructure" should include the air and water infrastructure that is polluted when you burn the gasoline and the army infrastructure that is busy battling around the world because of the money funneled into terrorist hands.
Gas isn't expensive but it should be. It's time that Americans paid the true cost of what they are receiving.
If what you want is to travel abroad, then do that. You want to fly to a foreign country and spend your time in front of a computer? The computers there look just like the computers here!
Instead: Finish your degree or skip a semester, pack a backpack, and fly to Argentina. Though unpopular among Americans, there are tons of Israelis (like me), British, Australians, and some other Europeans backpacking through South and Central America.
Travel in South America cost me about $700/month (~6 months), living reasonably, seeing all the sights, and taking buses from place to place. This is probably less than you're currently spending on rent. Starting your career a few months or a year late will have no long-term effect on your career. You could do it all on loans and have it paid back within months of work.
Don't go to Europe. Your money won't go nearly as far. And if you're looking for culture, leaving one westernized country of wealthy inhabitants for another westernized country of wealthy inhabitants seems bland. "Oh gee, here they say tea instead of dinner. How totally worldly I am now!" Bah!
A neat feature of the pssword manager is that you can use a master password. Without a master password, a trojan horse running on your system can steal all your passwords.
How come there is no master password to protect the cookies? Nowadays as most sites remember who I am in a cookie, a cookie seems just as useful as a password. Did no one else figure this out or did I get it wrong?
There's a time and place for each project. The Tetris game and most video games are not good FPGA projects. Tetris is something you could write just as well or better using a general purpose processor. No one would ever turn to an FPGA to write a video game nowadays. Implementing a CPU and searching DES space seem reasonable, though. When I took a similar course I remembering thinking, "What a waste," back then, too. We only had 10k gate-equivalents on our Xilinx FPGAs, though.
Would this work in an engine? Is someone already doing it? Maybe changing the number of electromagnets in use to spin a motor could be used in lieu of gears to achieve different levels of power at different speeds in a car?
I remember reading once about a solution to the TCP SYN flood problem. I think that one of the RSA guys (maybe Rivest) wrote it. Client puzzles. Would it work in lieu of CAPTCHA?
In addition to checking that the user is human, slow him down. Send his computer a cryptographic puzzle, something like, "What is the DES key that can decrypt _______ to get message text ______?" Now you've got his computer busy for a few seconds trying to break a password. If a spammer has to break a CAPTCHA but his computer can ony muster up the CPU to do it a few times a minute, will that slow him down enough?
You'd have to install some add-on into browsers to accept and try to solve client puzzles. The point is, maybe distinguishing humans from computers isn't enough.
As noted in the RFC, the ability to spoof addresses is there. From a practical standpoint, a vendor has little incentive to filter source addresses. Source-address filtering slows down the router for the people that are paying for it (indirectly or otherwise) and provides protection for the people that aren't paying for it.
Put another way, imagine a product that you could buy that would ensure that you don't rob anyone else's house but doesn't provide you any protection. How much money and inconvenience would you put up with for such a device?
No one source filters. End-to-end crypto is the real solution.
Who the hell wants to be in IT anyway? When I think of IT, I often think of the guy that played a lot of video games when he was younger, so he liked computers, but didn't quite get the push into programming/engineering. So he spent his teens reconfiguring the computer and the modem for a few hours and now he's made a career out of it as sysadmin. On the other hand, you've got the guys that spent their teens soldering BJTs or writing open source code into project that span months. Those guys go on to be engineers.
The engineers make more money and have more fun. It's a rare company where the guy making the product is getting paid less or doing less interesting work than the guy supporting him. The closer you are to being the guy that generates revenue, the more respect that you'll command. And respect in the industry lets you demand a more interesting job.
This is common at the beach. Before you head out surfing or sailing you want to know the size, direction and shape of the waves on the beach. No machine will measure how sloppy the break is, so the easiest solution is a webcam aimed at the water.
Once you've already got a webcam, putting a windsock on a pole in the water in view of the camera is often as good or better than an electronic anemometer. The electronics often corrode in the salty air, the anemometer often needs oiling, and you can't post it 20 yards offshore unless it's wireless and you want to be replacing batteries all the time. And then it's going to get stolen or broken. Or corrode first.
In my experience, the odds that the anemometer that you need at any given moment is running and accurate when you need it is about 50%. Sometimes the old ways are the best ways.
I didn't think about it that way, but you're right, it's true. If you don't care how hot or big your chip gets, give your user as many instructions as you can. Having a bunch of little instructions means that they all take as long as the slowest one, even if most of them don't need a full clock cycle.
The interesting part of the article is about the process. Intel's domination has been in their process, always a few steps ahead of the competition (maybe just a half step ahead of TSMC). Newer processes have always yielded faster, smaller, and cooler chips. Not anymore. 60nm didn't make chips use less power and 45nm doesn't help either.
In a sense, one dimension of the playing-field has become level for Intel and the custom fabs. And that's the level in which embedded plays.
If you act exactly as your enemy, do you become your enemy? Is the experiment to see if the tactic will stop the bombing, or to determine if you begin to hate yourself for the same reasons you hate the enemy?
IT is an over-hyped career about 5 years away from being a commodity. I haven't worked at a company that has in-house IT in a long time. No wonder the IT guys are the first ones laid-off. The question is where to learn EE or just CS. I studied EE/CS and highly recommend integrating hardware studies into your degree.
1) All the engineers that I know that studied CS without EE are lousy engineers. Just as understanding assembly makes you a better C programmer, so does understanding logic design make you a better assembly programmer.
2) Both CS and EE you can learn from a book, but only CS can you download and practice at home. EE requires expensive equipment and a lab that the university can provide.
3) Because CS is easier to learn on your own, everyone and his mother can get started. Competition will get fierce. Good electrical engineers are always more difficult to find. Low supply means better wages.
In short: If you had a "quantum hard drive", not only would you be unable to detect the error, if you even tried, you would ruin the data. And the article is still total FUD.
Let me try a simplified explanation using my limited quantum computing knowledge.
There are a bunch of parallel universes and every possible thing that could ever possibly happen is going on in one of them right now. (There's a universe in which you run into a wall and go right through it. In university-level physics you can calculate the odds of being in that universe.)
A regular bit is either 0 or 1 in our universe and almost all other universes. When the value of anything that you can observe is the same in almost all universes, that's called a "sharp" value. There are universes in which you wrote a 1 to your hard drive and got a 0, but the hard drive companies make sure that they're bits act according to classical physics.
A qubit acts sometimes like that bit when it's value is very "sharp", but sometimes is "unsharp" and carrying 0 in half the universes and 1 in half the universes. (Doesn't have to be a 50-50 split, mind you.) If we take a peek in our universe it's still either 0 or 1.
Quantum computing takes advantage of the qubit being unsharp and how unsharp it is. But there's no way to know if the qubit is sharp by looking at it, because in our universe it's always either 0 or 1. Worse still, looking at a qubit would make it more sharp in all the universes and ruin your computation.
How do you detect a virus that changes data in other universes? Still, total FUD.
Slashdot Mirror
The ~50 engineers at my workplace switched to NX 2 years ago. There have been many complaints about the instability of NX ever since. Sometimes engineers lose a handle on a simulation that has been running for more than 24 hours and have to start over. I stayed with VNC and lately people are asking me how to switch back.
For me, (Ultra)VNC has been rock-solid. If Google can make NX reliable, maybe I'll make the move, too.
Anyone else have complaints about NX stability?
This isn't hackable but it's awesome:
http://www.waze.com/faq/
Works in Israel so far and they say that it will roll out to the USA soon. Doesn't require special hardware and the software is free (as in beer).
I admit no knowledge of Blackberry phones or their apps. Your detailed account of development was super-interesting and I wish you the best of luck. Just one comment:
You've been trying to contact Audible but they're not returning your calls. Reminds me of this:
http://www.joelonsoftware.com/items/2009/06/10c.html
Maybe Audible doesn't want to help you because they want to implement what you've already done?
Two years ago one of my hard drives started making strange noises like a grinding motor. When trying to read some of the bits on the disk, the hard drive would fail for a minute or two before giving up.
I used Handy Recovery. It scans the drive and gives a file explorer similar to the one in WindowsXP. Recovery is easy. I've also been able to recover files from an old file system even after formating a disk and putting files on the new filesystem.
I realize that many tools exist. Are most able to recover files from disks with PHYSICAL malfunctions? That seems important.
Also agree. VHDL is the dinosaur. Verilog is the way to go. SystemVerilog is also a bonus. It's now a standard and while not as powerful as Vera or SpecMan, it's getting better and still much less expensive. Many companies are switching to SystemVerilog for cost. Recent graduates will probably write more SystemVerilog than Verilog in their first years at work.
In the USA, reselling a book is totally legal. I imagine that renting one is, too. Which part of the copyright law are they accused of breaking?
When you reward a company with money per traffic violation, obviously it will be in their interest for there to be more traffic violations. And the traffic laws are there to protect lives. Basically, governments are rewarding companies for killing people.
http://www.thenewspaper.com/news/06/602.asp
How about giving the companies a bonus relative to the decrease in the number of traffic accidents in an intersection? Now that seems smarter.
It seems like it would be trivial to implement a system to make gift cards. Having thought about it for just five minutes:
On each gift card you print a serial number and then the HMAC of thatserial number using a super-secret key. (Maybe as simple as concatenating the super-secret key and the serial number then hashing with SHA-1 or whatever is strongest today.) The iTunes store takes serial number and hash result as input.
That's it. To break the algorithm you either have to discover the secret key, which means that you've broken SHA-1. Good luck breaking SHA-1.
The point is, if hacking Apple's algorithm involved anything less than breaking a cryptographic algorithm that is believed unbreakable then Apple screwed up.
This sounds familiar. Like when there are too many insects in some farmer's field so he brings in some frog from a far-away country to eat all the insects. But then the frog becomes a nuisance and has side effects of his own so they bring in some cats that eat this frog, but the cats...
I wouldn't want to be the first to try out this new bacterium.
I think that he and other futurists need to be judged on two scales: Accuracy and relevance. Accuracy is easy. Wait 10 years and see what has come true and what hasn't. For relevance, the idea is to measure the implication of the prediction. To somehow give a grade to how radical it is. That should be done as soon as the prediction is written down. Hell, I can predict the time of day and season 10-years from now but who cares?
So he was wrong about the voice-recognition but he was right about the notebooks. You might say that the notebooks seem obvious. *Now* they seem obvious. Did they seem obvious back then?
In all, I'm unimpressed. I definitely wouldn't pay him $25,000 for dinner.
http://money.cnn.com/2007/05/01/magazines/fortune/kurzweil.fortune/index.htm
"Your usage of the infrastructure" should include the air and water infrastructure that is polluted when you burn the gasoline and the army infrastructure that is busy battling around the world because of the money funneled into terrorist hands.
Gas isn't expensive but it should be. It's time that Americans paid the true cost of what they are receiving.
If what you want is to travel abroad, then do that. You want to fly to a foreign country and spend your time in front of a computer? The computers there look just like the computers here!
Instead: Finish your degree or skip a semester, pack a backpack, and fly to Argentina. Though unpopular among Americans, there are tons of Israelis (like me), British, Australians, and some other Europeans backpacking through South and Central America.
Travel in South America cost me about $700/month (~6 months), living reasonably, seeing all the sights, and taking buses from place to place. This is probably less than you're currently spending on rent. Starting your career a few months or a year late will have no long-term effect on your career. You could do it all on loans and have it paid back within months of work.
Don't go to Europe. Your money won't go nearly as far. And if you're looking for culture, leaving one westernized country of wealthy inhabitants for another westernized country of wealthy inhabitants seems bland. "Oh gee, here they say tea instead of dinner. How totally worldly I am now!" Bah!
A neat feature of the pssword manager is that you can use a master password. Without a master password, a trojan horse running on your system can steal all your passwords.
How come there is no master password to protect the cookies? Nowadays as most sites remember who I am in a cookie, a cookie seems just as useful as a password. Did no one else figure this out or did I get it wrong?
There's a time and place for each project. The Tetris game and most video games are not good FPGA projects. Tetris is something you could write just as well or better using a general purpose processor. No one would ever turn to an FPGA to write a video game nowadays. Implementing a CPU and searching DES space seem reasonable, though. When I took a similar course I remembering thinking, "What a waste," back then, too. We only had 10k gate-equivalents on our Xilinx FPGAs, though.
From the article: "Was the massive patching effort justified, or was Kaminsky just an arrogant, media-hungry braggart?"
Yes and yes.
Would this work in an engine? Is someone already doing it? Maybe changing the number of electromagnets in use to spin a motor could be used in lieu of gears to achieve different levels of power at different speeds in a car?
I remember reading once about a solution to the TCP SYN flood problem. I think that one of the RSA guys (maybe Rivest) wrote it. Client puzzles. Would it work in lieu of CAPTCHA?
In addition to checking that the user is human, slow him down. Send his computer a cryptographic puzzle, something like, "What is the DES key that can decrypt _______ to get message text ______?" Now you've got his computer busy for a few seconds trying to break a password. If a spammer has to break a CAPTCHA but his computer can ony muster up the CPU to do it a few times a minute, will that slow him down enough?
You'd have to install some add-on into browsers to accept and try to solve client puzzles. The point is, maybe distinguishing humans from computers isn't enough.
As noted in the RFC, the ability to spoof addresses is there. From a practical standpoint, a vendor has little incentive to filter source addresses. Source-address filtering slows down the router for the people that are paying for it (indirectly or otherwise) and provides protection for the people that aren't paying for it.
Put another way, imagine a product that you could buy that would ensure that you don't rob anyone else's house but doesn't provide you any protection. How much money and inconvenience would you put up with for such a device?
No one source filters. End-to-end crypto is the real solution.
Who the hell wants to be in IT anyway? When I think of IT, I often think of the guy that played a lot of video games when he was younger, so he liked computers, but didn't quite get the push into programming/engineering. So he spent his teens reconfiguring the computer and the modem for a few hours and now he's made a career out of it as sysadmin. On the other hand, you've got the guys that spent their teens soldering BJTs or writing open source code into project that span months. Those guys go on to be engineers.
The engineers make more money and have more fun. It's a rare company where the guy making the product is getting paid less or doing less interesting work than the guy supporting him. The closer you are to being the guy that generates revenue, the more respect that you'll command. And respect in the industry lets you demand a more interesting job.
This is common at the beach. Before you head out surfing or sailing you want to know the size, direction and shape of the waves on the beach. No machine will measure how sloppy the break is, so the easiest solution is a webcam aimed at the water.
Once you've already got a webcam, putting a windsock on a pole in the water in view of the camera is often as good or better than an electronic anemometer. The electronics often corrode in the salty air, the anemometer often needs oiling, and you can't post it 20 yards offshore unless it's wireless and you want to be replacing batteries all the time. And then it's going to get stolen or broken. Or corrode first.
In my experience, the odds that the anemometer that you need at any given moment is running and accurate when you need it is about 50%. Sometimes the old ways are the best ways.
I didn't think about it that way, but you're right, it's true. If you don't care how hot or big your chip gets, give your user as many instructions as you can. Having a bunch of little instructions means that they all take as long as the slowest one, even if most of them don't need a full clock cycle.
The interesting part of the article is about the process. Intel's domination has been in their process, always a few steps ahead of the competition (maybe just a half step ahead of TSMC). Newer processes have always yielded faster, smaller, and cooler chips. Not anymore. 60nm didn't make chips use less power and 45nm doesn't help either.
In a sense, one dimension of the playing-field has become level for Intel and the custom fabs. And that's the level in which embedded plays.
...iet Russia, everyone is welcome. Unlike the Romanians, who would not welcome her because of her der
If you act exactly as your enemy, do you become your enemy? Is the experiment to see if the tactic will stop the bombing, or to determine if you begin to hate yourself for the same reasons you hate the enemy?
(writing from Israel)
IT is an over-hyped career about 5 years away from being a commodity. I haven't worked at a company that has in-house IT in a long time. No wonder the IT guys are the first ones laid-off. The question is where to learn EE or just CS. I studied EE/CS and highly recommend integrating hardware studies into your degree.
1) All the engineers that I know that studied CS without EE are lousy engineers. Just as understanding assembly makes you a better C programmer, so does understanding logic design make you a better assembly programmer.
2) Both CS and EE you can learn from a book, but only CS can you download and practice at home. EE requires expensive equipment and a lab that the university can provide.
3) Because CS is easier to learn on your own, everyone and his mother can get started. Competition will get fierce. Good electrical engineers are always more difficult to find. Low supply means better wages.
In short: If you had a "quantum hard drive", not only would you be unable to detect the error, if you even tried, you would ruin the data. And the article is still total FUD.
Let me try a simplified explanation using my limited quantum computing knowledge.
There are a bunch of parallel universes and every possible thing that could ever possibly happen is going on in one of them right now. (There's a universe in which you run into a wall and go right through it. In university-level physics you can calculate the odds of being in that universe.)
A regular bit is either 0 or 1 in our universe and almost all other universes. When the value of anything that you can observe is the same in almost all universes, that's called a "sharp" value. There are universes in which you wrote a 1 to your hard drive and got a 0, but the hard drive companies make sure that they're bits act according to classical physics.
A qubit acts sometimes like that bit when it's value is very "sharp", but sometimes is "unsharp" and carrying 0 in half the universes and 1 in half the universes. (Doesn't have to be a 50-50 split, mind you.) If we take a peek in our universe it's still either 0 or 1.
Quantum computing takes advantage of the qubit being unsharp and how unsharp it is. But there's no way to know if the qubit is sharp by looking at it, because in our universe it's always either 0 or 1. Worse still, looking at a qubit would make it more sharp in all the universes and ruin your computation.
How do you detect a virus that changes data in other universes? Still, total FUD.