With type safe languages, they really should be compiling all the components into the same memory space and then they would be able to use the programming language to create the API between components instead of IPC.
Which would basically be Linux written in a type safe language. I can't find it, but IIRC, Linus commented once that Linux is like a microkernel with how the components of the system are separated, but they're using C function calls to initiate inter-communication and run in the same memory/process space. The same memory/process space is something microkernels use to get better performance, but that destroys the separation advatages.
Sure, you can VLAN it off, but those switches still have a finite amount of bandwidth available
Not to take too much from your thunder (where I work, we use Fibre Channel for SAN), but all of our core ethernet switches are non-blocking; when we give them more ports, they get more bandwidth. We could not only segment the traffic onto separate VLANs, but we could also put that traffic onto separate trunks. Ethernet's poor multi-pathing support is what deters us the most from iSCSI.
I would say sports adds some other aspect to it. It might be that sports teams are the new armies, they are "The New York Yankees" going up against "The Boston Red Sox".
I just think that when it's sports, it becomes more personal to people. Computers are just tools to most people, as are many other things.
Well, you glossed over an important point: how is your system going to check for authorization?
You have to contact the credit card company and they'll have to confirm that you are who you say you are, which would rely on password/smartcard/biometric. Guess what, those are all numbers! That's what happens when something is digitized, it becomes numbers.
So the credit card holder must know something that amounts to a number in computer terms, so there is always sensitive information external to the credit card company.
All systems will depend on a "stupid number". The credit card system is broken because it relies on a shared number. It should instead work like a smart card where the number is pretty much (not absolutely) unobtainable, and then do a challenge-response based authentication between the credit company and the card to authorize the purchase (which unlike what some other poster thinks, would only require computers, humans would be too likely to screw up the numbers). Another addition would be to have a keypad on the card to input a pin number to activate the card.
There is an extreme amount of misplaced trust in so many systems these days. Don't trust the network, don't trust the terminal, don't trust the holder (not necessarily the proper owner) of the card.
This is why I don't want an RFID credit card, it's still based off a security broken system.
Most of the letter patterns are different only because they optimized the typing of those words for the QWERTY layout. The original developers realized how bad QWERTY was for typing actual words, so they made most of the commands non-words.
Which seems to be changing now with newer languages, they're now using the whole word instead of the word minus the vowels.
At least he turned on his lights to signal that he was going to break the normal traffic laws.
I've seen cops roll through red lights and nearly hit pedestrians. This was because they gave no warning that they were going to do that. I feel like that would require a Cheney-like apology for getting in the police officer's way.
I believe most states have laws about exempting emergency vehicles when responding from emergencies from many traffic laws. Again, there is warning that they are about to break normal traffic laws so people know to get out of the way. There should also be a record of emergencies so they can be exempted from such tickets, otherwise they should just take it like the rest of us mere mortals.
The biggest oversight on Theo's part is that this wouldn't have made Slashdot if he hadn't replied. It would have just been some message on a mailing list, and most of humanity, the Internet, Linux people, would have never noticed or cared.
I now want to have as little as possible with Theo, and it'd be hilarious if he intentionally, knowingly, and purposefully put holes in OpenSSH to exploit Linux only. That would get him sued (because of those three great lawyer words I used). That and some of his comments in the thread about how to get around copyright would suggest a very poor understanding of law.
Not all languages have or need do/while, while, or for loops.
No, standards are a waste of time
on
IT, Be Free!
·
· Score: 1
I've read RFCs and other standards documents and have come to the conclusion that they are useless. They use overly verbose English, so implementations usually deviate from the specification. With all that verbosity, you'd think they wouldn't allow things to be unspecified, but it happens anyway.
So much time spent just to specify something, then on top of that, many people spend time actually making an implementation so a specification can actually be useful.
How many people actually implement on a standard? Very few, that's why portability doesn't work. The whole point of a standard is to help portability!
Someone will probably point out TCP or IP has being a successful standard, but they have the two main problems with standards: try using out-of-band data in TCP. Can't? Oh, that's right, few people actually use it because few have implemented it. Running out of addresses in IPv4? Time to write up a new standard that actually takes current necessities. When will there be an implementation? Few years. When will it actually be in use? Even longer.
I would say the root cause for needing standards is the erection of the code/data boundary because current operating systems can't trust random code.
You obviously have not studied computer science or software engineering, otherwise you would know that there is no way to prevent this kind of stuff for any OS
Please point me to this "insecurity theorem" of yours.
Have you looked at the patch? It's 250KB, 8800 lines, which includes documentation, copyright headers for every file, and all the diff commands.
Knowledgeable people have said before that a driver does not tell anything about how to make the hardware. It's still going to cost tons of money to develop a chip that's compatible with someone else's. Only a graphics card company could have an excuse since the drivers have become so complex. And even the DRI people have said that a company could either spend millions on reverse engineering or millions on developing their own card that wouldn't be behind the competition by the huge amount of time to reverse engineer such a complex piece of silicon.
Sure, it theoretically gives the competition a help, but practically it actually does very little.
It seems like much of your problems were the professor did not spend any time teaching, more time fixing the book.
I've learned Calculus three different ways. I had originally learned it from the simplification of the usual limit approach (which did not really have any proofs, just "convincing arguments"). Then later I learned the formalization of the Riemann integral and epsilon-delta notation, while at the same time I read this book on nonstandard analysis. I then later took a graduate analysis course, which used the Lebesgue integral on top of measure theory.
Nonstandard analysis has the simplest proofs once you have certain facts about the Hyperreals, the Riemann approach requires tons of epsilon-deltas and definitions of lower and upper limits, while the measure theory approach requires both epsilon-delta and set theory mastery (but lets you do things Riemann can't make sense of).
The Hyperreal approach is elegant in many ways since proofs that are hard using epsilon-deltas become one-line proofs (For instance the proof of the equivalence of Cauchy and convergent sequences is a one line proof in NSA using basic properties of infinitesimals, while it takes two half page theorems in SA). It basically gets all the hard work done in constructing the Hyperreals and then the rest is for free (but many properties and the construction of the Reals are ignored in SA), but it's not a well-accepted way of doing calculus. Which is the biggest point it has against it (I read a paper about using nonstandard analysis in computer proving systems since it only needs statements with one quantifier instead of the usual three).
Given that, I would imagine that it would be a really bad way to learn calculus not because it's a hard approach, but that it's an arcane one.
How long did it take to convert people from insecure telnet to ssh? There are even systems still that do not have encryption imposed that some twit I have to deal with doesn't see any reason for setting it up. The main problem is that if policy is not dictated people will be lazy. It's easier to not have encryption in the protocol and make devices that don't use encryption, or maybe they do encrypt but it's useless (WEP), or maybe they then require drivers to support it (so no Linux support or some such nonsense).
I know not to run services that don't use encryption over the internet or over wireless, but remember that there are lots of stupid people out there.
I've looked at getting a Tablet PC. Everyone I've seen has been underpowered and overpriced. For the same price, a notebook would be a better computer but not in the tablet form factor.
I also believe Tablet PC makers have been saying that the sales are not agreeing with the hype.
The problem seems to be that they are trying to sell them as notebook replacements instead of a computer that's a tablet. They're too thick, too heavy, don't have enough battery power, or are overpriced.
I'd like a Tablet PC to do those things you list, but they all suck.
as the creator of the work, I naturally expect to be able to control said work.
Which is to say that you want to control what everyone else can do. You only naturally expect it because you grew up with copyrights and patents existing. They've only existed from around when the US was founded.
Look where all those people with no control over "their" creations got us.
Now look where "control" over creations has got us:
We have a movie industry which makes bad movies.
We have a recording industry which makes bad recodings.
We have a publication industry wich makes bad publications.
We have a software industry which makes bad software.
The API is documented and exported but the actual behavior is not consistent with the documentation.
I think it's terrible when people perserve the functionality of the implementation instead of perserving the behavior described by the documentation. If a program relies on undocumented behavior, it is relying on a bug. Bugs get fixed, programs break.
I don't know if this is the case since I've not seen a list of apps that break and then what it took to fix them, but I've seen it happen before with glibc and sendmail and it doesn't surprise me since there are lots of people with the attitude "It runs so nothing is wrong."
Slashdot Mirror
Microkernels should die.
With type safe languages, they really should be compiling all the components into the same memory space and then they would be able to use the programming language to create the API between components instead of IPC.
Which would basically be Linux written in a type safe language. I can't find it, but IIRC, Linus commented once that Linux is like a microkernel with how the components of the system are separated, but they're using C function calls to initiate inter-communication and run in the same memory/process space. The same memory/process space is something microkernels use to get better performance, but that destroys the separation advatages.
Not to take too much from your thunder (where I work, we use Fibre Channel for SAN), but all of our core ethernet switches are non-blocking; when we give them more ports, they get more bandwidth. We could not only segment the traffic onto separate VLANs, but we could also put that traffic onto separate trunks. Ethernet's poor multi-pathing support is what deters us the most from iSCSI.
I would say sports adds some other aspect to it. It might be that sports teams are the new armies, they are "The New York Yankees" going up against "The Boston Red Sox".
I just think that when it's sports, it becomes more personal to people. Computers are just tools to most people, as are many other things.
Well, you glossed over an important point: how is your system going to check for authorization?
You have to contact the credit card company and they'll have to confirm that you are who you say you are, which would rely on password/smartcard/biometric. Guess what, those are all numbers! That's what happens when something is digitized, it becomes numbers.
So the credit card holder must know something that amounts to a number in computer terms, so there is always sensitive information external to the credit card company.
All systems will depend on a "stupid number". The credit card system is broken because it relies on a shared number. It should instead work like a smart card where the number is pretty much (not absolutely) unobtainable, and then do a challenge-response based authentication between the credit company and the card to authorize the purchase (which unlike what some other poster thinks, would only require computers, humans would be too likely to screw up the numbers). Another addition would be to have a keypad on the card to input a pin number to activate the card.
There is an extreme amount of misplaced trust in so many systems these days. Don't trust the network, don't trust the terminal, don't trust the holder (not necessarily the proper owner) of the card.
This is why I don't want an RFID credit card, it's still based off a security broken system.
Most of the letter patterns are different only because they optimized the typing of those words for the QWERTY layout. The original developers realized how bad QWERTY was for typing actual words, so they made most of the commands non-words.
Which seems to be changing now with newer languages, they're now using the whole word instead of the word minus the vowels.
At least he turned on his lights to signal that he was going to break the normal traffic laws.
I've seen cops roll through red lights and nearly hit pedestrians. This was because they gave no warning that they were going to do that. I feel like that would require a Cheney-like apology for getting in the police officer's way.
I believe most states have laws about exempting emergency vehicles when responding from emergencies from many traffic laws. Again, there is warning that they are about to break normal traffic laws so people know to get out of the way. There should also be a record of emergencies so they can be exempted from such tickets, otherwise they should just take it like the rest of us mere mortals.
The biggest oversight on Theo's part is that this wouldn't have made Slashdot if he hadn't replied. It would have just been some message on a mailing list, and most of humanity, the Internet, Linux people, would have never noticed or cared. I now want to have as little as possible with Theo, and it'd be hilarious if he intentionally, knowingly, and purposefully put holes in OpenSSH to exploit Linux only. That would get him sued (because of those three great lawyer words I used). That and some of his comments in the thread about how to get around copyright would suggest a very poor understanding of law.
Offering it through the store and steam means the stores will have shorter lines because people bought it through steam.
So even if you have a 300 baud connection, you still win with steam!
Not all languages have or need do/while, while, or for loops.
So much time spent just to specify something, then on top of that, many people spend time actually making an implementation so a specification can actually be useful.
How many people actually implement on a standard? Very few, that's why portability doesn't work. The whole point of a standard is to help portability!
Someone will probably point out TCP or IP has being a successful standard, but they have the two main problems with standards: try using out-of-band data in TCP. Can't? Oh, that's right, few people actually use it because few have implemented it. Running out of addresses in IPv4? Time to write up a new standard that actually takes current necessities. When will there be an implementation? Few years. When will it actually be in use? Even longer.
I would say the root cause for needing standards is the erection of the code/data boundary because current operating systems can't trust random code.
Why not just hire live performances? You'd get no distortion from anything but the room.
Have you looked at the patch? It's 250KB, 8800 lines, which includes documentation, copyright headers for every file, and all the diff commands.
Knowledgeable people have said before that a driver does not tell anything about how to make the hardware. It's still going to cost tons of money to develop a chip that's compatible with someone else's. Only a graphics card company could have an excuse since the drivers have become so complex. And even the DRI people have said that a company could either spend millions on reverse engineering or millions on developing their own card that wouldn't be behind the competition by the huge amount of time to reverse engineer such a complex piece of silicon.
Sure, it theoretically gives the competition a help, but practically it actually does very little.
I've learned Calculus three different ways. I had originally learned it from the simplification of the usual limit approach (which did not really have any proofs, just "convincing arguments"). Then later I learned the formalization of the Riemann integral and epsilon-delta notation, while at the same time I read this book on nonstandard analysis. I then later took a graduate analysis course, which used the Lebesgue integral on top of measure theory.
Nonstandard analysis has the simplest proofs once you have certain facts about the Hyperreals, the Riemann approach requires tons of epsilon-deltas and definitions of lower and upper limits, while the measure theory approach requires both epsilon-delta and set theory mastery (but lets you do things Riemann can't make sense of).
The Hyperreal approach is elegant in many ways since proofs that are hard using epsilon-deltas become one-line proofs (For instance the proof of the equivalence of Cauchy and convergent sequences is a one line proof in NSA using basic properties of infinitesimals, while it takes two half page theorems in SA). It basically gets all the hard work done in constructing the Hyperreals and then the rest is for free (but many properties and the construction of the Reals are ignored in SA), but it's not a well-accepted way of doing calculus. Which is the biggest point it has against it (I read a paper about using nonstandard analysis in computer proving systems since it only needs statements with one quantifier instead of the usual three).
Given that, I would imagine that it would be a really bad way to learn calculus not because it's a hard approach, but that it's an arcane one.
If you can prove to me that there is only one "intuition", I will then believe that your position is not utter bull shit.
I know not to run services that don't use encryption over the internet or over wireless, but remember that there are lots of stupid people out there.
Nah, I just think it's everyone else looking out for themselves. Not everyone drives so badly and those people are looking out for someone who is.
I also believe Tablet PC makers have been saying that the sales are not agreeing with the hype.
The problem seems to be that they are trying to sell them as notebook replacements instead of a computer that's a tablet. They're too thick, too heavy, don't have enough battery power, or are overpriced.
I'd like a Tablet PC to do those things you list, but they all suck.
Look where all those people with no control over "their" creations got us.
Now look where "control" over creations has got us:
Plex86 has to be run on x86.
Plex86 is not an emulator.
Looks like he doesn't like PCs, so he doesn't want to have one in his house. Looks like you're an anti-mac zealot.
The API is documented and exported but the actual behavior is not consistent with the documentation.
I think it's terrible when people perserve the functionality of the implementation instead of perserving the behavior described by the documentation. If a program relies on undocumented behavior, it is relying on a bug. Bugs get fixed, programs break.
I don't know if this is the case since I've not seen a list of apps that break and then what it took to fix them, but I've seen it happen before with glibc and sendmail and it doesn't surprise me since there are lots of people with the attitude "It runs so nothing is wrong."