I think most of us with the low uids have moved into positions where we don't have as much time to post anymore.
I think it has more to do with there being no "moron" moderation tag. There used to be a few clued in people in most discussions but now there will almost always be a +5 rated off topic discussion about something else.
The 1st issue is that to be an Auditor you have to be in the business of selling security stuff. That is a serious conflict of interest.
The 2nd issue is that the PCI auditors are foolish enough to be set up to take the blame and provide insurance when a company fails. Lets assume that a processors gets hacked and is sending card numbers off to mob in a different country. How do banks cover reissuing the cards and recovering anything they don't stick the merchants with? In this case the processor that is handing off the numbers ends up bankrupt so there there is no blood left in that stone and the banks are just the members of the card schemes so the only ones left are the merchants and now thanks to PCI, the audit companies and their insurance policies.
Is there any wonder why most of the best groups that did past audits won't touch them anymore?
It started about 1990. In 1985 tuition was on the order of $45 per class at state schools and you were expected to take 8 to 10 a year. Then a bunch of federal programs got introduced and the size of administration got larger than the teaching staff and then the bigger loans were needed so the schools needed larger loan processing groups so they could get paid and with larger loans the could build newer buildings and the cycle keeps going. The student loans were fixed at 7 years max in 1989 but now they go on forever with 20 year terms not uncommon. The debt figures even get worse when you consider that once your out of school with your 1st job, there isn't enough cash flow to cover the start up costs so things like work suits get put on a credit card. Add in the poorer economic situation and home loans are for higher and longer amounts. The net result is that the extra administration is putting people in debt for decades. When my father went to school, he had to show up a week before classes to work on maintenance of the buildings.
If its so easy to parse, why are the libraries that parse it such a pain to use? Also why do they burn memory like crazy as well? Knuth has written a few things about why this style of parsing is bad (he used it in TeX) and why it should be avoided for anything but interactive stuff. If you do the proofs, you will see that XML errors diverge to two states, one takes an infinite amount of time and the other an infinite amount of memory.
The new long haul stuff means you don't need under sea repeaters at all if you stick to the Pacific rim and avoid Hawaii. Under sea fiber armored fiber runs about $7/m but the repeaters run about $1 million each which is why there tends to be only one or two pair used. When you can reduce the undersea infrastructure costs from about $2 billion the old way to $200m using on land repeaters, the ROI make sense for many major data users.
Its real role is to slow down starting new wars in a bureaucratic mess that takes a long time to work out so that new wars don't break out quickly like they did up to the early 20th century. There is no organization on earth that can create a bureaucratic mess like the UN can.
Freakin' Unix weenies, determined to forget the lessons of the past... As the 10th commandment of C goes: Thou shalt foreswear, renounce, and abjure the vile heresy which claimeth that ``All the world's a VAX^wSun^wLinux'', and have no commerce with the benighted heathens who cling to this barbarous belief, that the days of thy program may be long even though the days of thy current machine be short.
The latest apache 2.0 autoconf insists on gnu grep even though there is no requirement for grep to deal with large command lines at all other than the test in the configure script.
And with Solaris, I'm not happy with what is now on the 1st disk since its way too obese.
What do you mean compilers are never "done"? I use finished compilers every time I have to fix something that was written in the academic or trendy language of the day.
Well Sergey won the NSF grant that was instrumental in his staying at Stanford. That grant is a result of a random selection out of the best candidates. To even get that far you have to know the right people or else you can't get the right recommendations no matter how hard you worked or how smart you are. If he didn't go to Stanford would he be a billionaire now? How does one even get into the Stanford PhD program anyway? They get far more applications than they have positions and everyone that is considered will be very well qualified. There is some luck into be accepted since I expect many of the final votes come down to random choices made by the selection committee while selecting between two equal candidates. Page had a huge advantage by the luck of being born to the parents he had. Remember his brother is also a billionaire.
All the very rich people I know worked about as hard as most of my successful friends. Thats based on a small sample sizes (no billionaires but a handful of those who got 9 digit checks). They all were very lucky to be in the right place at the right time. I know lots of others others who worked hard and had it all destroyed by bad luck.
Google making money out of the idea was a result in being able to talk to the right people at the right time. They didn't have any magic technology at hand but they were unique compared to their competition in that they had enough resources to demo their early work. They could pull that off because they had been in the right place at the right time a bit before they where in the right place earlier. Most what is now considered their innovation was all discussed on usenet news groups long before their research was done. You can even look it up in google groups if you want.
Somewhere I have my rejection letter from UC Berkeley. It said they had over a million applications since they filled up and they didn't even bother considering my application. That was when UCB had a better comp-sci program than MIT did. I applied the 1st day I could based on when my high school did its testing. That was just bad luck and completely out of my control. I had a better chance of getting into Stanford or MIT than most students in my state but I went to one of the best high schools in the state and both schools limited how much aid could go to people from one high school and I lost out even if my parents had enough cash to pay for either school. There is a huge amount of luck required to get into the best schools in the world.
Its because making it big is 99% luck and less than 1% hard work. These guys made their money because they were in the right place at the right time which lead to meeting the right people (when the people with the money were willing to spend some). They also lucked out getting into a university that helped get them into the right place. Look at all the other dot com millionaires and look at how lucky they were to be in the right place at the right time. Even BillyG lucked out to have contacts into major companies like IBM thanks to his mother. With out her assistance, there would have been no way his company could have ever gotten the meetings that landed them their big contracts.
I know plenty of people who worked harder but got no where mostly due to things out of their control.
This is good because it causes the FAA to concentrate on the toys rather that whatever stupid thing they were meeting to collaborate on. If it makes their top level bureaucracy even less efficient, it means they won't be making new rules to break their existing sub-optimal but working system.
The fsck is so the partition is mapped as clean if you need to use it. It also lets you know if the copy was done at a bad time when the super-block was in a funny state so if fsck can't fix it automatically you might want to try again. With some OS's you can create a snapshot and dd that as the source. On large partitions, dump/restore or rsync will be faster in many cases but I find the low level copy works well and is better behaved when things go funny.
RAID was originally about Redundancy using cheap hardware. I'm doing the same thing but with a different and non-real time algorithm. The performance gains from striping come at a risk of data corruption that I have never found to be worth the tradeoff.
If you look at MTBF on modern systems, any system where the higher RAID levels make sense will always have something broken. Many problems can be re-factored into a way so that the Google hardware model makes more sense than spending an exponential amount of money to ensure you never lose any data.
You load balance over the disk using partitions. For example my web/news server has web stuff on disk 0 and news on disk 1 but there is a backup for the web stuff on disk 1 and a backup of the news (config at least) on disk 0. One trick to keep in mind is your partition mapping may look odd or unusual.
The reason for the dd to/dev/null is to make sure the disk is sill good. If that fails, I do a manual sync between the soon to be dead disk and the last known good disk using something like rsync --show_me_but_don't_do_it and then manually copy files making sure I can read them 100% first.
When expert systems were all the rage in the late 80's no one thought of hooking an expert system to an email router? I'm guessing 20 minutes in the masters thesis section of UCB or MIT would provide more than enough prior art to kill this dead.
Your right that vacation(1) won't meet all of the claims however vacation gets tired into sendmail's "deliver to a pipe" system and a bit of research on that show lots of programs that start breaking down most (if not all) of the other claims way before the patent was filed.
The real problem here is that in 1979 (when I think most of this was well known for people "Practiced in the Art", the patent office wouldn't allow people to patent software so no one was sending them applications. A few decades later and they decide its ok to patent software but they didn't have anything in their prior art library and they didn't have any examiners in the field so they lat anything go.
Consumer protection laws in most countries require Microsoft to recall their software due to damage its done to innocent 3rd parties yet where in the world did that happen? How about free (or $2) CDs at the local computer shop that will reinstall and patch whatever disks people are likely to have.
Remember grandma with the hacked computer is running software that is owned by Microsoft. She only licensed it and the owner is still to blame.
Many of the safety features you mention are there because of UL. UL is a lab set up by insurance companies to encourage safer products to save the insurance company money.
I wonder what will happen if a bunch of insurance companies all got hit with suites going after home owner liability insurance payoffs. Would the insurance companies then got after MS or would they just force all insured home owners to run the latest version of their favorite corps bad anti-virus code?
And they bought a license from AT&T and UCB and thats just the majors. They also did deals with IBM, DEC, HP, Apollo, PE, Pyramid and others. Just because their deal with SCO is a bit fishy doesn't mean they aren't fully covered.
They can't fake an IP address if you have pipelining turned on either but you have to flush the input buffer before you send back the response to the EHLO. I did a simple hack that looked at the queue size at the time the greeting was issued and purged lots of spam that way.
Slashdot Mirror
I think most of us with the low uids have moved into positions where we don't have as much time to post anymore.
I think it has more to do with there being no "moron" moderation tag. There used to be a few clued in people in most discussions but now there will almost always be a +5 rated off topic discussion about something else.
The 1st issue is that to be an Auditor you have to be in the business of selling security stuff. That is a serious conflict of interest.
The 2nd issue is that the PCI auditors are foolish enough to be set up to take the blame and provide insurance when a company fails. Lets assume that a processors gets hacked and is sending card numbers off to mob in a different country. How do banks cover reissuing the cards and recovering anything they don't stick the merchants with? In this case the processor that is handing off the numbers ends up bankrupt so there there is no blood left in that stone and the banks are just the members of the card schemes so the only ones left are the merchants and now thanks to PCI, the audit companies and their insurance policies.
Is there any wonder why most of the best groups that did past audits won't touch them anymore?
It started about 1990. In 1985 tuition was on the order of $45 per class at state schools and you were expected to take 8 to 10 a year. Then a bunch of federal programs got introduced and the size of administration got larger than the teaching staff and then the bigger loans were needed so the schools needed larger loan processing groups so they could get paid and with larger loans the could build newer buildings and the cycle keeps going. The student loans were fixed at 7 years max in 1989 but now they go on forever with 20 year terms not uncommon. The debt figures even get worse when you consider that once your out of school with your 1st job, there isn't enough cash flow to cover the start up costs so things like work suits get put on a credit card. Add in the poorer economic situation and home loans are for higher and longer amounts. The net result is that the extra administration is putting people in debt for decades. When my father went to school, he had to show up a week before classes to work on maintenance of the buildings.
If its so easy to parse, why are the libraries that parse it such a pain to use? Also why do they burn memory like crazy as well?
Knuth has written a few things about why this style of parsing is bad (he used it in TeX) and why it should be avoided for anything but interactive stuff. If you do the proofs, you will see that XML errors diverge to two states, one takes an infinite amount of time and the other an infinite amount of memory.
The new long haul stuff means you don't need under sea repeaters at all if you stick to the Pacific rim and avoid Hawaii. Under sea fiber armored fiber runs about $7/m but the repeaters run about $1 million each which is why there tends to be only one or two pair used. When you can reduce the undersea infrastructure costs from about $2 billion the old way to $200m using on land repeaters, the ROI make sense for many major data users.
The UN does its intended role very well.
Its real role is to slow down starting new wars in a bureaucratic mess that takes a long time to work out so that new wars don't break out quickly like they did up to the early 20th century. There is no organization on earth that can create a bureaucratic mess like the UN can.
Freakin' Unix weenies, determined to forget the lessons of the past...
As the 10th commandment of C goes:
Thou shalt foreswear, renounce, and abjure the vile heresy which claimeth that ``All the world's a VAX^wSun^wLinux'', and have no commerce with the benighted heathens who cling to this barbarous belief, that the days of thy program may be long even though the days of thy current machine be short.
The latest apache 2.0 autoconf insists on gnu grep even though there is no requirement for grep to deal with large command lines at all other than the test in the configure script.
And with Solaris, I'm not happy with what is now on the 1st disk since its way too obese.
What do you mean compilers are never "done"?
I use finished compilers every time I have to fix something that was written in the academic or trendy language of the day.
Well Sergey won the NSF grant that was instrumental in his staying at Stanford. That grant is a result of a random selection out of the best candidates. To even get that far you have to know the right people or else you can't get the right recommendations no matter how hard you worked or how smart you are. If he didn't go to Stanford would he be a billionaire now? How does one even get into the Stanford PhD program anyway? They get far more applications than they have positions and everyone that is considered will be very well qualified. There is some luck into be accepted since I expect many of the final votes come down to random choices made by the selection committee while selecting between two equal candidates. Page had a huge advantage by the luck of being born to the parents he had. Remember his brother is also a billionaire.
All the very rich people I know worked about as hard as most of my successful friends. Thats based on a small sample sizes (no billionaires but a handful of those who got 9 digit checks). They all were very lucky to be in the right place at the right time. I know lots of others others who worked hard and had it all destroyed by bad luck.
Google making money out of the idea was a result in being able to talk to the right people at the right time. They didn't have any magic technology at hand but they were unique compared to their competition in that they had enough resources to demo their early work. They could pull that off because they had been in the right place at the right time a bit before they where in the right place earlier. Most what is now considered their innovation was all discussed on usenet news groups long before their research was done. You can even look it up in google groups if you want.
Somewhere I have my rejection letter from UC Berkeley. It said they had over a million applications since they filled up and they didn't even bother considering my application. That was when UCB had a better comp-sci program than MIT did. I applied the 1st day I could based on when my high school did its testing. That was just bad luck and completely out of my control. I had a better chance of getting into Stanford or MIT than most students in my state but I went to one of the best high schools in the state and both schools limited how much aid could go to people from one high school and I lost out even if my parents had enough cash to pay for either school. There is a huge amount of luck required to get into the best schools in the world.
Its because making it big is 99% luck and less than 1% hard work. These guys made their money because they were in the right place at the right time which lead to meeting the right people (when the people with the money were willing to spend some). They also lucked out getting into a university that helped get them into the right place. Look at all the other dot com millionaires and look at how lucky they were to be in the right place at the right time. Even BillyG lucked out to have contacts into major companies like IBM thanks to his mother. With out her assistance, there would have been no way his company could have ever gotten the meetings that landed them their big contracts.
I know plenty of people who worked harder but got no where mostly due to things out of their control.
This is good because it causes the FAA to concentrate on the toys rather that whatever stupid thing they were meeting to collaborate on. If it makes their top level bureaucracy even less efficient, it means they won't be making new rules to break their existing sub-optimal but working system.
No, NNTP works fine, you just have to grep the news spool for a recently posted article.
The fsck is so the partition is mapped as clean if you need to use it. It also lets you know if the copy was done at a bad time when the super-block was in a funny state so if fsck can't fix it automatically you might want to try again. With some OS's you can create a snapshot and dd that as the source. On large partitions, dump/restore or rsync will be faster in many cases but I find the low level copy works well and is better behaved when things go funny.
RAID was originally about Redundancy using cheap hardware. I'm doing the same thing but with a different and non-real time algorithm. The performance gains from striping come at a risk of data corruption that I have never found to be worth the tradeoff.
If you look at MTBF on modern systems, any system where the higher RAID levels make sense will always have something broken. Many problems can be re-factored into a way so that the Google hardware model makes more sense than spending an exponential amount of money to ensure you never lose any data.
You load balance over the disk using partitions. For example my web/news server has web stuff on disk 0 and news on disk 1 but there is a backup for the web stuff on disk 1 and a backup of the news (config at least) on disk 0. One trick to keep in mind is your partition mapping may look odd or unusual.
/dev/null is to make sure the disk is sill good. If that fails, I do a manual sync between the soon to be dead disk and the last known good disk using something like rsync --show_me_but_don't_do_it and then manually copy files making sure I can read them 100% first.
The reason for the dd to
This is why I use poor mans pseudo-RAID
Cron kicks off with a dd if=$SRC of=/dev/null && dd if=$DEST of=/dev/null && dd if=$SRC of=$DEST && fsck $DEST
With two disks in a system, one disk gets 1/2 the daily use and the other disk gets the other half.
I've seen ltos of disks die and this system has never let me down. I can't say that about hardware or
software RAID solutions.
I use rsync to copy other stuff off site.
When expert systems were all the rage in the late 80's no one thought of hooking an expert system to an email router? I'm guessing 20 minutes in the masters thesis section of UCB or MIT would provide more than enough prior art to kill this dead.
And fax machines predate the telephone.
Your right that vacation(1) won't meet all of the claims however vacation gets tired into sendmail's "deliver to a pipe" system and a bit of research on that show lots of programs that start breaking down most (if not all) of the other claims way before the patent was filed.
The real problem here is that in 1979 (when I think most of this was well known for people "Practiced in the Art", the patent office wouldn't allow people to patent software so no one was sending them applications. A few decades later and they decide its ok to patent software but they didn't have anything in their prior art library and they didn't have any examiners in the field so they lat anything go.
Consumer protection laws in most countries require Microsoft to recall their software due to damage its done to innocent 3rd parties yet where in the world did that happen? How about free (or $2) CDs at the local computer shop that will reinstall and patch whatever disks people are likely to have.
Remember grandma with the hacked computer is running software that is owned by Microsoft. She only licensed it and the owner is still to blame.
Many of the safety features you mention are there because of UL. UL is a lab set up by insurance companies to encourage safer products to save the insurance company money.
I wonder what will happen if a bunch of insurance companies all got hit with suites going after home owner liability insurance payoffs. Would the insurance companies then got after MS or would they just force all insured home owners to run the latest version of their favorite corps bad anti-virus code?
And they bought a license from AT&T and UCB and thats just the majors. They also did deals with IBM, DEC, HP, Apollo, PE, Pyramid and others.
Just because their deal with SCO is a bit fishy doesn't mean they aren't fully covered.
They can't fake an IP address if you have pipelining turned on either but you have to flush the input buffer before you send back the response to the EHLO. I did a simple hack that looked at the queue size at the time the greeting was issued and purged lots of spam that way.