How do we turn people in? I suspect there is a guy running a super bowl party on lots of big screens and he is doing it for political purposes which also violate the letter of the law. Maybe the NFL can raid his house? Its at 1600 Pennsylvania Avenue.
i think the point is to think small here... the idea is to allow the smallest set of changes that work to kernels and code that still works and increases real security. The patch I mentioned works well with most open source programs that don't have code like if(uid!=0) winge(must be run as root) when they should have been if(bind()==FAIL) {winge(must be run as roo)}. My early tests with this showed that it worked great for bind, most mtas, web servers (that don't run cgis as their users but thats a different mess). Sure its not the best way but it doesn't break anything (that at least one unix systems hasn't broken already) and it kills the number one reason for setuid code.
Have good passwords. I had a program long ago that would tell you how quickly it would guess your password. It was based on one of the programs like "John the Ripper" and it would tell you the detail of which pass and rule set it would generate your password so that it might report "l3tm31n" as 1) common password "letmein" with 2) 1 typewritter leters for numbers for the i=1 and 3) leet speek for e=3 too bad I can't find that program. I think it would make an excellent pam module.
I've been pushing for group mappings to map to ports for a while with no luck. The idea is that you put your apache startup user in group 80 and 443 and your email listern process in group 25 and 587 and then they don't have to be root at all to bind. It used to be about a 40 character patch in the Linux kernel but no one ever seemed to interested.
The best thing that could happen to most state universities would be for them and the nearby states to all lose their federal funding at the same time. There would be a year or two of real pain for some students and some professors and loads of pain for most of the administration but the resulting downsizing and refocussing on funding sources outside of the fed would result in better education and a far better deal for the students.
But if you run a few thousand dns queries against the domain, they buy it. It appears to take anything entered into their search box and if its made of dictionary words (random stuff didn't for me all the time). Also it appears you may need another registrar to do a whois for them to trigger their "buy".
We know they count DNS hits to figure out what to sell the domain for latter.
Someone needs to start dropping law suits on the Commerce Departments desk over their contract for this.
The terrorist aren't going to be using state of the art system. They will be using either point and shoot, fly by real wire or heat seekers. This system is pointless against two of those and will increase the odds of the missile causing critical damage to a plane in the other case.
Re:OOD has doomed us all to bad standards
on
RTF Vs. OOXML
·
· Score: 1
I have TeX documents from 1985 that still render the same way. LaTeX documents from then do not always render the same way they used to.
Standards don't work universally. Thats why there are so many of them. ASCII worked so well in the early days. And then there are the hacks.
ODF only works for a majority of documents but most people involved with that standard can create something that is legal but won't render the same way on many platforms. You can repeat that for HTML, PDF, GIF, JPG etc.
With a word process you can store either the input, output, memory or processing. Word Perfect up to 5.1 stored input (with hacks for specific printers). TeX stores a reflection of the internal processing model. PDF stores the output model. Word stores its storage model.
Re:I'm sorry, I don't get your point.
on
RTF Vs. OOXML
·
· Score: 2, Interesting
Output on different devices is what a word processor does. Other wise its "process" step wouldn't be anything at all. And the 1950's model doesn't consider "storage" part of the I/O model but modern sanity sort of would imply it unless its direct memory or object dump.
Words storage model is odd. It doesn't match the input model but is sort of based on the output model assuming line printers with loads of hacks to make it work with modern printers. As far as I know its been that way since I was using Word (for Unix on a 3b2) in 1987.
OOD has doomed us all to bad standards
on
RTF Vs. OOXML
·
· Score: 1, Troll
Computers do Input, Output, Processing and Storage. Its been that way for more than a half centuy. At some point people thought it would be cool if that wasn't the case and dreamed up lots of crud to put in text books sold to college students and they made lots of money but hasn't changed a thing.
It still doesn't fix the problem that a word processor has an internal model of what the user typed. Its job is to output that in a way that is consistent with what it's showing the user and what the user told it to do. Now for some odd reason a large group of people come along and say "we want magic" and expect the input/output and storage models to be disassociated. How is that supposed to work? Remapping input? More levels of indirection? It sill doesn't fix the core problem.
One example would be 8 mb of cache tag ram can deal with the core look ups for a packet routing engine in under 2ns. The idea is "Content Addressable Memory"
About 15 years ago I found that some 32 bit CPUs at the time will let you use their cache as content addressable memory for tagged look ups which were incredibly fast. I even hacked on an access system to the sys_v ipc drivers to allow user level apps to get access. This is a feature that could be very useful for so much software yet it never went anywhere and the hardware will still do it yet nothing exists for getting at that from userland programs. I could see a brand new OS that used that concept as having huge promise but only in niche areas like router or car OS not an OS to run my word processor and spreadsheet.
The US military has been promoting American style football for years since it is the best tool they have to make soldiers starting at a very young age. It teaches respect for the command structure and teamwork as well as the lesson that not everyone is cut out to be the leader. Back in the days in the Army Navy game was big, they were tweaking the rules that later ended up in collage and pro football.
Too bad the hard drives in apples are not easy to get at. It takes me about 1/2 hour to swap a disk in a laptop I know and its taken about 2 hours for one I couldn't find good take apart instructions.
Out of 20+ bulbs I have purchased in Australia, their power factors run from.2 to.6ish. That means the power grid losses on some of the 20w CFL are the same as the power grid losses of a 100W incandescent bulb. How do you like that 17% increase in power bills starting next month? Part of that is a result of grid inefficiencies thanks to the ever increasing load devices with bad power factors.
Except that the mercury put in the air during the manufacture of the bulbs isn't included and the mercury put in the bulbs often comes from coal scrubbers. Wal-mart is calming to reduce the amount of mercury in its bulbs by 4 mg per bulb which would imply they use more than that.
The mercury argument assumes the power comes from coal. You might want to check the source of the mercury that is in the Chinese made bulbs. Most of it comes from mercury scrubbers in coal burning power plants. There is also that problem that mercury in CFLs is in a form where it can be rapidly absorbed by living tissue which makes its risks fit into one of those inverse distance squared problems. If these bulbs can be recycled, can you explain who does it? And I mean full recycled, not ship the bits off someplace else. As far as I know there is no equipment on the planet to recycle them since they have to be completely disassembled and they have to remain intact until then. Even the long tubes used in offices aren't recycled much and there are machines to take them apart and recapture their mercury. Even if you do manage to recycle the metal bits, no one wants to buy mercury and lead contaminated glass and metals so it goes off to a landfill.
I figure they have about 400 file cabinets worth of plans for the Sat V at the federal archive plus what ever is still in microfilm, microfiche and other nearly unreadable formats in other locations but most of the documentation has been trashed or otherwise destroyed. While even NASA claims to have lots of info, I don't think that would even be close to what would be needed to build another Sat V and it can't include details of normal off the shelf parts let alone the complex system. For example a radar altimeter used in the early 727 was designed by many of the same people who designed the one for lunar lander. The test procedure documentation for the 727 radar took a wall of file cabinets to hold all its documents. Repeat that for every subsystem and you end up with a lot of info.
Slashdot Mirror
Telstra handed out lots of sat phones because even their extended range GSM couldn't cover the range of the old AMPS system.
Australia is mix between places with no people and very high density where the US has much more area with very few people.
I can do live splices about 10% of the time without the link going down on 10 gig optics without the right tools on dry land.
How do we turn people in? I suspect there is a guy running a super bowl party on lots of big screens and he is doing it for political purposes which also violate the letter of the law. Maybe the NFL can raid his house? Its at 1600 Pennsylvania Avenue.
i think the point is to think small here... the idea is to allow the smallest set of changes that work to kernels and code that still works and increases real security. The patch I mentioned works well with most open source programs that don't have code like if(uid!=0) winge(must be run as root) when they should have been if(bind()==FAIL) {winge(must be run as roo)}. My early tests with this showed that it worked great for bind, most mtas, web servers (that don't run cgis as their users but thats a different mess). Sure its not the best way but it doesn't break anything (that at least one unix systems hasn't broken already) and it kills the number one reason for setuid code.
Have good passwords.
I had a program long ago that would tell you how quickly it would guess your password. It was based on one of the programs like "John the Ripper" and it would tell you the detail of which pass and rule set it would generate your password so that it might report "l3tm31n" as
1) common password "letmein" with 2) 1 typewritter leters for numbers for the i=1 and 3) leet speek for e=3
too bad I can't find that program. I think it would make an excellent pam module.
I've been pushing for group mappings to map to ports for a while with no luck. The idea is that you put your apache startup user in group 80 and 443 and your email listern process in group 25 and 587 and then they don't have to be root at all to bind. It used to be about a 40 character patch in the Linux kernel but no one ever seemed to interested.
The best thing that could happen to most state universities would be for them and the nearby states to all lose their federal funding at the same time. There would be a year or two of real pain for some students and some professors and loads of pain for most of the administration but the resulting downsizing and refocussing on funding sources outside of the fed would result in better education and a far better deal for the students.
But if you run a few thousand dns queries against the domain, they buy it.
It appears to take anything entered into their search box and if its made of dictionary words (random stuff didn't for me all the time).
Also it appears you may need another registrar to do a whois for them to trigger their "buy".
We know they count DNS hits to figure out what to sell the domain for latter.
Someone needs to start dropping law suits on the Commerce Departments desk over their contract for this.
The party that lost will hand count the votes and if their numbers don't match the machine, all votes are recounted.
The terrorist aren't going to be using state of the art system. They will be using either point and shoot, fly by real wire or heat seekers. This system is pointless against two of those and will increase the odds of the missile causing critical damage to a plane in the other case.
I have TeX documents from 1985 that still render the same way. LaTeX documents from then do not always render the same way they used to.
Standards don't work universally. Thats why there are so many of them. ASCII worked so well in the early days. And then there are the hacks.
ODF only works for a majority of documents but most people involved with that standard can create something that is legal but won't render the same way on many platforms. You can repeat that for HTML, PDF, GIF, JPG etc.
With a word process you can store either the input, output, memory or processing. Word Perfect up to 5.1 stored input (with hacks for specific printers). TeX stores a reflection of the internal processing model. PDF stores the output model. Word stores its storage model.
Output on different devices is what a word processor does. Other wise its "process" step wouldn't be anything at all. And the 1950's model doesn't consider "storage" part of the I/O model but modern sanity sort of would imply it unless its direct memory or object dump.
Words storage model is odd. It doesn't match the input model but is sort of based on the output model assuming line printers with loads of hacks to make it work with modern printers. As far as I know its been that way since I was using Word (for Unix on a 3b2) in 1987.
Computers do Input, Output, Processing and Storage. Its been that way for more than a half centuy.
At some point people thought it would be cool if that wasn't the case and dreamed up lots of crud to put in text books sold to college students and they made lots of money but hasn't changed a thing.
It still doesn't fix the problem that a word processor has an internal model of what the user typed. Its job is to output that in a way that is consistent with what it's showing the user and what the user told it to do. Now for some odd reason a large group of people come along and say "we want magic" and expect the input/output and storage models to be disassociated. How is that supposed to work? Remapping input? More levels of indirection? It sill doesn't fix the core problem.
One example would be 8 mb of cache tag ram can deal with the core look ups for a packet routing engine in under 2ns.
The idea is "Content Addressable Memory"
I hope their training covers the differences in expressions between terrorist activity and the expressions of disdain and contempt.
About 15 years ago I found that some 32 bit CPUs at the time will let you use their cache as content addressable memory for tagged look ups which were incredibly fast. I even hacked on an access system to the sys_v ipc drivers to allow user level apps to get access. This is a feature that could be very useful for so much software yet it never went anywhere and the hardware will still do it yet nothing exists for getting at that from userland programs. I could see a brand new OS that used that concept as having huge promise but only in niche areas like router or car OS not an OS to run my word processor and spreadsheet.
The US military has been promoting American style football for years since it is the best tool they have to make soldiers starting at a very young age. It teaches respect for the command structure and teamwork as well as the lesson that not everyone is cut out to be the leader. Back in the days in the Army Navy game was big, they were tweaking the rules that later ended up in collage and pro football.
Too bad the hard drives in apples are not easy to get at. It takes me about 1/2 hour to swap a disk in a laptop I know and its taken about 2 hours for one I couldn't find good take apart instructions.
Blocking the UV is only part of the problem for art. The UV from CFLs makes ozone and that is very bad for bright color pigments.
Out of 20+ bulbs I have purchased in Australia, their power factors run from .2 to .6ish. That means the power grid losses on some of the 20w CFL are the same as the power grid losses of a 100W incandescent bulb. How do you like that 17% increase in power bills starting next month? Part of that is a result of grid inefficiencies thanks to the ever increasing load devices with bad power factors.
Except that the mercury put in the air during the manufacture of the bulbs isn't included and the mercury put in the bulbs often comes from coal scrubbers. Wal-mart is calming to reduce the amount of mercury in its bulbs by 4 mg per bulb which would imply they use more than that.
How does 118 parts per billion relate to .2mg/kg?
The mercury argument assumes the power comes from coal.
You might want to check the source of the mercury that is in the Chinese made bulbs. Most of it comes from mercury scrubbers in coal burning power plants. There is also that problem that mercury in CFLs is in a form where it can be rapidly absorbed by living tissue which makes its risks fit into one of those inverse distance squared problems. If these bulbs can be recycled, can you explain who does it? And I mean full recycled, not ship the bits off someplace else. As far as I know there is no equipment on the planet to recycle them since they have to be completely disassembled and they have to remain intact until then. Even the long tubes used in offices aren't recycled much and there are machines to take them apart and recapture their mercury.
Even if you do manage to recycle the metal bits, no one wants to buy mercury and lead contaminated glass and metals so it goes off to a landfill.
I figure they have about 400 file cabinets worth of plans for the Sat V at the federal archive plus what ever is still in microfilm, microfiche and other nearly unreadable formats in other locations but most of the documentation has been trashed or otherwise destroyed. While even NASA claims to have lots of info, I don't think that would even be close to what would be needed to build another Sat V and it can't include details of normal off the shelf parts let alone the complex system. For example a radar altimeter used in the early 727 was designed by many of the same people who designed the one for lunar lander. The test procedure documentation for the 727 radar took a wall of file cabinets to hold all its documents. Repeat that for every subsystem and you end up with a lot of info.