Dude, no reasonable person could read the linked document and conclude that accessing wikileaks on a personal computer and/or down time was anything other than contraindicated. You have to -want- to misunderstand in order to reach another conclusion.
It seemed pretty clear to me: intentionally access wikileaks, lose your clearance lose your job. Inadvertently access wikilieaks, immediately file an security report to avoid losing your clearance and job.
Wireless is a competitive market so carriers are permitted to do what they want with the presumption that if the customer doesn't like it they'll go to another vendor.
Wired broadband typically suffers from a monopoly or duopoly. There typically isn't a lot of choice available to the consumer, so there are strong restrictions on what the vendor is permitted to do.
You don't have to buy the entire game, You can generally find the boards that were removed from the cabinets when the cabinets were upgraded to some other game. Prices range from $20 to several hundred dollars depending on the popularity of the game.
The legal status of copying those roms on to modern hardware is in doubt. The moral status is more clear: of course you can put the board in the closet and do the minimum it takes to run the software on modern hardware.
Nobody with a brain ever said Marx was wrong about capitalism. Were he so clearly wrong, he wouldn't have inspired a quarter of the world to revolution. What they said about Marx was:
1. His alternative, communism, suffered from problems far more grave than the system it was proposed to replace. Communism's structure is fundamentally unstable -- it tends to shift into despotism. That's a double-whammy for communism because it's supposed to start with a revolution where a society is already very unstable. And that's only the beginning of its problems... Central planning for large systems rarely works in the first place and tends to devolve over time when it does. And any computer person has an intuitive understanding of the euphemism "design by committee."
2. Capitalism's worst problems could be mitigated. Things like the Sherman Antitrust Act and the SEC are responsive to the problems Marx examined. Did we get it wrong last decade? Yes, Bush screwed the pooch with a particularly ignorant version of deregulation while the world was busy living beyond its means. But unlike communism, the capitalist systems system will recover over the course of the next few years and, for a while at least, will be stronger for it.
They were censored because the ISPs are fucking retarded.
You wouldn't say that had you ever worked as a network engineer at a large ISP.
First, you'd have to route IP packets for the impacted address to an internal filtering machine. What filtering machine? Well, that's a rub too... you'd have to build one and while it's possible with open source, it isn't easy or particularly cheap.
Then once you've "transparently proxied" the HTTP requests you want to block, you have to somehow send those packets merrily on their way... except the route for that IP address leads back to you. So, you have to tunnel it out to an system beyond your routing domain. Which means you'll need to NAT the source for anything that isn't outright proxied. That's more money.
And then you have to very strongly log the proxied and/or NATed packets because any abuse is going to lead back to your filter machine instead of back to the customer and when the policia come knocking, by God they're going to want to know who did it and they're not going to accept the answer that the Judge-ordered filtering obscured the activity near the site ordered filtered.
The ISPs aren't retarded here. The judge ordering an ISP to filter on a criteria ISPs aren't equipped to filter on is the retarded one.
Because the character array runs on past the official end of the structure, this is a very advanced topic in C. Hence much more complex that a simple character array.
It probably wasn't about the bytes. The factors are:
1. Complexity. Without exception, every variable in C is an integer, a pointer or a struct. A null terminated string is a pointer to a series of integers -- barely one step more complex than a single integer. To keep the string length, you'd have to employ a struct. That or you'd have to create a magic type for strings that's on the same level as integers, pointers and structs. And you don't want to use a magic type because then you can't edit it as an array. Simplicity was important in C -- keep it close to the metal.
2. Computational efficiency. Many if not most operations on strings don't need to know how long they are. So why suffer the overhead of keeping track? That makes string operations on null terminated strings on average faster than string operations on a string bounded by an integer.
3. Bytes. It's only one extra byte with a magic type or an advanced topic struct. In both cases with an assumption that the maximum possible length on which the standard string functions will work is 64kb. If you're talking about a more mundane struct then you're talking about an int and a pointer to a block of memory which has an extra set of malloc overhead. That's a lot of extra bytes, not just one.
For the kind of language C aimed to be -- a replacement for assembly language -- the choice of null terminated strings was both obvious and correct.
A 6-disk raid-5 has at least 12 times the activity on the memory bus and at least 6 times the activity on the card bus as the same 6-disk hardware raid 5: one pass computing the parity bits and a second pass pushing the block out to all the disks. That's assuming you're reading sectors linearly. If you peck a sector here and a sector there your software raid 5 pulls the entire block into main memory and change it.
So no, it isn't the same I/O in and out of main memory.
If you sprung for hardware raid without the battery/capacitor needed for the write back cache then you're a fool.
With the write back cache, the instant return for data committed to the cache makes the hardware raid insanely faster than software raid for most disk loads.
With a write-through cache (no battery) a software raid 1 is usually faster. You have to have a pretty crappy card before a raid 5 or raid 6 is faster on the main CPU than on the dedicated card. The extra computing on the main CPU and the extra I/O overhead to multiple devices instead of one takes its toll.
Burnable CDs and DVDs depend on your brand. I have collected thousands of Verbatims over the last decade and have lost data on very few. The discs in my software kit from 1998 all still work. On the other hand, I bought some Windatas because I wanted a few 9gb discs and was too cheap to cough up the cash. About a 80% failure rate on those after 1 year... not total but portions of the disc are irretrievable.
Nevertheless, hard disks are the way to go. Much lower hassle factor and the price per gigabyte is approaching parity with burnable DVDs.
I've used a dozen different brands and models of hardware raid cards. The only ones that haven't lost my data are the defunct Mylex DAC960 and HP's Smartarray controllers.
Had a problem with the mail server recently where a malfunctioning drive kept killing the scsi bus, causing the Smartarray controller to see every drive except that one as faulty. Once I figured out the bad drive, recovered without losing the data. Had similar problems using Adaptec and LSI cards a few years ago. Data gone. Maybe the current generation of Adaptec and LSI cards is better, but the ones I trust are the HP Smararray.
Linux software raid is also superb for reliability if you can tolerate the performance of software raid. Which for backups you generally can.
And for the record, an Adaptec 2940 is not a raid controller. It's a plain jane SCSI card -- something which unlike RAID controllers Adaptec did very well.
For the OP - go with hard disks. Use different brands and models and always use more than one. Some production runs are defective; I had a 10 disk raid 5 once where 6 went bad, all from the same production run. Killed me because they went bad faster than I replaced them in the raid 5. Putting copies of the file on at least two different brands and models will save you from that. Use different sizes too... if you put one copy on a 2TB seagate drive, put the other copy on a 1TB WD drive.
Keep some at home, some at work, some in the bank safe deposit box.
They're either charging you tax at origin (which they're allowed but not required to do -- they can treat it as a local sale and then ship the merchandise from the completed sale to you as opposed to treating it as an interstate sale) or they're making a best guess (which when they guess wrong is unlawful and leaves them liable both to you and the locality in which they allege to have performed the sale).
Remember: just because you can find a store doing it doesn't mean it's lawful.
For example, an Exxon station down the street from me charges a different price for cash and credit card sales. That hasn't been lawful in a long time but they do it anyway.
Exactly. Taxes on Interstate commerce is a Federal issue. It *is* unfair that local retailers have to collect sales tax while folks like Amazon don't. But that won't be solved by insisting that a company comply with or even know about the tax codes of every tiny municipality they affix an address label for. Fix it at the federal level with a law and free-to-business simplified and standardized database of tax collection requirements.
GoDaddy had every intention of going public but choose not to because of how they would have had to report their earnings/recognize revenue. From what I remember they would essentially split the revenue of a domain registration out over the life of the domain registration as opposed to immediately upon payment.
The only thing I use them for is the ability to edit directly on the web site. For that they work just fine.
Caveat: I heavily tweaked the installation back in the day, changing how the FP extensions integrate with the server. They were terribly insecure as installed. In my install they're just a generic CGI script same as any other.
Slashdot Mirror
Dude, no reasonable person could read the linked document and conclude that accessing wikileaks on a personal computer and/or down time was anything other than contraindicated. You have to -want- to misunderstand in order to reach another conclusion.
"This requirement applies to accessing or downloading classified information that occurs
using [...] employees' personally owned
computers"
Work on that reading comprehension thing.
This was the warning I received:
http://www.dss.mil/documents/NISP-Contractors-Protecting-Class-Info-%20coordinated-final-7-Feb.pdf
It seemed pretty clear to me: intentionally access wikileaks, lose your clearance lose your job. Inadvertently access wikilieaks, immediately file an security report to avoid losing your clearance and job.
Wireless is a competitive market so carriers are permitted to do what they want with the presumption that if the customer doesn't like it they'll go to another vendor.
Wired broadband typically suffers from a monopoly or duopoly. There typically isn't a lot of choice available to the consumer, so there are strong restrictions on what the vendor is permitted to do.
Does your estimate include power consumption for the computers and the requisite cooling for that many computers?
IIRC, she was selling fake Cisco-branded SFP GBICS and the like, stuff that has no software of its own.
The irony is that nowadays folks legally sell the same equipment as "Cisco Compatible." She went to jail over a sticker.
You don't have to buy the entire game, You can generally find the boards that were removed from the cabinets when the cabinets were upgraded to some other game. Prices range from $20 to several hundred dollars depending on the popularity of the game.
The legal status of copying those roms on to modern hardware is in doubt. The moral status is more clear: of course you can put the board in the closet and do the minimum it takes to run the software on modern hardware.
Nobody with a brain ever said Marx was wrong about capitalism. Were he so clearly wrong, he wouldn't have inspired a quarter of the world to revolution. What they said about Marx was:
1. His alternative, communism, suffered from problems far more grave than the system it was proposed to replace. Communism's structure is fundamentally unstable -- it tends to shift into despotism. That's a double-whammy for communism because it's supposed to start with a revolution where a society is already very unstable. And that's only the beginning of its problems... Central planning for large systems rarely works in the first place and tends to devolve over time when it does. And any computer person has an intuitive understanding of the euphemism "design by committee."
2. Capitalism's worst problems could be mitigated. Things like the Sherman Antitrust Act and the SEC are responsive to the problems Marx examined. Did we get it wrong last decade? Yes, Bush screwed the pooch with a particularly ignorant version of deregulation while the world was busy living beyond its means. But unlike communism, the capitalist systems system will recover over the course of the next few years and, for a while at least, will be stronger for it.
Zero. If you opt-in then it it's marketing, not spam.
They were censored because the ISPs are fucking retarded.
You wouldn't say that had you ever worked as a network engineer at a large ISP.
First, you'd have to route IP packets for the impacted address to an internal filtering machine. What filtering machine? Well, that's a rub too... you'd have to build one and while it's possible with open source, it isn't easy or particularly cheap.
Then once you've "transparently proxied" the HTTP requests you want to block, you have to somehow send those packets merrily on their way... except the route for that IP address leads back to you. So, you have to tunnel it out to an system beyond your routing domain. Which means you'll need to NAT the source for anything that isn't outright proxied. That's more money.
And then you have to very strongly log the proxied and/or NATed packets because any abuse is going to lead back to your filter machine instead of back to the customer and when the policia come knocking, by God they're going to want to know who did it and they're not going to accept the answer that the Judge-ordered filtering obscured the activity near the site ordered filtered.
The ISPs aren't retarded here. The judge ordering an ISP to filter on a criteria ISPs aren't equipped to filter on is the retarded one.
Before C, OS kernels were laboriously written in assembly language. The kernel is an excellent place for the use of C.
In C, that "special semantic" to store the length in the first bytes looks like this:
typedef struct {
unsigned short max;
unsigned short length;
char s[1];
} string;
string *newstring(unsigned short max) {
string *s;
s=malloc(sizeof(string)+(sizeof (char)*(max-1)));
s->max=max;
s->length=0;
return s;
}
Because the character array runs on past the official end of the structure, this is a very advanced topic in C. Hence much more complex that a simple character array.
It probably wasn't about the bytes. The factors are:
1. Complexity. Without exception, every variable in C is an integer, a pointer or a struct. A null terminated string is a pointer to a series of integers -- barely one step more complex than a single integer. To keep the string length, you'd have to employ a struct. That or you'd have to create a magic type for strings that's on the same level as integers, pointers and structs. And you don't want to use a magic type because then you can't edit it as an array. Simplicity was important in C -- keep it close to the metal.
2. Computational efficiency. Many if not most operations on strings don't need to know how long they are. So why suffer the overhead of keeping track? That makes string operations on null terminated strings on average faster than string operations on a string bounded by an integer.
3. Bytes. It's only one extra byte with a magic type or an advanced topic struct. In both cases with an assumption that the maximum possible length on which the standard string functions will work is 64kb. If you're talking about a more mundane struct then you're talking about an int and a pointer to a block of memory which has an extra set of malloc overhead. That's a lot of extra bytes, not just one.
For the kind of language C aimed to be -- a replacement for assembly language -- the choice of null terminated strings was both obvious and correct.
The source code is included. Just not the source for the source code.
A 6-disk raid-5 has at least 12 times the activity on the memory bus and at least 6 times the activity on the card bus as the same 6-disk hardware raid 5: one pass computing the parity bits and a second pass pushing the block out to all the disks. That's assuming you're reading sectors linearly. If you peck a sector here and a sector there your software raid 5 pulls the entire block into main memory and change it.
So no, it isn't the same I/O in and out of main memory.
Not really.
If you sprung for hardware raid without the battery/capacitor needed for the write back cache then you're a fool.
With the write back cache, the instant return for data committed to the cache makes the hardware raid insanely faster than software raid for most disk loads.
With a write-through cache (no battery) a software raid 1 is usually faster. You have to have a pretty crappy card before a raid 5 or raid 6 is faster on the main CPU than on the dedicated card. The extra computing on the main CPU and the extra I/O overhead to multiple devices instead of one takes its toll.
Burnable CDs and DVDs depend on your brand. I have collected thousands of Verbatims over the last decade and have lost data on very few. The discs in my software kit from 1998 all still work. On the other hand, I bought some Windatas because I wanted a few 9gb discs and was too cheap to cough up the cash. About a 80% failure rate on those after 1 year... not total but portions of the disc are irretrievable.
Nevertheless, hard disks are the way to go. Much lower hassle factor and the price per gigabyte is approaching parity with burnable DVDs.
I've used a dozen different brands and models of hardware raid cards. The only ones that haven't lost my data are the defunct Mylex DAC960 and HP's Smartarray controllers.
Had a problem with the mail server recently where a malfunctioning drive kept killing the scsi bus, causing the Smartarray controller to see every drive except that one as faulty. Once I figured out the bad drive, recovered without losing the data. Had similar problems using Adaptec and LSI cards a few years ago. Data gone. Maybe the current generation of Adaptec and LSI cards is better, but the ones I trust are the HP Smararray.
Linux software raid is also superb for reliability if you can tolerate the performance of software raid. Which for backups you generally can.
And for the record, an Adaptec 2940 is not a raid controller. It's a plain jane SCSI card -- something which unlike RAID controllers Adaptec did very well.
For the OP - go with hard disks. Use different brands and models and always use more than one. Some production runs are defective; I had a 10 disk raid 5 once where 6 went bad, all from the same production run. Killed me because they went bad faster than I replaced them in the raid 5. Putting copies of the file on at least two different brands and models will save you from that. Use different sizes too... if you put one copy on a 2TB seagate drive, put the other copy on a 1TB WD drive.
Keep some at home, some at work, some in the bank safe deposit box.
Sure. But that takes FEDERAL action.
They're either charging you tax at origin (which they're allowed but not required to do -- they can treat it as a local sale and then ship the merchandise from the completed sale to you as opposed to treating it as an interstate sale) or they're making a best guess (which when they guess wrong is unlawful and leaves them liable both to you and the locality in which they allege to have performed the sale).
Remember: just because you can find a store doing it doesn't mean it's lawful.
For example, an Exxon station down the street from me charges a different price for cash and credit card sales. That hasn't been lawful in a long time but they do it anyway.
Exactly. Taxes on Interstate commerce is a Federal issue. It *is* unfair that local retailers have to collect sales tax while folks like Amazon don't. But that won't be solved by insisting that a company comply with or even know about the tax codes of every tiny municipality they affix an address label for. Fix it at the federal level with a law and free-to-business simplified and standardized database of tax collection requirements.
GoDaddy had every intention of going public but choose not to because of how they would have had to report their earnings/recognize revenue. From what I remember they would essentially split the revenue of a domain registration out over the life of the domain registration as opposed to immediately upon payment.
Yeah, that's how the GAAP says you do it. http://en.wikipedia.org/wiki/Generally_Accepted_Accounting_Principles
That's how you avoid a pyramid scheme where the finances fall apart when there's no longer enough new revenue to fund existing service commitments.
Not so much in the US.
In the US it's illegal for a for-profit company to accept donations and it's illegal to pay someone less than minimum wage.
So, the unpaid internships (at least the legal ones) are only in the non-profit sector.
The only thing I use them for is the ability to edit directly on the web site. For that they work just fine.
Caveat: I heavily tweaked the installation back in the day, changing how the FP extensions integrate with the server. They were terribly insecure as installed. In my install they're just a generic CGI script same as any other.