I think that much of your pain is due to Debian, which didn't have a real concept of "an OS release", but encouraged haphazard melding of stable and untested components. That's asking for death on servers that need to work.
If you get the chance, take a look at one of the more "packaged" Linux releases, such as CentOS if you have no money to buy with. (It's a free rebundling of RedHat Enterprise Linux, and kept well organized and up-to-date to reduce the necessary updates after installation.)
En masse is sometimes the only way to do it. You go from supporting a dozen different ways of doing things, to supporting one *CORE* technology and a few exceptions, keeping your resources focused on making that core work well and migrating the exceptions ASAP. Anyone who's worked with more than a dozen systems at a time knows how painful dangling OS versions and deployments can be.
We need the big success stories to encourage hardware vendors and major software vendors to take us seriously. Unless we see Photoshop for Linux, for example, there are entire markets that are closed to the OS.
They gave up on breaking Lotus via DOS years ago. They still break it wherever possible via MS Office "features". And they most certainly have a habit of "embrace and extend" that involves breaking other people's software, or even simply breaking their own software's ability to operate with the published standards, such as the Java and Kerberos lawsuits established.
This is not a nice company. Their fraudulent and criminal behavior is well-established in court, year after year.
That is exactly right. Remember, "Office ain't done 'til Lotus won't run".
Microsoft likes to patent XML technologies. By incorporating XML into the filesystem itself, they can patent-encumber their filesystems and prevent Linux file systems and Samba-like network file systems from other operating systems to ever interoperate with it.
Don't rely on this punk to tell the truth about why he wrote a worm. If he can sell out some spamware salesman to lessen his own jail time in a Moroccan jail, you can bet that he'll bargain to keep his precious 18-year-old ass out of a jail cell with some very, very protective cellmate doing hard time who will consider him very, very precious.
No, cat and ls are pretty stable and not that fancy. I don't consider the feature additions in ls over the last 5 years to be an improvement, nor to put ls anywhere ahead of commercial tools like "dir". There's just not enough material there to improve.
HURD has never worked: it took too much trouble to set up, and never worked on enough hardware for enough time without crashing to even want to use it, despite it's theoretical benefits.
I urge you to look at what open source advocates and authors actually get paid for. It's quite capitalist, and profitable, and doesn't steal a darned thing. Also, far more new ideas and development are coming out of the open source world, on a programmer by programmer basis, then ever came out of the corporate software world.
The fiscal incentives do CHANGE and are displaced, from middle management's ability to seal the box and not have their clients able to use any other product and thus the growth of monopolistic and anti-competitive, even non-capitalist companies, and allowing a much smaller start-up cost in buying the software licenses to do development. The money instead goes in-house to local developers, and far more smaller opportunities for local variation is created.
It's fun, it's profitable, and I'm certainly making a living at working with tools at least 5 years ahead of where they'd be without such open source tools.
And the "run as other other user" trick works fine, until it doesn't. Software installers are notoriously bad about this, I've seen several in the last year that would crash badly when run that way. Game installers in particular are bad. And then there's scanners, especially old ones, which simply refuse to operate correctly unless you have Administrator privileges, and which are out of date enough that the vendor has no updates for it anymore.
With this layout, there's only one approach for ODSL that works. Do the testing and development honestly, and when Microsoft tries to lie or skew them, publish the details with references of exactly what Microsoft tried to screw up and how.
It will tarnish ODSL somewhat, but much less than refusing the implicit challenge of Microsoft's attempts to enter it. They've also got a great chance to scare Microsoft off, publicly, by insisting on a GPL rather than BSD or other license that would allow Microsoft to embrace, extend, steal, and deliberately break technologies to make incompatible with other people's software, such as they tried with Kerberos and Java.
That's absolutely correct. Take a look at SPF over at http://spf.pobox.com/ and how their embrace and extend approach was used to fracture an approach that would lighten the burden of spam for mail servers worldwide and break their SenderID model of selling keys to authorize users. Then notice that AOL has thrown out SenderID, but held up full implementation of SPF for nearly a year while the development and legal arguments were going on.
3: Take the results, skew them wherever possible, lie about them, and take sole credit for anything even vaguely positive that comes out of it.
Go take a look at Microsoft's attempts to enforce their SenderID "buy a Microsoft stamp on all your email" licensing scheme on top of the SPF anti-forgery email scheme, at http://spf.pobox.com/ for an example of how they not only manipulate development for their own ends but of how now they're taking credit for SPF, and how they actually prevented SPF from getting RFC's published so it can have standards implemented into new mail servers.
Implementing SPF widely would entirely block most modern email worms, and could reduce phishing schemes by a huge factor. But no, Microsoft had to "embrace and extend" its development and push most developers right out of it by supergluing their own already-proven-useless project onto it.
The absolute garbage involved in managing them, however, is nasty. I've seen plenty of sites where every single user runs with "Administrator" privileges in Windows every single day, because running or installing simple software like MP3 players or CD burners requires it.
Microsoft may have a very sophisticated user permission management system, one almost as fine grained as Kerberos, AFS, and NIS in the UNIX and Linux worlds offered 10 years ago. But way too much Windows software just ignores it.
On top of that, even in the Linux and UNIX world, you may notice how little people actually use the more subtle features and rely on the old "you're a guest with no privileges, you're a distinct authorized user, or you are god" levels of authority.
It's worse. The clause that a lot of bulk mail includes about how it is legal under the CAN-SPAM act, and how you can opt out of it and see how legal their spam is, is one of the strongest indicators that the message is actually spam that I have seen in years.
I have not seen a single non-spam message that discussed its compliance with CAN-SPAM. In fact, out of the dozen or so I've seen this week, every single one that talked about CAN-SPAM was in fact in violation of CAN-SPAM because they were actually using CAN-SPAM unsubscribe messages to record new mail addresses for other spam, but usually from another address. (I add a few throwaway accounts every month for precisely such tests of spam behavior.)
Worse. They're concerned that it will cut into their profits on selling spam-filters, such as their patented and amazingly stupid SenderID concept, and that it will interfere with the bulk mailing list management tools they sell tightly integrated for use in Microsoft Outlook.
Couple that with their need for your name and personal details with every product registration, and the default settings of those forms to permit them to advertise at you, and we're seeing a company geared up for bulk marketing under the excuse of "customer notices".
Start with http://www.thecanadapage.org/Machiavelli.htm, and work your way out. There are plenty of historical cases of real election fraud in Canada: perhaps not as common as in the US, which is to their credit, but believing that it never happens is silly to the point of being laughable.
Then I suggest you learn to look at what other federal offices do, in particular the Carnivore email monitoring (described at http://news.zdnet.com/2100-9595_22-522071.html?leg acy=zdnn). And despite public claims to the contrary, the system is still alive and in use.
State laws and local city or county policies in the US can get pretty weird. There's usually a good right to appeal, and have what's called a "competency hearing". I've seen some seriously paranoid senior citizens challenge their status and lose, calling the court corrupt, even though it seemed blatantly obvious as a distant family member that they were in fact incompetent to manage their lives. (One would forget to eat for days, another couldn't remember their dead spouse's name.)
It's a good case. But I urge you to take a look at the US TeleCommunications Privacy Act. That piece of nastiness tried to forbid telephone companies from implementing new technologies without, integrated into them, the means to do undetectable wiretapping. That part of the act was dropped when the phone companies balked at the expense implied by its provisions, without federal funding to provide network to the home, which is what they were willing to trade for that provision.
Similarly, the "Clipper Chip" encryption designed by the NSA years ago for use in cell phone and in data communications was to have its encryption keys stored in federal hands, for release "when required for law enforcement". No warrant, no public record of their release for review b by the courts, the pair of keys to decrypt such communications could have been kept in the same computer for easy theft and duplication by other federal offices for all we knew. What finally killed that chip was when it turned out they violated 3 patents due to developing it in secret, and when it turned out their "Law Enforcement Agency Field" checksum was so short you could generate a fake key in about 40 minutes and use that instead.
Using strong encryption for email, however, presents a legal problem. It's still illegal in the US, the main software exporting country especially with Microsoft here, to ship encryption to certain countries (such as Cuba), and illegal in certain countries to use encryption this way (France and mainland China leap to mind). So you can't easily make well encrypted email universal.
Just like in Florida in the US, right?
In a close election, even a little bit of electrion fraud is both easy to do and has tremendous leverage. In the length of Canada's history, I'm sure it's happened at least once. Like, say, in Newfoundland in the early 80's?
The courts are a different department than the police. They're supposed to be somewhat opposed to each other in their goals, so that a crooked judge or a bad one is turned in by the police, and so that a bad cop who lies on the stand gets convicted by the judge. It's certainly true that judges aren't perfect: take a look at the rulings by Judge White in the Microsoft anti-trust case for examples of a judge convinced by flashy lawyers to condone criminal activity, even within his own courtroom.
But it's too easy for a cop to get committed to convicting one particular person and, without any oversite, use warrantless searches to arrest them for unrelated offenses or even fraudulent ones. The judge reviewing the case early helps keep the police focused on the real cases, and it provides a paper trail for the investigation that's vital for a defense attorney to verify the evidence, and for other attorneys to make sure the police aren't randomly searching for cute girls' and boys' phone numbers, insider stock information, dirt on the mayor's political enemies, etc.
If you don't think such abuses occur, I strongly urge you to work as a secretary in a social services department for a summer. A good one stops the abuses almost before they occur, but an overworked and bureaucratical fiefdom can permit all sorts of abuses.
Using a shuttle on autopilot is a complete waste of money. The very heavy and expensive equipment to support human life would be a complete waste of lift capacity and fuel, since it's very inefficiently using a considerably lower thrust than an unmanned craft can use.
Instead, in the short term, use Titans for basic equipment launches. The technology is established, it's robust, and it's a lot cheaper and uses less fuel per pound to get to orbit. Don't waste shuttle launches on raw shipping, use it only for material that can't stand high accelerations.
Re:Gotta use it right
on
Ending Spam
·
· Score: 2, Informative
No, SenderID tags have to be purchased from Microsoft, and can only be parsed by mail software from Microsoft due to the encumbering XML patents it uses. Take a look at the patent issues surrounding the RFC's for SPF, which Microsoft tried to "embrace and extend" into patented and proprietary uselessness.
The current result is that the SenderID keys are not purchased by spammers: they're usually stolen by using the SenderID key's machine as a spam zombie, and it serve the admins of Microsoft mail servers right for believing in such a stupid approach.
Slashdot Mirror
For Linux, UNIX, and the CygWin users out there, there are also the "zip" and "unzip" utilities, which work quite well from the command line.
A nice, friendly built-in GUI is worth a few bucks, but that's all you really get for your few bucks.
I think that much of your pain is due to Debian, which didn't have a real concept of "an OS release", but encouraged haphazard melding of stable and untested components. That's asking for death on servers that need to work. If you get the chance, take a look at one of the more "packaged" Linux releases, such as CentOS if you have no money to buy with. (It's a free rebundling of RedHat Enterprise Linux, and kept well organized and up-to-date to reduce the necessary updates after installation.)
En masse is sometimes the only way to do it. You go from supporting a dozen different ways of doing things, to supporting one *CORE* technology and a few exceptions, keeping your resources focused on making that core work well and migrating the exceptions ASAP. Anyone who's worked with more than a dozen systems at a time knows how painful dangling OS versions and deployments can be.
We need the big success stories to encourage hardware vendors and major software vendors to take us seriously. Unless we see Photoshop for Linux, for example, there are entire markets that are closed to the OS.
They gave up on breaking Lotus via DOS years ago. They still break it wherever possible via MS Office "features". And they most certainly have a habit of "embrace and extend" that involves breaking other people's software, or even simply breaking their own software's ability to operate with the published standards, such as the Java and Kerberos lawsuits established.
This is not a nice company. Their fraudulent and criminal behavior is well-established in court, year after year.
That is exactly right. Remember, "Office ain't done 'til Lotus won't run". Microsoft likes to patent XML technologies. By incorporating XML into the filesystem itself, they can patent-encumber their filesystems and prevent Linux file systems and Samba-like network file systems from other operating systems to ever interoperate with it.
Don't rely on this punk to tell the truth about why he wrote a worm. If he can sell out some spamware salesman to lessen his own jail time in a Moroccan jail, you can bet that he'll bargain to keep his precious 18-year-old ass out of a jail cell with some very, very protective cellmate doing hard time who will consider him very, very precious.
No, cat and ls are pretty stable and not that fancy. I don't consider the feature additions in ls over the last 5 years to be an improvement, nor to put ls anywhere ahead of commercial tools like "dir". There's just not enough material there to improve.
Make. Sed. Less. Grep. Bash. Apache. Firefox. Kerberos was 5 years ahead at its release.
Well, OK, Firefox is only 2 years ahead.
HURD has never worked: it took too much trouble to set up, and never worked on enough hardware for enough time without crashing to even want to use it, despite it's theoretical benefits.
I urge you to look at what open source advocates and authors actually get paid for. It's quite capitalist, and profitable, and doesn't steal a darned thing. Also, far more new ideas and development are coming out of the open source world, on a programmer by programmer basis, then ever came out of the corporate software world.
The fiscal incentives do CHANGE and are displaced, from middle management's ability to seal the box and not have their clients able to use any other product and thus the growth of monopolistic and anti-competitive, even non-capitalist companies, and allowing a much smaller start-up cost in buying the software licenses to do development. The money instead goes in-house to local developers, and far more smaller opportunities for local variation is created.
It's fun, it's profitable, and I'm certainly making a living at working with tools at least 5 years ahead of where they'd be without such open source tools.
And the "run as other other user" trick works fine, until it doesn't. Software installers are notoriously bad about this, I've seen several in the last year that would crash badly when run that way. Game installers in particular are bad. And then there's scanners, especially old ones, which simply refuse to operate correctly unless you have Administrator privileges, and which are out of date enough that the vendor has no updates for it anymore.
With this layout, there's only one approach for ODSL that works. Do the testing and development honestly, and when Microsoft tries to lie or skew them, publish the details with references of exactly what Microsoft tried to screw up and how. It will tarnish ODSL somewhat, but much less than refusing the implicit challenge of Microsoft's attempts to enter it. They've also got a great chance to scare Microsoft off, publicly, by insisting on a GPL rather than BSD or other license that would allow Microsoft to embrace, extend, steal, and deliberately break technologies to make incompatible with other people's software, such as they tried with Kerberos and Java.
That's absolutely correct. Take a look at SPF over at http://spf.pobox.com/ and how their embrace and extend approach was used to fracture an approach that would lighten the burden of spam for mail servers worldwide and break their SenderID model of selling keys to authorize users. Then notice that AOL has thrown out SenderID, but held up full implementation of SPF for nearly a year while the development and legal arguments were going on.
Or perhaps:
3: Take the results, skew them wherever possible, lie about them, and take sole credit for anything even vaguely positive that comes out of it.
Go take a look at Microsoft's attempts to enforce their SenderID "buy a Microsoft stamp on all your email" licensing scheme on top of the SPF anti-forgery email scheme, at http://spf.pobox.com/ for an example of how they not only manipulate development for their own ends but of how now they're taking credit for SPF, and how they actually prevented SPF from getting RFC's published so it can have standards implemented into new mail servers.
Implementing SPF widely would entirely block most modern email worms, and could reduce phishing schemes by a huge factor. But no, Microsoft had to "embrace and extend" its development and push most developers right out of it by supergluing their own already-proven-useless project onto it.
The absolute garbage involved in managing them, however, is nasty. I've seen plenty of sites where every single user runs with "Administrator" privileges in Windows every single day, because running or installing simple software like MP3 players or CD burners requires it. Microsoft may have a very sophisticated user permission management system, one almost as fine grained as Kerberos, AFS, and NIS in the UNIX and Linux worlds offered 10 years ago. But way too much Windows software just ignores it. On top of that, even in the Linux and UNIX world, you may notice how little people actually use the more subtle features and rely on the old "you're a guest with no privileges, you're a distinct authorized user, or you are god" levels of authority.
It's worse. The clause that a lot of bulk mail includes about how it is legal under the CAN-SPAM act, and how you can opt out of it and see how legal their spam is, is one of the strongest indicators that the message is actually spam that I have seen in years. I have not seen a single non-spam message that discussed its compliance with CAN-SPAM. In fact, out of the dozen or so I've seen this week, every single one that talked about CAN-SPAM was in fact in violation of CAN-SPAM because they were actually using CAN-SPAM unsubscribe messages to record new mail addresses for other spam, but usually from another address. (I add a few throwaway accounts every month for precisely such tests of spam behavior.)
Worse. They're concerned that it will cut into their profits on selling spam-filters, such as their patented and amazingly stupid SenderID concept, and that it will interfere with the bulk mailing list management tools they sell tightly integrated for use in Microsoft Outlook.
Couple that with their need for your name and personal details with every product registration, and the default settings of those forms to permit them to advertise at you, and we're seeing a company geared up for bulk marketing under the excuse of "customer notices".
Start with http://www.thecanadapage.org/Machiavelli.htm, and work your way out. There are plenty of historical cases of real election fraud in Canada: perhaps not as common as in the US, which is to their credit, but believing that it never happens is silly to the point of being laughable.
Then I suggest you learn to look at what other federal offices do, in particular the Carnivore email monitoring (described at http://news.zdnet.com/2100-9595_22-522071.html?leg acy=zdnn). And despite public claims to the contrary, the system is still alive and in use.
State laws and local city or county policies in the US can get pretty weird. There's usually a good right to appeal, and have what's called a "competency hearing". I've seen some seriously paranoid senior citizens challenge their status and lose, calling the court corrupt, even though it seemed blatantly obvious as a distant family member that they were in fact incompetent to manage their lives. (One would forget to eat for days, another couldn't remember their dead spouse's name.)
It's a good case. But I urge you to take a look at the US TeleCommunications Privacy Act. That piece of nastiness tried to forbid telephone companies from implementing new technologies without, integrated into them, the means to do undetectable wiretapping. That part of the act was dropped when the phone companies balked at the expense implied by its provisions, without federal funding to provide network to the home, which is what they were willing to trade for that provision. Similarly, the "Clipper Chip" encryption designed by the NSA years ago for use in cell phone and in data communications was to have its encryption keys stored in federal hands, for release "when required for law enforcement". No warrant, no public record of their release for review b by the courts, the pair of keys to decrypt such communications could have been kept in the same computer for easy theft and duplication by other federal offices for all we knew. What finally killed that chip was when it turned out they violated 3 patents due to developing it in secret, and when it turned out their "Law Enforcement Agency Field" checksum was so short you could generate a fake key in about 40 minutes and use that instead. Using strong encryption for email, however, presents a legal problem. It's still illegal in the US, the main software exporting country especially with Microsoft here, to ship encryption to certain countries (such as Cuba), and illegal in certain countries to use encryption this way (France and mainland China leap to mind). So you can't easily make well encrypted email universal.
Just like in Florida in the US, right? In a close election, even a little bit of electrion fraud is both easy to do and has tremendous leverage. In the length of Canada's history, I'm sure it's happened at least once. Like, say, in Newfoundland in the early 80's?
The courts are a different department than the police. They're supposed to be somewhat opposed to each other in their goals, so that a crooked judge or a bad one is turned in by the police, and so that a bad cop who lies on the stand gets convicted by the judge. It's certainly true that judges aren't perfect: take a look at the rulings by Judge White in the Microsoft anti-trust case for examples of a judge convinced by flashy lawyers to condone criminal activity, even within his own courtroom.
But it's too easy for a cop to get committed to convicting one particular person and, without any oversite, use warrantless searches to arrest them for unrelated offenses or even fraudulent ones. The judge reviewing the case early helps keep the police focused on the real cases, and it provides a paper trail for the investigation that's vital for a defense attorney to verify the evidence, and for other attorneys to make sure the police aren't randomly searching for cute girls' and boys' phone numbers, insider stock information, dirt on the mayor's political enemies, etc.
If you don't think such abuses occur, I strongly urge you to work as a secretary in a social services department for a summer. A good one stops the abuses almost before they occur, but an overworked and bureaucratical fiefdom can permit all sorts of abuses.
Using a shuttle on autopilot is a complete waste of money. The very heavy and expensive equipment to support human life would be a complete waste of lift capacity and fuel, since it's very inefficiently using a considerably lower thrust than an unmanned craft can use.
Instead, in the short term, use Titans for basic equipment launches. The technology is established, it's robust, and it's a lot cheaper and uses less fuel per pound to get to orbit. Don't waste shuttle launches on raw shipping, use it only for material that can't stand high accelerations.
No, SenderID tags have to be purchased from Microsoft, and can only be parsed by mail software from Microsoft due to the encumbering XML patents it uses. Take a look at the patent issues surrounding the RFC's for SPF, which Microsoft tried to "embrace and extend" into patented and proprietary uselessness. The current result is that the SenderID keys are not purchased by spammers: they're usually stolen by using the SenderID key's machine as a spam zombie, and it serve the admins of Microsoft mail servers right for believing in such a stupid approach.