I think you're merging together two different rides. Splash mountain is a flume ride, with conveyor lifts and tall drops. There is no guest or cast member power. There is also a canoe ride in the Rivers of America, which works as you described, minus the flume and elevation change. The biggest drop in Splash Mountain has a runout adjacent to the Rivers of America, so I can see how the two would blur together.
My only point was that the foreign vendors aren't particularly concerned about Microsoft's pocketbook, and will respond to the dictates of the consumer of their products...
Microsoft and Intel decide the direction of the PC. The manufacturers must follow or die. There can be differentiation within the field established by Wintel, such as SFF PC's, but a new mainboard must run the newest version of Windows. I roughly agree with you about the trend of IT departments and the end of the truly Personal Computer within modern organizations. In fairness, a lot of what was done on an early PC with BASIC can be done within Excel. Even working within the narrowest confines of a locked-down machine, users can achieve substantial automation of repetitive work.
The market has decided (after repeated waves of Windows worms successfully invading millions of machines, my firewall logs still show several thousand attempted propagations every day) that security is officially a "big deal." Therefore, Microsoft would like processor and chipset makers to come in, after the fact, and make their software more "secure" rather than changing their coding practices.
No! That's emphatically not what Microsoft is doing. In fact, if they continue their shoddy coding practices, this hardware-based security won't be worth much. Palladium (which this seems to part of) allows an application to encrypt, decrypt and "attest" via an API to the hardware that bypasses the OS. So while it's true that OS bugs should be removed from the critical path, application bugs remain in the critical path. If Phoenix's BIOS is providing TCPA-like authentication of the bootloader, that only proves the authorized bootloader takes control. Any flaws in the bootloader or subsequent OS could still be exploited. So this new BIOS won't protect a shaky OS or application.
Microsoft is about one thing, and one thing only, and anything they say is "good for the user" should be scrutinized very, very carefully. You can call Slashdotters paranoid if you like, but we're at least looking at the issues. The rest of the media appears willing to simply follow the party line.
If I try to inject a note of reality into a slashdot discussion, apparently it looks like I'm "taking the side of" the relevant Bad Guy. Paranoia does not help one to look at the issues - it blinds one with imaginary issues. If you were hunting a huge animal with only a.22 rifle, it would be wise to learn everything you can about the anatomy and habits of that animal. But paranoid people might see the animal behind every bush, and prefer to trade talll tales about how huge and evil the animal is. Then if these paranoid hunters tried to enlist the aid of nearby villagers, they would properly be laughed at, because they told obviously false tales about the animal. But these stories (the 10 inch teeth, the appetite for babies) would sound true to the paranoid hunters because they'd repeated and embellished them for so long.
An excess of cynicism is exactly the same as an excess of naivete. Both render one incapable of decisive action.
And your argument about DRM is specious. There is already a lot of entertainment sold, online and otherwise. Billions of dollars of it.
Whether or not that's true, the majority of mainstream entertainment is not legitimately available online. Apple Music Store is a huge exception - it remains to be seen how it will fare once aac's start showing up on p2p. On the whole, the entertainment industry is waiting for stronger protections before throwing their content on the net.
I have no interest whatsoever in paying more money for a computer that will do even less for me in order to guarantee a group of corporate thugs a neverending revenue stream.
First, do you really see the cost of PC's going up? Maybe a brief spike until this technology is compl
I'm not too happy with 2.0. I used it because it came packaged with an OS and just for once, in the interest of speed I relaxed my rule of always building Apache from source (which would have been 1.x, of course) and accepted the package. I had several weird problems with 2.x on 2 different machines. Random errors in mod_perl, which I spent a lot of time trying to nail down until I switched to 1.3.x, solving it. And the part that really pisses me off - Apache 2.x inserts a bunch of garbage into each line in the error log, and the docs say you can't turn this off. That was the last straw for me. I sincerely hope 1.3.x is maintained forever.
In all my experience with 1.3.x, I never felt the anger and frustration that commercial software can cause, but 2.x did it with that brain-dead misfeature.
Slashdot has this really stupid, shallow way of covering kernels.
Or anything, really. Let's see:
Patents: Grab a random patent and greatly overstate its scope, so it sound like you can hardly breathe without infringing it. Enjoy sheeplike chorus of whining, which drowns out anyone pointing out that the patent is much narrower than stated.
Nifty device/invention: Take something that's neither the newest nor the best in its field, nor particularly interesting, and: "Drool. I want one. But where do I get 5 megawatts and 16 tons of cooling to run it?"
Scary legislation story: Take a new proposed/passed law. Don't read it. Use rumours about the law to find the most appropriate section of 1984 and spin paranoid theories from there. Don't try to understand what problem the legislature was trying to solve, or how it could be better solved.
Worst of all: the professional troll/journalist/analyst who throws some poop at Linux/Open Source in the hope of getting page hits or notoriety. Slashdot cooperates grandly, with a comment like, "Joe Schmoe says Linux is obsolete. So what do you guys think?"
Obviously I like something about it, or I'd have left by now.
I don't really see why they would give a rat's ass about "trusted computing" unless the marketplace demands it.
Who do you think the marketplace is? A bunch of disgrunted hobbyists? No, the serious market is system integrators - companies, large and small that assemble computers. If they are selling to corporate customers, these system integrators may want to deliver computers that can't be tampered with by users. Many corporate sysadmins might welcome additional weapons to fight against viruses, pirated software, etc. Unlike the slashdot crowd, they won't be viewing this technology through paranoid eyes, but rather asking, "What can it do for me?" And they'll see a lot of potential. A lot of help in keeping PCs in a known, trusted state, rather than corrupted by user actions.
The other major market is retail PCs. If a strong DRM solution becomes widely used, it will enable lots of entertainment content to be sold online. Everyone (except slashdot) knows this, so everyone is scrambling like mad to become that solution. So if this system is called "HappyPuppy" for example, consumers shopping for a new PC will make sure it has HappyPuppy because that lets them download their favorite songs cheaply. No retailer will buy any more PCs without HappyPuppy because they wouldn't sell.
To a normal person, HappyPuppy is an additional capability, like having a DVD drive. It is not a restriction. It doesn't stop him from doing anything he could do before. Contrary to slashdot mythology, it doesn't stop him from downloading, using or sharing illegal mp3s. Of course, there is no way to extract the HappyPuppy content into something like mp3s, but there never was.
Actually, if the (h|cr|att)acker were clever, he could use this scenario to camouflage a theft. Design the attack so half the accounts are net gainers and half are net losers. Imagine a bell curve centered around 0. Maybe 3% of accounts gained or lost more than $10,000. The (h|cr|att)acker's accounts just happen to be in the upper 3%.
Still wouldn't work in the long run, for lots of reasons. The banks have backups, and when ATMs have erroneously given money to customers, the banks have chased it down.
Re:The main issue with XML is performance
on
Effective XML
·
· Score: 1
I deal with very large amounts of data (as do many programmers that read/. I'm sure)
That's probably the key issue. I wouldn't want to move large amounts of tabular information with XML. Most of the XML messages I handle are quite small, and the database processing time dwarfs the time to generate and transmit the XML. Bloating 1 KB to 3 KB doesn't matter, but bloating 100G to 300G would definitely matter.
Re:The main issue with XML is performance
on
Effective XML
·
· Score: 1
I think your view is a little simplistic. You seem to be limiting the discussion to communicating tables, which is an area where CSV can work pretty well. But what if there is hierarchical structure? An XML document could have a list of PC's, each containing a list of peripherals with attributes.
And if you don't know what they are (and no data dictionary is provided) then you don't need to mess with the file.
This doesn't match my experience. I maintain a complex app that receives XML requests and returns XML responses. The ability of a human to quickly read the data is invaluable. Also, as messages evolve, there is no chance of a data item being misinterpreted due to being in the wrong column.
My XML is simple and readable, not SOAP or anything like that.
As time goes on, optional tags get added to requests to cause special behavior. Only the programmer who needs that behavior needs to care about the existence of the new tag. Likewise, more info gets added to responses, but the new info can be ignored by existing clients.
CSV has its place - importing and exporting table-formatted data. But for cross-platform messaging, XML is the best right now.
Can I claim all RIAA songs are public domain, and sell bootleg CDs that way?
No. But if there is one song on an RIAA album which you claim to own, you can sue the record company. During the lawsuit, both you and the record company could continue to distribute the song unless the judge issues an injunction barring one of the parties from doing so. Once the case is decided, the infringing party could be liable for damages. But if you lose the case, I don't think you could be prosecuted criminally. That requires willful infringement, but since you were actively asserting your purported rights via the lawsuit, it looks like you acted in good faith.
Law enforcement does not usually wait for civil cases to end before charging someone with a crime; on the contrary, it's usually the other way around.
Can you give me some examples? Specifically, when have they charged the plaintiff in a lawsuit with a crime that is factually equivalent to losing the lawsuit? If neighbors Smith and Jones are embroiled in a lawsuit over a road between their properties, do the authorities ever charge Smith or Jones with trespassing for using that road?
If they can track down few bootleg ISOs of Adobe's software and prosecute the "pirate" they should definitely investigate SCO's criminal copyright infringement on behalf of hundreds (if not thousands) of organizations and individuals.
I wonder how such an investigation actually begins. I'm guessing that it begins with a complaint from Adobe, and that a certain amount of expertise might be needed to make the complaint compelling. There are a lot of former prosecutors in private practice. It would probably cost about $200 to get some clues on the complaint process, or to be told why it won't work.
Well, the situation being so public and so persistent, it doesn't take a magician to figure that out. Authors filing complaints would be nice but is not required.
I don't agree at all. If someone is publicly reputed to have murdered his wife, and the wife is missing, maybe the authorities would take the initiative. Although even there, the story usually starts with someone phoning the police to report the missing person. In this case, a casual read of any SCO story in the papers does not reveal any criminal actions. I think you are too close to this story to see it objectively. There are many lawsuits in the courts, and almost all of them cause great hatred. Almost everyone embroiled in a lawsuit would like to see his adversaries thrown in prison. You may be able to construct a valid theory for prosecution of SCO, but it's a reach. It should be constructed by a criminal attorney and patiently explained to the relevant authorities. I tend to think that if it can be done, IBM's lawyers will do it.
I share your evident desire to see McBride severely punished for his actions. And don't forget co-conspirators Ralph Yarrow and Chris Sontag. I just don't see any realistic hope for it, as the legal system has not adapted yet to handle this kind of abuse. These people are criminals in all but name, but since their chosen weapon is the lawsuit rather than the gun, they are probably immune to prosecution. I think they will lose their suit, but land on their feet with enough money for a modest retirement.
Anyone running any operating system can be attacked and comprimized.
This is the kind of statement some geeks love to make. While it's literally true, it's nearly meaningless. Let's try a few variations:
Q:Should I bring my umbrella on Thursday? A:Rain could occur on any day.
Q:Which car is more reliable, a Toyota Corolla or a BMW 325? A:Anyone driving any car can experience a mechanical breakdown.
Q:Should I drink this six month old milk in the fridge or buy some new milk? A:Anyone drinking any milk can die of food poisoning.
Microsoft's approach to security is deeply flawed. Again and again, they made visibly wrong decisions which any experienced network programmer could see as wrong. Their permissions system sucks, for example. They thought they would be clever and leapfrog Unix - they would go from no security to fancy ACL's. Unfortunately, almost none of their customer base can figure out the fancy ACL's, and most of their ISV's are not cooperating. Unix ugo perms are already at the outer limit of what most people can understand. On a typical Unix box, the majority of files have correct permissions, and a minority have a non-disastrous error in perms. On a typical Windows box, either everything is wide open, or everything is locked down so Administrator intervention is needed for almost anything.
Both Windows and Linux contain flaws in execution. But Windows contains severe flaws in design.
You people sometimes seem to forget that despite MS's faults, they do employ some of the best and brightest in the world. I imagine some of you may not believe that, but I do.
If you define the top x (fraction) as "the best and brightest" and Microsoft has n employees with the same brightness distribution as the general population, it follows that they employ roughly nx of the "best and brightest." If nx > 3 I think everyone would agree that Microsoft employs some of the best and brightest. E.g. 65,000 *.001 = 65; 65 > 3. Congratulations on another reliably true statement.
Of course this assertion, like the one above, is intended to connote more than it denotes. Since you raise this alleged brightness as a defense against accusations of making bad software, you appear to argue that the software isn't really bad, since bright people can't make bad software. This seems like a curiously indirect way of evaluating products that have inflicted real-world pain on many of us.
Your argument makes sense. Unfortunately, SCO seems to be saying that GPL'd code is effectively in the public domain. If so, that would also constitute their defense against copyright infringement. So it seems to make sense to evaluate that argument just once, not in two different courts.
Where I completely disagree with you is your assumption that the government operates only at the behest of special interests. Some people are so politically tied-in that they could be shielded from prosecution. Darl McBride is not one of them.
Prosecutors want to put people in prison. If they could put someone in prison with five minutes work, they would do it, whether that person is Linus Torvalds or Darl McBride. But if they have to devote huge resources to a conviction that carries a light sentence, they'd rather skip it and work on easier cases.
Have any Linux authors complained to their local US attorney's office? I think that generally the victim has to take the initiative and complain, rather than waiting for the feds to magically figure out the situation.
I have no idea if your figures are correct, but there's no way you can simply look at a number and scream "too much!" You have to understand what is bought for that number.
Sun Tzu wrote something like "It is cheaper to pay the best spies lavishly than to pay the worst army stingily." NSA earned its huge budget by producing priceless intelligence that advanced US interests and saved US lives. The history of signals intelligence shows that it is a good investment for nations.
The DOJ doesn't generally prosecute violent crime.
And they are weak at white-collar crime, not for sinister reasons but because these cases are very hard and expensive to try. A prosecutor would rather put five people in prison, and let one go, than spend all his time putting that guy in prison and let five go.
The validity of SCO's claims is still up in the air, and will be settled in court. Do you seriously expect the Justice Department to initiate a criminal prosecution based on an issue that is pending settlement in civil court? Can you think of any example where this happened? And just because SCO is constant front-page news here doesn't mean DOJ knows or cares about them.
Can you find a single criminal attorney in the US, either prosecutor or defender, who believes there is a solid criminal case against SCO or some of their employees? Remember, DOJ is run by lawyers, not geeks. They only pursue cases that they think they can win. What is your basis for thinking they could win such a case? What would the charges be exactly?
Even if SCO loses, I doubt that they will be prosecuted for copyright infringement, because they arguably acted in good faith. They distributed code which they thought they owned. Copyright infringement must be wilful to be criminal.
I agree that there's too much tinfoil-hattism here, but I think there's a big difference between a Grand Jury demanding information and an FBI Agent demanding the information. Nobody can reasonably object to a judge or Grand Jury extracting private information - without that power injustice would flourish everywhere in secrecy. We worry about the executive branch silently and easily vacuuming up our personal info.
Even though I really don't care about Ashcroft or his critics, this annoyed me enough to reply. You point out correctly that terrorism is a vanishingly small problem. Then you drag out the tired cliche of Ashcroft "trampling on our civil liberties." Any such trampling is also vanishingly small. Want to see serious trampling? The war on drugs. That has had vastly more impact on people's lives than the current terrorism panic.
Nothing Ashcroft does can prevent a desperate individual willing to die for his cause.
Except imprisoning or deporting that individual before he can act. But I think your logic could be applied to all law enforcement efforts, ending in the conclusion that no laws should be enforced.
The saddest thing is that so many otherwise intelligent people are hypnotized by the Republican-Democrat pingpong, and try to politicize even mundane governmental functions like catching crooks. Ashcroft is a bogey man to the left, yet I doubt that he is doing his job any differently from any other Attorney General. In fact, I doubt that the AG has much impact on Justice - he's just a short-term figurehead who represents career civil servants to the public.
I say this because you didn't really reply to my post...
OK, re-reading your post, I see that you think the Unix userland is more important and long-lasting than the kernel. I pretty much agree, but applying the label "GNU" to that userland rubs me the wrong way, since a lot of the programmers who created it thought they were contributing to Linux. People have an unfortunate need to see some hero or leader behind everything, when really the "GNU/Linux" of today is the product of many programmers, not of Linus or Stallman. Face it: if those two got hit by a bus tomorrow, we'd continue without missing a step. So we're looking at a vital, growing body of software and debating which "patriarch" gets to stamp his logo on it based on bygone accomplishments.
XFree86 is a massive contribution, certainly comparable to both the GNU and Linux parts. Yet we don't happen to have a loud-voiced alpha male asserting its importance.
I think you might underestimate what Linus did. GNU was a handy package of utilities for sysadmins, but it wasn't creating the kind of development community that Linux created. Rather, it was a nice set of software that came from a particular group. Part of the Linux magic probably came from the ambitious and frustrated Minix community, which had tasted a bit of Unix on their PCs and wanted more. Compared to GNU, it was more down-to-earth, decentralized, and appreciative of the individual programmer.
First of all without GNU, there wouldn't have been a GCC...
You're saying that nobody would ever have written a useful free C compiler? Maybe you're right, though I doubt it. My point is that saying a piece of software is written by "GNU" or "Microsoft" is incorrect. Individuals write code. In the case of GNU, not all those contributors are happy with the way their work has gone to strengthen a certain power base. For such people, the most painful feeling must be that people like you give them no credit, but rather credit a collective entity. The fact that today's popular compiler is called "GCC" is a historical accident - if Stallman had not accepted the EGCS code, everyone would be using EGCS.
GNU is a brand name, like Microsoft or Apple or Google, designed to stoke positive feelings in those who fall under its spell. People infused with these positive feelings overlook the gaps in the brand story, and defend "their" brand against critics.
Drepper, if anything, is only talking about the C library which has changed in major ways over the years.
Yes. His point is that today's glibc came from his desire to build a worthy libc for Linux. Initially he didn't think it was important that it was called glibc. Now he probably wishes he had started from scratch rather than inherit the ideological baggage. Glibc is more a part of Linux than of GNU.
I hope you read the linked page. That is the pain of a programmer who has poured his heart into a project only to find it effectively snatched from him. It's one of the things the GPL was intended to prevent.
You seem to think I'm bashing Stallman, which I'm not. I'm grateful for his vision and hard work in getting the ball rolling.
The problem is in the label GNU. It makes people assume that Stallman either wrote or played a major role in creating the software. However today's GCC is really Cygnus's EGCS - the original GCC development could not keep up. EGCS embraced a more open, Linux-like development model and thereby outpaced GCC, eventually replacing it. So the name GCC is now a bit of a misnomer - GNU didn't make it, but rather hindered it.
I find this completely unacceptable and can assure everybody that I consider none of the code I contributed to glibc (which is quite a lot) to be as part of the GNU project and so a major part of what Stallman claims credit for is simply going away.
So the bottom line is that GNU, like Microsoft, takes credit for a lot of people's work, sometimes with their complete approval, sometimes against their will. In reality, most of the energy in free software came from Linux and people's desire to get Unixy things working on Linux.
Sounds good. I infer that you have no backup. Would it make sense to have 2X RAID0, one for production and one nightly copied for backup? That way, a single mistake in the workday couldn't do substantial harm.
And, you coulda created an account. That way your posts will be seen by more people.
The mini-ITX spec only calls for 55W of power, which isn't enough for a bunch of disks. Of course you could use a mini-ITX mainboard with an ATX power supply in some custom or semi-custom case.
As for the rest, I agree the author's goals are unclear.
I wonder if you can devise a reliable system where two employees alternate taking a disk home, so there is always at least one disk off site. For example, Alice brought in her disk today, and Bob left his home.
The backups could be encrypted so that if the employee disks are lost or stolen there's no harm done.
Slashdot Mirror
I think you're merging together two different rides. Splash mountain is a flume ride, with conveyor lifts and tall drops. There is no guest or cast member power. There is also a canoe ride in the Rivers of America, which works as you described, minus the flume and elevation change. The biggest drop in Splash Mountain has a runout adjacent to the Rivers of America, so I can see how the two would blur together.
Microsoft and Intel decide the direction of the PC. The manufacturers must follow or die. There can be differentiation within the field established by Wintel, such as SFF PC's, but a new mainboard must run the newest version of Windows.
I roughly agree with you about the trend of IT departments and the end of the truly Personal Computer within modern organizations. In fairness, a lot of what was done on an early PC with BASIC can be done within Excel. Even working within the narrowest confines of a locked-down machine, users can achieve substantial automation of repetitive work.
No! That's emphatically not what Microsoft is doing. In fact, if they continue their shoddy coding practices, this hardware-based security won't be worth much. Palladium (which this seems to part of) allows an application to encrypt, decrypt and "attest" via an API to the hardware that bypasses the OS. So while it's true that OS bugs should be removed from the critical path, application bugs remain in the critical path. If Phoenix's BIOS is providing TCPA-like authentication of the bootloader, that only proves the authorized bootloader takes control. Any flaws in the bootloader or subsequent OS could still be exploited. So this new BIOS won't protect a shaky OS or application.
If I try to inject a note of reality into a slashdot discussion, apparently it looks like I'm "taking the side of" the relevant Bad Guy. Paranoia does not help one to look at the issues - it blinds one with imaginary issues. If you were hunting a huge animal with only a .22 rifle, it would be wise to learn everything you can about the anatomy and habits of that animal. But paranoid people might see the animal behind every bush, and prefer to trade talll tales about how huge and evil the animal is. Then if these paranoid hunters tried to enlist the aid of nearby villagers, they would properly be laughed at, because they told obviously false tales about the animal. But these stories (the 10 inch teeth, the appetite for babies) would sound true to the paranoid hunters because they'd repeated and embellished them for so long.
An excess of cynicism is exactly the same as an excess of naivete. Both render one incapable of decisive action.
Whether or not that's true, the majority of mainstream entertainment is not legitimately available online. Apple Music Store is a huge exception - it remains to be seen how it will fare once aac's start showing up on p2p. On the whole, the entertainment industry is waiting for stronger protections before throwing their content on the net.
First, do you really see the cost of PC's going up? Maybe a brief spike until this technology is compl
I'm not too happy with 2.0. I used it because it came packaged with an OS and just for once, in the interest of speed I relaxed my rule of always building Apache from source (which would have been 1.x, of course) and accepted the package. I had several weird problems with 2.x on 2 different machines. Random errors in mod_perl, which I spent a lot of time trying to nail down until I switched to 1.3.x, solving it. And the part that really pisses me off - Apache 2.x inserts a bunch of garbage into each line in the error log, and the docs say you can't turn this off. That was the last straw for me. I sincerely hope 1.3.x is maintained forever.
In all my experience with 1.3.x, I never felt the anger and frustration that commercial software can cause, but 2.x did it with that brain-dead misfeature.
Or anything, really. Let's see:
Obviously I like something about it, or I'd have left by now.
Who do you think the marketplace is? A bunch of disgrunted hobbyists? No, the serious market is system integrators - companies, large and small that assemble computers. If they are selling to corporate customers, these system integrators may want to deliver computers that can't be tampered with by users. Many corporate sysadmins might welcome additional weapons to fight against viruses, pirated software, etc. Unlike the slashdot crowd, they won't be viewing this technology through paranoid eyes, but rather asking, "What can it do for me?" And they'll see a lot of potential. A lot of help in keeping PCs in a known, trusted state, rather than corrupted by user actions.
The other major market is retail PCs. If a strong DRM solution becomes widely used, it will enable lots of entertainment content to be sold online. Everyone (except slashdot) knows this, so everyone is scrambling like mad to become that solution. So if this system is called "HappyPuppy" for example, consumers shopping for a new PC will make sure it has HappyPuppy because that lets them download their favorite songs cheaply. No retailer will buy any more PCs without HappyPuppy because they wouldn't sell.
To a normal person, HappyPuppy is an additional capability, like having a DVD drive. It is not a restriction. It doesn't stop him from doing anything he could do before. Contrary to slashdot mythology, it doesn't stop him from downloading, using or sharing illegal mp3s. Of course, there is no way to extract the HappyPuppy content into something like mp3s, but there never was.
Actually, if the (h|cr|att)acker were clever, he could use this scenario to camouflage a theft. Design the attack so half the accounts are net gainers and half are net losers. Imagine a bell curve centered around 0. Maybe 3% of accounts gained or lost more than $10,000. The (h|cr|att)acker's accounts just happen to be in the upper 3%.
Still wouldn't work in the long run, for lots of reasons. The banks have backups, and when ATMs have erroneously given money to customers, the banks have chased it down.
That's probably the key issue. I wouldn't want to move large amounts of tabular information with XML. Most of the XML messages I handle are quite small, and the database processing time dwarfs the time to generate and transmit the XML. Bloating 1 KB to 3 KB doesn't matter, but bloating 100G to 300G would definitely matter.
This doesn't match my experience. I maintain a complex app that receives XML requests and returns XML responses. The ability of a human to quickly read the data is invaluable. Also, as messages evolve, there is no chance of a data item being misinterpreted due to being in the wrong column.
My XML is simple and readable, not SOAP or anything like that.
As time goes on, optional tags get added to requests to cause special behavior. Only the programmer who needs that behavior needs to care about the existence of the new tag. Likewise, more info gets added to responses, but the new info can be ignored by existing clients.
CSV has its place - importing and exporting table-formatted data. But for cross-platform messaging, XML is the best right now.
Can you give me some examples? Specifically, when have they charged the plaintiff in a lawsuit with a crime that is factually equivalent to losing the lawsuit? If neighbors Smith and Jones are embroiled in a lawsuit over a road between their properties, do the authorities ever charge Smith or Jones with trespassing for using that road?
I wonder how such an investigation actually begins. I'm guessing that it begins with a complaint from Adobe, and that a certain amount of expertise might be needed to make the complaint compelling. There are a lot of former prosecutors in private practice. It would probably cost about $200 to get some clues on the complaint process, or to be told why it won't work.
I don't agree at all. If someone is publicly reputed to have murdered his wife, and the wife is missing, maybe the authorities would take the initiative. Although even there, the story usually starts with someone phoning the police to report the missing person. In this case, a casual read of any SCO story in the papers does not reveal any criminal actions. I think you are too close to this story to see it objectively. There are many lawsuits in the courts, and almost all of them cause great hatred. Almost everyone embroiled in a lawsuit would like to see his adversaries thrown in prison. You may be able to construct a valid theory for prosecution of SCO, but it's a reach. It should be constructed by a criminal attorney and patiently explained to the relevant authorities. I tend to think that if it can be done, IBM's
lawyers will do it.
I share your evident desire to see McBride severely punished for his actions. And don't forget co-conspirators Ralph Yarrow and Chris Sontag. I just don't see any realistic hope for it, as the legal system has not adapted yet to handle this kind of abuse. These people are criminals in all but name, but since their chosen weapon is the lawsuit rather than the gun, they are probably immune to prosecution. I think they will lose their suit, but land on their feet with enough money for a modest retirement.
Q:Should I bring my umbrella on Thursday?
A:Rain could occur on any day.
Q:Which car is more reliable, a Toyota Corolla or a BMW 325?
A:Anyone driving any car can experience a mechanical breakdown.
Q:Should I drink this six month old milk in the fridge or buy some new milk?
A:Anyone drinking any milk can die of food poisoning.
Microsoft's approach to security is deeply flawed. Again and again, they made visibly wrong decisions which any experienced network programmer could see as wrong. Their permissions system sucks, for example. They thought they would be clever and leapfrog Unix - they would go from no security to fancy ACL's. Unfortunately, almost none of their customer base can figure out the fancy ACL's, and most of their ISV's are not cooperating. Unix ugo perms are already at the outer limit of what most people can understand. On a typical Unix box, the majority of files have correct permissions, and a minority have a non-disastrous error in perms. On a typical Windows box, either everything is wide open, or everything is locked down so Administrator intervention is needed for almost anything.
Both Windows and Linux contain flaws in execution. But Windows contains severe flaws in design.
If you define the top x (fraction) as "the best and brightest" and Microsoft has n employees with the same brightness distribution as the general population, it follows that they employ roughly nx of the "best and brightest." If nx > 3 I think everyone would agree that Microsoft employs some of the best and brightest. E.g. 65,000 *
Of course this assertion, like the one above, is intended to connote more than it denotes. Since you raise this alleged brightness as a defense against accusations of making bad software, you appear to argue that the software isn't really bad, since bright people can't make bad software. This seems like a curiously indirect way of evaluating products that have inflicted real-world pain on many of us.
Yes! And make it withstand a 6 foot drop onto concrete. Lots of things, like handheld radios already meet this spec. Why not computers?
Your argument makes sense. Unfortunately, SCO seems to be saying that GPL'd code is effectively in the public domain. If so, that would also constitute their defense against copyright infringement. So it seems to make sense to evaluate that argument just once, not in two different courts.
Where I completely disagree with you is your assumption that the government operates only at the behest of special interests. Some people are so politically tied-in that they could be shielded from prosecution. Darl McBride is not one of them.
Prosecutors want to put people in prison. If they could put someone in prison with five minutes work, they would do it, whether that person is Linus Torvalds or Darl McBride. But if they have to devote huge resources to a conviction that carries a light sentence, they'd rather skip it and work on easier cases.
Have any Linux authors complained to their local US attorney's office? I think that generally the victim has to take the initiative and complain, rather than waiting for the feds to magically figure out the situation.
I have no idea if your figures are correct, but there's no way you can simply look at a number and scream "too much!" You have to understand what is bought for that number.
Sun Tzu wrote something like "It is cheaper to pay the best spies lavishly than to pay the worst army stingily." NSA earned its huge budget by producing priceless intelligence that advanced US interests and saved US lives. The history of signals intelligence shows that it is a good investment for nations.
The DOJ doesn't generally prosecute violent crime.
And they are weak at white-collar crime, not for sinister reasons but because these cases are very hard and expensive to try. A prosecutor would rather put five people in prison, and let one go, than spend all his time putting that guy in prison and let five go.
The validity of SCO's claims is still up in the air, and will be settled in court. Do you seriously expect the Justice Department to initiate a criminal prosecution based on an issue that is pending settlement in civil court? Can you think of any example where this happened? And just because SCO is constant front-page news here doesn't mean DOJ knows or cares about them.
Can you find a single criminal attorney in the US, either prosecutor or defender, who believes there is a solid criminal case against SCO or some of their employees? Remember, DOJ is run by lawyers, not geeks. They only pursue cases that they think they can win. What is your basis for thinking they could win such a case? What would the charges be exactly?
Even if SCO loses, I doubt that they will be prosecuted for copyright infringement, because they arguably acted in good faith. They distributed code which they thought they owned. Copyright infringement must be wilful to be criminal.
I agree that there's too much tinfoil-hattism here, but I think there's a big difference between a Grand Jury demanding information and an FBI Agent demanding the information. Nobody can reasonably object to a judge or Grand Jury extracting private information - without that power injustice would flourish everywhere in secrecy. We worry about the executive branch silently and easily vacuuming up our personal info.
Except imprisoning or deporting that individual before he can act. But I think your logic could be applied to all law enforcement efforts, ending in the conclusion that no laws should be enforced.
The saddest thing is that so many otherwise intelligent people are hypnotized by the Republican-Democrat pingpong, and try to politicize even mundane governmental functions like catching crooks. Ashcroft is a bogey man to the left, yet I doubt that he is doing his job any differently from any other Attorney General. In fact, I doubt that the AG has much impact on Justice - he's just a short-term figurehead who represents career civil servants to the public.
OK, re-reading your post, I see that you think the Unix userland is more important and long-lasting than the kernel. I pretty much agree, but applying the label "GNU" to that userland rubs me the wrong way, since a lot of the programmers who created it thought they were contributing to Linux. People have an unfortunate need to see some hero or leader behind everything, when really the "GNU/Linux" of today is the product of many programmers, not of Linus or Stallman. Face it: if those two got hit by a bus tomorrow, we'd continue without missing a step. So we're looking at a vital, growing body of software and debating which "patriarch" gets to stamp his logo on it based on bygone accomplishments.
XFree86 is a massive contribution, certainly comparable to both the GNU and Linux parts. Yet we don't happen to have a loud-voiced alpha male asserting its importance.
I think you might underestimate what Linus did. GNU was a handy package of utilities for sysadmins, but it wasn't creating the kind of development community that Linux created. Rather, it was a nice set of software that came from a particular group. Part of the Linux magic probably came from the ambitious and frustrated Minix community, which had tasted a bit of Unix on their PCs and wanted more. Compared to GNU, it was more down-to-earth, decentralized, and appreciative of the individual programmer.
The fact that today's popular compiler is called "GCC" is a historical accident - if Stallman had not accepted the EGCS code, everyone would be using EGCS.
GNU is a brand name, like Microsoft or Apple or Google, designed to stoke positive feelings in those who fall under its spell. People infused with these positive feelings overlook the gaps in the brand story, and defend "their" brand against critics.
Yes. His point is that today's glibc came from his desire to build a worthy libc for Linux. Initially he didn't think it was important that it was called glibc. Now he probably wishes he had started from scratch rather than inherit the ideological baggage. Glibc is more a part of Linux than of GNU.
I hope you read the linked page. That is the pain of a programmer who has poured his heart into a project only to find it effectively snatched from him. It's one of the things the GPL was intended to prevent.
You seem to think I'm bashing Stallman, which I'm not. I'm grateful for his vision and hard work in getting the ball rolling.
Read this statement by Ulrich Drepper, glibc maintainer. Among other things, he says:
So the bottom line is that GNU, like Microsoft, takes credit for a lot of people's work, sometimes with their complete approval, sometimes against their will. In reality, most of the energy in free software came from Linux and people's desire to get Unixy things working on Linux.
Oh, and log in so more people see your posts.
Good link.
Sounds good. I infer that you have no backup. Would it make sense to have 2X RAID0, one for production and one nightly copied for backup? That way, a single mistake in the workday couldn't do substantial harm.
And, you coulda created an account. That way your posts will be seen by more people.
The mini-ITX spec only calls for 55W of power, which isn't enough for a bunch of disks. Of course you could use a mini-ITX mainboard with an ATX power supply in some custom or semi-custom case.
As for the rest, I agree the author's goals are unclear.
I care. No need to get defensive. Have you considered using rsync? It's --delete option will delete files that don't exist on the sending side.
And why don't you create an account so more people can view your posts?
I wonder if you can devise a reliable system where two employees alternate taking a disk home, so there is always at least one disk off site. For example, Alice brought in her disk today, and Bob left his home.
The backups could be encrypted so that if the employee disks are lost or stolen there's no harm done.