I tried JFS, and it handled power interruptions very poorly.
Essentially, I liked philosophically that the act of mounting and journal replay are separated, it really makes sense. Journal replay should be more an fsck option, thought that was neat. And when you mount read-only, you *mean* read-only, no journal reply or anything even on a 'dirty' filesystem.
However, I found all too frequently that after power failures, it would replay the journal and think everything was fine, until a few hours of usage later when it figures out that it left something in an inconsistant state and remounts read only all of a sudden. Then you fsck and watch lost+found get a few more files. As long as I could recognize the files, I could put them back fairly easily, but I haven't had issue with ext3 yet. Have had similar issues to this with XFS, and, admittedly, far worse with Reiser.
Anyway, returning to topic, GPFS is a filesystem for shared-storage SANs and for aggregating individual node storage into a potentially fault tolerant filesystem (or filesystems). Since they ditched the RSCT stuff a while back, I've found them to be fairly robust and not overly difficult to configure (Lustre I found significantly harder than new GPFS, but lustre is easier than old GPFS to get running). It is not suitable for desktop systems.
Hard to dismiss cronyism legitimately for the ports deal, or, conversely, if you reject the concept of cronyism you can't accept that other activities of the US government are not malicious in intent.
This administration has overseen fairly egregious violations of citizen's rights in the wake of the 9/11 attacks in the name of national security. We have been asked to sacrifice a significant deal of privacy and had privacy in some cases taken without knowledge to them for the purposes of national security. I disagree with these policies, but can logically see the correlation between what they demand and the justification, though it appears to be more a believable excuse than true justification. Despite my worries about these occurances, this in and of itself doesn't prove any malicious intent rather than an overly paranoid reaction with good intentions.
Now this ports deal comes along. I won't say outright that the deal decreases our security enough to seriously worry, but it at the very least has the potential to increase it enough to give us pause. At the same time over-zealous wire tapping and arrests without due process are occuring in the name of national security (in many cases to no effect at all on our safety), the same administration turns around and say a purchase of port terminals by a foreign body is nothing to worry about and all, and in fact so little to worry about they should skip the requisite review process. Add to this that while the foreign entity has been cooperative and an ally, suspicious things have happened around them that either suggest that some people within this entity are not friendly, or that they are legitimately friendly, but unable to properly implement sufficient security measures, either way means extra care should be taken with such a potential deal. This is logically opposed to the behavior with respect to everything else this administration has done/called for along these lines. The very fact the administration involved itself so quickly and so adamantly to the point of waiving standard procedures for the UAE just seems highly improper, even without the context of being otherwise over-zealous paranoid everywhere else.
If the administration hadn't taken an intense interest in this, and if it wasn't occuring in a context of overly paranoid national security measures, I wouldn't think twice about it, but all the indications are there that something significantly improper is going on with this ports deal. Dubai may be correct in every way how they deal with it, but Dubya is certainly making the deal look bad by meddling with it so much.
Whether they should have blocked or not, who knows, but a lot of IBM employees can testify that a) the government reviewed the deal with interest and b) steps have been taken/are being continuously taken to comply with government security demands with respect to this deal. Considering all the stuff that went Lenovo wasn't exactly anywhere near special (pretty much southeast asian made systems to specifications and testing in the US), and the steps being taken are doing a pretty good job of keeping Lenovo and IBM separate, there is admittedly little risk with respect to that deal.
Ultimately I suspect the same sort of thing will happen with this case, but it won't be so tricky since it doesn't involve a foreign owned company taking ownership of part of a company and sharing a campus with an American company with some potentially sensitive data. I guess the deal itself may be more questionable due to the potentially more sensitive nature of the transfer...
In any event, neither occurance has a potential for security issues similar to the ports deal. This article was little more than an excuse to discuss the ports deal on slashdot, IMHO.
For google, a core part of their PR strategy is 'do no evil', and therefore any opportunity to grandstand in a way that appears to comply with this core promise is gold for Google.
The other sites don't have that as a PR strategy at the moment. Therefore, they would perceive little to no value compared to their costs.
Of course, it does sound good to stand up to the government lately with all the negative trends against privacy going on, but as many have pointed out, google themselves is using the data in ways not that much different from the government plans, so it isn't 100% as good as they like everyone to think...
No, a lot of people use commonly inexpensive, but that is just really silly the more you think about it. What about the technology dictates that the devices used as members should be inexpensive? There is a fair amount of independence implied by the approach, so it is fair to use independent but it is not defensible to say it is inherently inexpensive disks used.
I would like to replace 'disks' with 'devices' or 'drives', because there is also nothing disk-specific about the technology. Flash memory or whatever could be used as members.
And a complete dismissal of so-called RAID-0, since it makes no sense when the acronym is expanded.
And I would say even if you have, say, a mirrored set of RAID5 arrays (RAID1 of RAID5 arrays), you wouldn't say Redundant Rundundant Array of Independent Drives, you would say Redundant Arrays of Independent Drives, so english wise you wouldn't add more 'redundants' to describe that behavior.
Notice all these particularly high end car companies don't name their cars? i.e. BMW 330i
The problem they had before was they tried to have the product number stand on its own, so the marketing was focused on the 486 processor, for example. Other companies did 486s, and intel ran into issues, and so they wanted a trademarkable product name, 'Pentium'.
Now, they look at those car companies, and there is a key difference. This isn't the 'D processor', it would be the "Intel D processor" In other words, the product-specific name is too short/unintelligble to be usefully distinguishable, and the market is forced to have the Intel brand name in too. They want to enhance and leverage their brand versus the product like BMW, Lexus, et al do. If they had thought this 10 years ago, we wouldn't have the Pentium we might have been emphasized as 'The Intel 586', though 586 might have been made less predictable, useful, or generally made unable to stand on it's own as a product family identifier without the Intel name to have any clue as to what context to consider it in.
I think its interesting the inferred deferral of responsibility in the response to the snake feeding example.
It's only natural for the snake to eat a mouse, but the point is the human is putting the mouse in the position of being eaten, knowing precisely what the consequence of putting that situation together, but since the final stroke is not done by their hands, they are less responsible, and therefore less cruel. If they had to hypothetically kill the mouse for the snake before feeding it, some may have more problems with doing that.
Similarly, the average person eats meat, but wouldn't kill an animal and eat it because the experience seems horrific. Again, the actual burden of the act of killing is deferred, but the person benefits from and to some extent can be considered responsible for the act.
It's fascinating how for a lot of people is a larger measure of cruelty is how dirty the person's hands directly get in the act versus how responsible they are for the act.
Of course, I'm one of the people who eat meat but wouldn't kill an animal, but at least I recognize my psuedo-hypocrisy for what it is.
But is it any more cruel than the typical use of mice as snake food where they are fed live to a snake? Also undoubtedly would induce as much fear as strangulation would, if not more so since the snake situation is exacerbated by facing a natural predator. Personally, I couldn't do that, but it is a widespread accepted practice that seems not very different from this experiment.
However, it does seem rather pointless, considering how specific the test is and it doesn't reflect how useful this would be in humans. I would think it easy to collect samples from cadavers with well known causes of death and test those. Maybe they need shortly before to compare against?
There are degrees of apparent privacy. Before home video, there were theaters, and therefore home video market when available was greatly influenced by porn availability. You have to go to a public store and be relatively public or receive a potentially conspicuous package, but have to wait and still risk embarassment. The home video market exploded, decreasing the theater market to nil and growing the overall market for porn in general.
Nowadays, how sizable is the home porn video market compared to the more anonymous, the more instantly 'gratifying' internet porn market that has presumably overwhelmed DVD/VHS distribution due to the immediacy and anonymity the computer offers. If nothing else, seeing all the computers I've dealt with where people stick porn in places they perceive as obscure suggests they have higher confidence in hiding files on a computer than hiding tapes or discs in their home. Even for the television channels, I would wager people feel safer buying some porn network/pay-per-view and hiding the charges on their credit card they find easier than hiding discs/tapes.
In essence, as amusing it is to think of porn as a huge market force in such a context, it probably isn't realistic to consider it a 'killer app' this time around. However, I doubt Sony will be so prudish this time compared to the Betamax fiasco, just to be on the safe side.
If I did do Windows more: I would replace SSH.com's client with PuTTY, it does the job and has a much nicer license. I would supplement it with WinSCP for file transfers.
I would not recommend ClamAV at all, absolutely worthless in my testing on a Windows box, instead I'd probably recommend Avast. Slow, but actually catches things which I never saw ClamAV do. Just seems like ClamAV didn't have a nearly complete enough definition database compared to Avast.
And on limited privilege accounts, I'd feel relatively helpless as the home user applications more times than not are so badly done that they won't run without Administrator level privileges.
Sounds all nice and good, but when you stop to think about it, they weren't helped so much by the desktop aspects of linux, but server side aspects. In that line of reasoning, how much testing, bug reports, and fixes have they provided back to the kernel and relevant untilities to them? I honestly have no idea, but the areas in which Google would be 'paying back' their benefit of linux is in places very mundane and boring to the linux desktop market, and therefore for a great deal of users so low profile as to appear ungrateful despite efforts they may be making to really bolster the enterprise-capable aspects of Linux.
My bet is that they have made significant QA and development contributions back, simply because any large scale user of any technology contributes QA back, and if an open technology and they have technical skills available, they will be impatient enough to make progress in fixing it themselves.
All this aside, if nothing else consider the marketing leverage Google provides by advocates being able to point at google as a successful extremely large deployment of Linux. Not so significant nowadays since Linux is taken seriously, but when Google first started deploying with Linux, a great deal of the market still considered it unproven, and moves like Google's served to help convince skeptical would-be users that there is value and maturity in the Linux platform.
What I've always heard is that it is more cost effective to do it this way than maintain old product lines. This is because, for the most part, the budget processors are chips off the line that have flaws preventing reliable use of the full cache, and so that cache is disabled and shipped. Occasionally there are incidents of particularly good yield and they will take otherwise good processors. Simply because they know they have more product than their demand dictates at the desired price point for the high-end, and vice-versa on the low end. Though not cache, there was a long run of Celeron's that could easily and reliably be clocked up that someone had said was due to this phenomenon.
Of course, this could all be tech urban legend, but it fits in pretty well with the practices I do know about in the business.
Yes, pedal to the floor turning into a banked turn when suddenly the car decides to use that as actual input and your garage gets screwed..
I know, engine would have to start and all and thus this scenario is less likely to proceed all the way through without driver notice, but a fun thought.
More net productivity is the point and is valid. But it is really hard to convince a lot of managers.
For example, I have a sort of shared leadership role with a small set of people. At the times I'm pretty much in charge, there will be an occasional bout of gaming (first person shooter, whatever so long as it can be done quickly in the period of time like a typical break or lunch). Others who have discovered this became infuriated, and complained we are spread too thin to afford any time whatsoever to goof off, ignoring the fact that we had been acheiving more than previously done.
Managers on average can't understand the concept that anything but working during a work day can possibly yield benefit. In this case once I started the gaming on occasion, it was clear how more focused and productive people were outside those breaks. It seemed like the obvious motivational and relaxed association between workers did have an effect pretty much along the lines of what people normally expect. But perhaps more importantly, the amount of individual goofing off diminished. I.e. people would spend a fair amount of time browsing forums, news, and such. Part of that may simply be part of improved morale/motivation, but another part seemed like there was an attitude of 'they just explicitly paid me to goof off, I need to work to make up for it and to show they don't need to stop the practice for me to do well'. It's like focusing the goofing off into manageable blocks and making people want to prove they can goof off and do good work.
Does that surprise you or is it in any way different from 95% of other companies out there today?
Sorry, just irritated that not only is this strategy so widespread, but that it is so effective in the market. Why are people generally more caught up in a brand than the actual product?
Think their marketing message is that it is 'revolutionary', and hence 360. xbox 180 would be saying whatever we did last was wrong and we are going a totally different direction, but that isn't the message they feel is correct. They want people to perceive that they have been getting it right, but this next console will revolutionize things along the path they demonstrated, but not abandoning the overall strategic direction.
Yes, it marketize and is confusing, but xbox 180 clearly would send the wrong message regarding their confidence/commitment when the last platform didn't totally bomb. Now other companies that have had horribly bad reputations have done things like say 'we screwed up and learned, try us again', but if their isn't the wide perception of screwing up, the message would just be bizarre.
It is a hash algo. It's used not to protect the content of anything, just to provide a method to validate content integrity, to show nothing accidental or intentional happened to change it.
Every dealing I have with them is a strange mixture of trying to be too cutting edge at times, and at others so terrified of even straightforward patches required to avoid panics. In general when I deal with them, I invariably get someone who has a huge ego that exceeds their ability. The fact I've had so many dealings with them speaks volumes on the QA issues. Once I did deal with some RedHat folks in the UK, and they seemed more level headed than the US RedHat folks, admittedly.
By far, I have had immense success dealing with and deploying SuSE. I haven't had to actually deal with them since the Novell buyout, but at least back then the support and path to someone appropriately competent was reasonable. They were prepared to take very sane courses of action to get problem resolution. They never seemed too paranoid of patches nor did they try to push things too far feature-wise. These occurances where I have to deal with them are once in a blue moon, which speaks highly of their support (use RHEL4 and SLES9 in various capacities to this day, but have only had to deal with RedHat support despite heavy SLES usage).
On a technical and support level in an enterprise, SuSE by far has my recommendation.
Slashdot Mirror
If you RTFA, it also applies to the user password that gets unlimited sudo access. Which means, by default, you still get screwed.
That is a lot more expensive than a magic marker or hole punch.
I tried JFS, and it handled power interruptions very poorly.
Essentially, I liked philosophically that the act of mounting and journal replay are separated, it really makes sense. Journal replay should be more an fsck option, thought that was neat. And when you mount read-only, you *mean* read-only, no journal reply or anything even on a 'dirty' filesystem.
However, I found all too frequently that after power failures, it would replay the journal and think everything was fine, until a few hours of usage later when it figures out that it left something in an inconsistant state and remounts read only all of a sudden. Then you fsck and watch lost+found get a few more files. As long as I could recognize the files, I could put them back fairly easily, but I haven't had issue with ext3 yet. Have had similar issues to this with XFS, and, admittedly, far worse with Reiser.
Anyway, returning to topic, GPFS is a filesystem for shared-storage SANs and for aggregating individual node storage into a potentially fault tolerant filesystem (or filesystems). Since they ditched the RSCT stuff a while back, I've found them to be fairly robust and not overly difficult to configure (Lustre I found significantly harder than new GPFS, but lustre is easier than old GPFS to get running). It is not suitable for desktop systems.
http://www.bash.org/?13213
Fun keyword filtering.
Hard to dismiss cronyism legitimately for the ports deal, or, conversely, if you reject the concept of cronyism you can't accept that other activities of the US government are not malicious in intent.
This administration has overseen fairly egregious violations of citizen's rights in the wake of the 9/11 attacks in the name of national security. We have been asked to sacrifice a significant deal of privacy and had privacy in some cases taken without knowledge to them for the purposes of national security. I disagree with these policies, but can logically see the correlation between what they demand and the justification, though it appears to be more a believable excuse than true justification. Despite my worries about these occurances, this in and of itself doesn't prove any malicious intent rather than an overly paranoid reaction with good intentions.
Now this ports deal comes along. I won't say outright that the deal decreases our security enough to seriously worry, but it at the very least has the potential to increase it enough to give us pause. At the same time over-zealous wire tapping and arrests without due process are occuring in the name of national security (in many cases to no effect at all on our safety), the same administration turns around and say a purchase of port terminals by a foreign body is nothing to worry about and all, and in fact so little to worry about they should skip the requisite review process. Add to this that while the foreign entity has been cooperative and an ally, suspicious things have happened around them that either suggest that some people within this entity are not friendly, or that they are legitimately friendly, but unable to properly implement sufficient security measures, either way means extra care should be taken with such a potential deal. This is logically opposed to the behavior with respect to everything else this administration has done/called for along these lines. The very fact the administration involved itself so quickly and so adamantly to the point of waiving standard procedures for the UAE just seems highly improper, even without the context of being otherwise over-zealous paranoid everywhere else.
If the administration hadn't taken an intense interest in this, and if it wasn't occuring in a context of overly paranoid national security measures, I wouldn't think twice about it, but all the indications are there that something significantly improper is going on with this ports deal. Dubai may be correct in every way how they deal with it, but Dubya is certainly making the deal look bad by meddling with it so much.
Whether they should have blocked or not, who knows, but a lot of IBM employees can testify that a) the government reviewed the deal with interest and b) steps have been taken/are being continuously taken to comply with government security demands with respect to this deal. Considering all the stuff that went Lenovo wasn't exactly anywhere near special (pretty much southeast asian made systems to specifications and testing in the US), and the steps being taken are doing a pretty good job of keeping Lenovo and IBM separate, there is admittedly little risk with respect to that deal.
Ultimately I suspect the same sort of thing will happen with this case, but it won't be so tricky since it doesn't involve a foreign owned company taking ownership of part of a company and sharing a campus with an American company with some potentially sensitive data. I guess the deal itself may be more questionable due to the potentially more sensitive nature of the transfer...
In any event, neither occurance has a potential for security issues similar to the ports deal. This article was little more than an excuse to discuss the ports deal on slashdot, IMHO.
For google, a core part of their PR strategy is 'do no evil', and therefore any opportunity to grandstand in a way that appears to comply with this core promise is gold for Google.
The other sites don't have that as a PR strategy at the moment. Therefore, they would perceive little to no value compared to their costs.
Of course, it does sound good to stand up to the government lately with all the negative trends against privacy going on, but as many have pointed out, google themselves is using the data in ways not that much different from the government plans, so it isn't 100% as good as they like everyone to think...
> Renders a GUI using graphics hardware.
How are you supposed to have a GUI without graphics hardware? 'Render' does not automatically mean 3D, you need to explicitly state that...
No, a lot of people use commonly inexpensive, but that is just really silly the more you think about it. What about the technology dictates that the devices used as members should be inexpensive? There is a fair amount of independence implied by the approach, so it is fair to use independent but it is not defensible to say it is inherently inexpensive disks used.
I would like to replace 'disks' with 'devices' or 'drives', because there is also nothing disk-specific about the technology. Flash memory or whatever could be used as members.
And a complete dismissal of so-called RAID-0, since it makes no sense when the acronym is expanded.
And I would say even if you have, say, a mirrored set of RAID5 arrays (RAID1 of RAID5 arrays), you wouldn't say Redundant Rundundant Array of Independent Drives, you would say Redundant Arrays of Independent Drives, so english wise you wouldn't add more 'redundants' to describe that behavior.
An Insightful comment can contain no information. If it was scored informative, you have a valid beef.
Some of the most insightful statements in the world are simply important questions.
Now this comment, -1 Offtopic.
Notice all these particularly high end car companies don't name their cars? i.e. BMW 330i
The problem they had before was they tried to have the product number stand on its own, so the marketing was focused on the 486 processor, for example. Other companies did 486s, and intel ran into issues, and so they wanted a trademarkable product name, 'Pentium'.
Now, they look at those car companies, and there is a key difference. This isn't the 'D processor', it would be the "Intel D processor" In other words, the product-specific name is too short/unintelligble to be usefully distinguishable, and the market is forced to have the Intel brand name in too. They want to enhance and leverage their brand versus the product like BMW, Lexus, et al do. If they had thought this 10 years ago, we wouldn't have the Pentium we might have been emphasized as 'The Intel 586', though 586 might have been made less predictable, useful, or generally made unable to stand on it's own as a product family identifier without the Intel name to have any clue as to what context to consider it in.
I think its interesting the inferred deferral of responsibility in the response to the snake feeding example.
It's only natural for the snake to eat a mouse, but the point is the human is putting the mouse in the position of being eaten, knowing precisely what the consequence of putting that situation together, but since the final stroke is not done by their hands, they are less responsible, and therefore less cruel. If they had to hypothetically kill the mouse for the snake before feeding it, some may have more problems with doing that.
Similarly, the average person eats meat, but wouldn't kill an animal and eat it because the experience seems horrific. Again, the actual burden of the act of killing is deferred, but the person benefits from and to some extent can be considered responsible for the act.
It's fascinating how for a lot of people is a larger measure of cruelty is how dirty the person's hands directly get in the act versus how responsible they are for the act.
Of course, I'm one of the people who eat meat but wouldn't kill an animal, but at least I recognize my psuedo-hypocrisy for what it is.
But is it any more cruel than the typical use of mice as snake food where they are fed live to a snake? Also undoubtedly would induce as much fear as strangulation would, if not more so since the snake situation is exacerbated by facing a natural predator. Personally, I couldn't do that, but it is a widespread accepted practice that seems not very different from this experiment.
However, it does seem rather pointless, considering how specific the test is and it doesn't reflect how useful this would be in humans. I would think it easy to collect samples from cadavers with well known causes of death and test those. Maybe they need shortly before to compare against?
There are degrees of apparent privacy. Before home video, there were theaters, and therefore home video market when available was greatly influenced by porn availability. You have to go to a public store and be relatively public or receive a potentially conspicuous package, but have to wait and still risk embarassment. The home video market exploded, decreasing the theater market to nil and growing the overall market for porn in general.
Nowadays, how sizable is the home porn video market compared to the more anonymous, the more instantly 'gratifying' internet porn market that has presumably overwhelmed DVD/VHS distribution due to the immediacy and anonymity the computer offers. If nothing else, seeing all the computers I've dealt with where people stick porn in places they perceive as obscure suggests they have higher confidence in hiding files on a computer than hiding tapes or discs in their home. Even for the television channels, I would wager people feel safer buying some porn network/pay-per-view and hiding the charges on their credit card they find easier than hiding discs/tapes.
In essence, as amusing it is to think of porn as a huge market force in such a context, it probably isn't realistic to consider it a 'killer app' this time around. However, I doubt Sony will be so prudish this time compared to the Betamax fiasco, just to be on the safe side.
If I did do Windows more:
I would replace SSH.com's client with PuTTY, it does the job and has a much nicer license.
I would supplement it with WinSCP for file transfers.
I would not recommend ClamAV at all, absolutely worthless in my testing on a Windows box, instead I'd probably recommend Avast. Slow, but actually catches things which I never saw ClamAV do. Just seems like ClamAV didn't have a nearly complete enough definition database compared to Avast.
And on limited privilege accounts, I'd feel relatively helpless as the home user applications more times than not are so badly done that they won't run without Administrator level privileges.
Sounds all nice and good, but when you stop to think about it, they weren't helped so much by the desktop aspects of linux, but server side aspects. In that line of reasoning, how much testing, bug reports, and fixes have they provided back to the kernel and relevant untilities to them? I honestly have no idea, but the areas in which Google would be 'paying back' their benefit of linux is in places very mundane and boring to the linux desktop market, and therefore for a great deal of users so low profile as to appear ungrateful despite efforts they may be making to really bolster the enterprise-capable aspects of Linux.
My bet is that they have made significant QA and development contributions back, simply because any large scale user of any technology contributes QA back, and if an open technology and they have technical skills available, they will be impatient enough to make progress in fixing it themselves.
All this aside, if nothing else consider the marketing leverage Google provides by advocates being able to point at google as a successful extremely large deployment of Linux. Not so significant nowadays since Linux is taken seriously, but when Google first started deploying with Linux, a great deal of the market still considered it unproven, and moves like Google's served to help convince skeptical would-be users that there is value and maturity in the Linux platform.
What I've always heard is that it is more cost effective to do it this way than maintain old product lines. This is because, for the most part, the budget processors are chips off the line that have flaws preventing reliable use of the full cache, and so that cache is disabled and shipped. Occasionally there are incidents of particularly good yield and they will take otherwise good processors. Simply because they know they have more product than their demand dictates at the desired price point for the high-end, and vice-versa on the low end. Though not cache, there was a long run of Celeron's that could easily and reliably be clocked up that someone had said was due to this phenomenon.
Of course, this could all be tech urban legend, but it fits in pretty well with the practices I do know about in the business.
Yes, pedal to the floor turning into a banked turn when suddenly the car decides to use that as actual input and your garage gets screwed..
I know, engine would have to start and all and thus this scenario is less likely to proceed all the way through without driver notice, but a fun thought.
More net productivity is the point and is valid. But it is really hard to convince a lot of managers.
For example, I have a sort of shared leadership role with a small set of people. At the times I'm pretty much in charge, there will be an occasional bout of gaming (first person shooter, whatever so long as it can be done quickly in the period of time like a typical break or lunch). Others who have discovered this became infuriated, and complained we are spread too thin to afford any time whatsoever to goof off, ignoring the fact that we had been acheiving more than previously done.
Managers on average can't understand the concept that anything but working during a work day can possibly yield benefit. In this case once I started the gaming on occasion, it was clear how more focused and productive people were outside those breaks. It seemed like the obvious motivational and relaxed association between workers did have an effect pretty much along the lines of what people normally expect. But perhaps more importantly, the amount of individual goofing off diminished. I.e. people would spend a fair amount of time browsing forums, news, and such. Part of that may simply be part of improved morale/motivation, but another part seemed like there was an attitude of 'they just explicitly paid me to goof off, I need to work to make up for it and to show they don't need to stop the practice for me to do well'. It's like focusing the goofing off into manageable blocks and making people want to prove they can goof off and do good work.
Does that surprise you or is it in any way different from 95% of other companies out there today?
Sorry, just irritated that not only is this strategy so widespread, but that it is so effective in the market. Why are people generally more caught up in a brand than the actual product?
You have entirely too much confidence in the slashdot readership...
The cost is $1080 dollars, since it is ~30/month minimum of 3 years.
Second, an Ultra 20 Opteron does not mean 20 Opterons, it means a workstation model 20 with one processor.
So, while reasonable, not nearly incredible.
Think their marketing message is that it is 'revolutionary', and hence 360. xbox 180 would be saying whatever we did last was wrong and we are going a totally different direction, but that isn't the message they feel is correct. They want people to perceive that they have been getting it right, but this next console will revolutionize things along the path they demonstrated, but not abandoning the overall strategic direction.
Yes, it marketize and is confusing, but xbox 180 clearly would send the wrong message regarding their confidence/commitment when the last platform didn't totally bomb. Now other companies that have had horribly bad reputations have done things like say 'we screwed up and learned, try us again', but if their isn't the wide perception of screwing up, the message would just be bizarre.
It is a hash algo. It's used not to protect the content of anything, just to provide a method to validate content integrity, to show nothing accidental or intentional happened to change it.
Every dealing I have with them is a strange mixture of trying to be too cutting edge at times, and at others so terrified of even straightforward patches required to avoid panics. In general when I deal with them, I invariably get someone who has a huge ego that exceeds their ability. The fact I've had so many dealings with them speaks volumes on the QA issues. Once I did deal with some RedHat folks in the UK, and they seemed more level headed than the US RedHat folks, admittedly.
By far, I have had immense success dealing with and deploying SuSE. I haven't had to actually deal with them since the Novell buyout, but at least back then the support and path to someone appropriately competent was reasonable. They were prepared to take very sane courses of action to get problem resolution. They never seemed too paranoid of patches nor did they try to push things too far feature-wise. These occurances where I have to deal with them are once in a blue moon, which speaks highly of their support (use RHEL4 and SLES9 in various capacities to this day, but have only had to deal with RedHat support despite heavy SLES usage).
On a technical and support level in an enterprise, SuSE by far has my recommendation.