The current system of more and more data collecting isn't for more security. That's just how it's sold. It is, bluntly, control. Over your data and you. It is easier to pinpoint and neutralize "troublemakers" before they start gaining a lot of support.
Who are these troublemakers being pinpointed and neutralized? What sort of trouble are they making? How is the pinpointing and neutralizing being accomplished? Ever known anyone who was pinpointed and neutralized? And why haven't they (the ominous 'they') pinpointed and neutralized you yet - for posting things like this?
The article is 48 paragraphs long. MS/Gates are mentioned in 3 paragraphs. Intel is mentioned in more than 15 paragraphs, and their bad effect on OLPC is singled out by Negroponte multiple times. Yet most of the/. replies single out MS as the bad guy here - for cutting price on their software. Hardly a mention of Intel, who have done far more to damage OLPCs chances in the market.
Hmm. Slashdot reading comprehension in top form, eh?
Way to go on that security model thing, Microsoft!
That security model thing doesn't work any better when you bypass it by being Administrator on Windows, than when you bypass it by being root on Linux.
Please stop trying to fit me into any of the pro- or anti-MS pigeonholes.
None of my post cared specifically who has the data. Microsoft, Canonical, God, the Devil - that wasn't the point. The point was, whomever you get your updates from can know what software and hardware you have. The question isn't what they can know, it's whether they bother to do anything with the knowledge.
You've misinterpreted my response. I have no desire to FUD for MS, only to clarify the technical picture. A more succinct version of what I said:
Windows Update works in a similar way to the Ubuntu update you are championing. It downloads a list of all updates along with some XML data about what they apply to.
But according to the linked article WU does send MS server a list of installed hardware; from this the server sends WU a list of applicable driver updates for that hardware.
In either case, if the update server wanted to track what software you have installed, it's still a trivial matter to parse the server-side logs of what was downloaded. Even though the initial apt-get update only pulls a list, after that you still have to download the actual updates your system needs. If someone downloads the update for SuperFoobar2001.3, it's a good deduction that this someone (at IP address 123.4.5.6 or whatever) has SuperFoobar2001.3 installed on his computer. Drivers too - a user is unlikely to download drivers via apt-get for hardware that's not installed.
What's the difference here? Either update mechanism can learn what you have on the PC. The WU setup does give MS one bit of info that the APT system doesn't pass: the unique Product Information number (PID) for that installation of Windows. ow that we know this data is collectable (but not necessarily saved or analyzed), the real question is: what are they doing with that data?
Well, I did read the linked article. They claim that Windows Update (WU) uploads a complete list of installed hardware to the MS server; and the server then sends WU a list of applicable updates for that hardware. They also claim (with less certainty) that the product identification key and a signed hash of that key are sent to Microsoft as a way of potentially denying updates to pirated copies of Windows.
These are possibly reasons for concern, but just to be clear they are a far cry from the upload everything!!!oneoneone!!! approach claimed by the grandparent post. Keep in mind that at the end of the day, any automatic update server (Windows, Ubuntu, insert your OS) can learn a lot about what's installed in the system being updated, if only by analyzing what gets downloaded. Or would we all be better served by an automatic update system which always downloaded every available update whether it was needed or not?
It's not actually 'right to make money'. It's 'right to set their own price and terms'.
In the case where they are giving you a $200 phone for "free", they get you to sign a contract for a number of months or years so that they can earn back their cash outlay for the "free" phone. They don't have a 'right to make money'... they could write the contract foolishly, leaving you a way to keep the phone without their being able to recoup costs. But they'd soon be out of business if they did, so they learn not to write foolish contracts.
They certainly do have a right to make their delivery of phone + service conditional on your entrance into a contract with them (as long as nothing in the contract is illegal). Just as you have a right to say no to the whole deal.
You've just rewritten the news. MS did not laugh; indeed they were the first to report that one of their employees had gone out of bounds with email promises. Literally within hours they contacted the two partners who had been sent the improper emails, saying basically 'disregared that, it's wrong'. If the Evil MS Borg had these companies so completely enthralled as some have suggested, it would have ended there and no one would be the wiser. But no... it was Microsoft who contacted the Swedish Institute of Standards to explain that an impropriety had occurred. SIS then abstained from the vote.
Through MS's own, conscious actions, they ended up losing Sweden's vote. Hardly something to laugh openly about.
It's sad that Gateway provides no cheap media replacement procedure - *against* MS recommended policy. Were Gateway to go out of business, MS would replace the disk for $30: http://support.microsoft.com/kb/326246#OEM
If I were you, I'd go ahead and install it, then. If it takes you to phone activation, tell 'em your computer was stolen, feeling free to provide the case number your local law enforcement gave you. betcha a buck they allow the activation.
As for file associations, there's no reason for this to be inaccessible by users. If I want to open.jpg images with "mirage" instead of "kview" by default, why should I not be able to set that? This is an issue purely about user preferences, just like what I want my screen saver and desktop background to be. How would "security vulnerabilities" have anything to do with this?
Because (in Windows at least) the user is changing system-wide file associations, not just his own. So, if the next user doubleclicks on foo.doc and it is passed as an argument to filedeleter.exe (just to make up a hideous but possible example), that does have security and usability repercussions. Screensaver and desktop backround are per-user changes, so in that case it's not an issue. However the sysadmin should still have the ability to override user choice of screensaver/desktop backround, for whatever reasons the organization deems appropriate.
"When I first started my experiment I was trying to keep it a secret out of fear of attacks from angry Microsoft worshipers (especially from the admins and desktop support). What I am finding out is that most of the folks that I was hiding from are sick and tired of supporting Windows and are proponents of Linux."
Granted my own experience is anecdotal, but this completely unsurprising. I wouldn't go so far as to say that most Windows-using folk are tired of their OS and actively seeking or evangelizing another, though. They mostly just don't really have a big interest in what the other guy is using.
Within the Linux communities it is pretty hard to miss the hatred and derision of Microsoft - it's everywhere. And so the Linux folk seem to assume that there will be a corresponding hatred of all things Linux in the MS-using communities. But by and large, that Linux-hatred just doesn't exist in the MS-using communities. Oh, you can dig up your 10 year old Halloween Documents, or you can violently twist the wording for a 'Get The Facts' whitepaper or something from the ODF-OOXML skirmishes. And once in awhile you'll find an anti-Linux zealot who loves Microsoft - but these guys are few and far between. What you won't find is a plethora of BadLinux sites to mirror BadVista.org, or an anti-Slashdot, where hundreds of thousands of Windows users go to regularly slag on Linux. Or a gazillion "Linux sucks and here is why!" posts from the MS-lovers in response to any blog post which even faintly praises anything from the F/OSS world.
By and large most Windows users have a pretty mild reaction to F/OSS software in general, or Linux in particular. Mention you're using Linux, and the average Windows-using person will shrug, or express mild interesst... then turn back to whatever they were doing before. It's just not a big deal. These guys aren't about to bend over backwards to learn Linux, but by the same token, they aren't interested enough to need to convince anyone of anything. Even the sysadmin who may reimage your work system back to Windows (from the Linux you had installed) isn't doing so from personal conviction; he's doing it because that's company policy.
In my experience most Windows-using folk are a little nonplussed by the defensiveness and anti-MS rants/jokes/propaganda coming from the Linux camp. They don't see any reason to get emotional over some software, unless the software isn't performing as expected.
I don't mind the element of DRM which tries to enforce payment for copyrighted content. I'm happy to look at the price of whatever content I am interested in, and make the decision to buy or not buy. If the price is too high, I won't buy and I won't even attempt to circumvent.
But, the element of DRM which makes it difficult or impossible for me to move that content onto other devices I own (now or in the future) - THAT is the problem these days. I want my digital copy of the content to be at least as durable and portable as an analog tape or record used to be.
Now, please note that I said *move* that content from device to device, not *copy* it. The difference is that a copy leaves behind an instance of the content, and a move does not. The copying is the problem for the folks who need to make enough money to support the work they put into making this (and future) content - if I can simply copy stuff, then I can give it to friends who didn't pay for it (And still enjoy it myself). But if I have to move it from device to device, such that I can only experience it from a device which is physically near me, then the 'giving a copy to my friend who didn't pay' problem is eliminated. I should still be able to give/loan it to my friend, but only on the condition that once I give it away, I don't have it anymore.
It feels a little silly writing this out, but the endless ranting over DRM convinces me that quite a few people have forgotten these fundamentals. But I would be satisfied with a universal DRM scheme that allowed me to easily *move* my paid content from device to device.
But things like indentation, bullets, numbered lists, etc... these can change the reader's understanding of what was written. You can't just write that stuff off as if it has no value.
The article definitely notes the need for perfect emulation, because of the need to perfectly convert all old data into the new format, and do so in a batchprocessing way, without requiring humans to touch up the results. But there's no reason you couldn't add more stuff into the format once it was capable of full emulation and automated translation from other formats.
The citizen does not have to pay an MS tax. As noted over and over in these debates, there are free-to-download viewers for Word, Excel, Powerpoint, Visio files (and probably others; these are just the ones I have used). Admittedly they only work on Windows; the Linux solution is to open those files in OO, though you will sometimes get misformatted data as a result.
So far I have gotten through about 75% of the comments, and it seems that everyone is willing to vent their personal experiences and vitriol, but no one is addressing the main point of the LinuxWorld article: the fact that if an entity wishes to change its data formats today, it will need to convert all of its legacy data into the new format. And it will need to do this in an automated way, with zero transcription errors. Rip and replace isn't going to get it done.
Goddamn, will people stop saying this?!? I've formatted a 200GB hard drive as FAT with the Windows XP installer. There is no 32GB limit.
OK, here is the Real Deal:
FAT itself can be up to 2 terabytes in size. FAT32: 2TB (theoretically 8 TB) FAT16: 4GB FAT12: 16MB.
Large FAT partitions can be hugely wasteful of disk space, because FAT has a limited number of possible entries in the file allocation table itself, and therefore must use ever-larger cluster sizes (think extents) for file storage if you wish to have a large partition. Much disk space is lost to the many resulting partially-filled clusters. We used to call it 'slack'.
The 32 GB limit (which MS admits is arbitrary) was imposed in the GUI partitioning tool in Windows 2000, and has persisted since. This only applies to partitions created with that formatter; W2000 and above will happily use much larger partitions.
But the gotcha is that if your FAT filesystem is larger than about 124 gigabytes and it breaks, you will not be able to fix it. Scandisk is the repair tool for FAT filesystems, and it simply cannot process a partition larger than 124.5GB.
If you want to create a >32GB FAT partition from within the Windows GUI, you can use fat32format.exe.
There are a few ways to attack this, but the most common is always going to be security and honestly, the biggest flaw for Windows (and most apps that use it) are the requirement for Administrator access for users.
Except that's not true; it never really was. I have been running Windows and all my daily use applications as nonadmin since the mid-90's on NT4. I also ran as nonadmin on Windows 2000 and XP; now I run as nonadmin on Vista. Granted, it hasn't always been as easy as falling off a log, since some apps are poorly written and tested, not following the MS guidelines which were there all along. Those guidelines are/were easy to summarize:
1) Apps should install to Program Files (Admin privs needed for install).
2) once the app is installed, it should do file writes to the current user's Documents and Settings directory only; not to Program Files
3) During normal program use, registry writes to the HKEY_CURRENT_USER hive (not HKEY_LOCAL_MACHINE).
Some programs (and not as many of them as you might think) broke these guidelines. The easy way to solve the problem was to just run as Admin; the harder way was to break out your copies of FileMon and RegMon, find out which specific folders/regkeys the program was writing to in violation of guideline, and fix the permissions. Yes this is geeky and painful, but once you've done it a few times, it gets easy, and rarely takes more than 20 minutes per program. Whenever I did it, I'd take a few more minutes to document the permissions changes I had to make, and send these off to the program's authors along with a copy of the guidelines. This accomplished two things for me: first I'd have some documentation of the issue in case I needed to do it for someone else, and there was a (slim) chance the coders would fix it in their next revision.
The root of all this traces back to a fairly simple and stupid thing. NT4 Workstation didn't sell that well. Users graduating to NT4 from the Win9x series found that a secure OS was very painful! So some bright bulb over at MS implemented a quick fix in Windows 2000: during installation (and only during installation) any new user you created would automtatically be made a member of the Administrators group. Worse, this happened without any sort of notification to the user. He never knew he'd just bypassed all security features of the OS. So, a lot of coding and testing took place in the Administrator context - bypassing all security. Those programs now depended on having Admin privs. Then more and more people went to 'always on' internet connections while running as Admin all the time, and Windows' reputation as an insecure platform was inevitable.
Vista attempts to make up for all this lost ground with several tricks, but really the big one is UAC. The point of UAC is simply to make it easier for people to run nonadmin all the time, only elevating privs when they have to. Sadly, the bad habit of invisibly making Admins out of all users created during installation has still not been kicked, but I have my hopes up for Seven. (sigh)
Try to run something that uses 100% CPU and then try to do anything else while that happens. What a great scheduler...
Works fine here. In basic concept, Windows' scheduler (priority values) isn't so different from the Linux scheduler (nice values). Obviously any system starts to bog when it has a high Load Average (Linux) or CPU Queue Length (Windows).
Also, try to fill up your RAM. Kind of hard, isn't it? Windows doesn't seem to think you have as much RAM as you do and starts to swap far too early to be considered useful. This is why people complain about Firefox using $x amount of RAM; Windows starts to swap way too early and causes slowdowns all around.
This comment shows an almost total lack of understanding of the Windows memory model. See any of the Inside Windows or Windows Internals books by Russinovich and Solomon for definitive reference; for a shorter slam-bang course, have a look at this Understanding Virtual Memory article. Pay close attention to the concept of the backing store. I took a quick look at an XP system with 2GB of RAM: 92% of memory in use. And a Vista system with 1GB RAM: 100% memory in use.
Try to delete a file that's in use (something you can do in any Unix-like system). File in use? Whoops, can't do that.
Agreed; this sucks. You can mitigate somewhat with utilities like MoveFile or Process Explorer, but again, I agree. Tracking down the process that locked your file, or scheduling a reboot for the rename or delete operation, is a little too baroque for my taste!
Also, Windows has jack shit support for more filesystems than their own FAT and NTFS families (both of which get fragmented; modern filesystems prevent that on the fly). Sure, you can get more support via plugins (I believe there are two different ways to make a filesystem plugin for Windows: kernel and shell), but that isn't as reliable as having native support for them.
You've noted that plugins exist. Which is how most new filesystem support for Linux originally evolved - either you had to manually compile it into the kernel (until Linus decided to just do that for you) or run a FUSE FS. MS doesn't build {otherFS} support into any Windows 'distro', but that's the only step missing. If you want {otherFS} support in Windows, you find or write it, then install that to your system. With IFS builtin, this can be just as reliable as native. In fact, NTFS/FAT are loaded via IFS. So if your {otherFS} via IFS is less reliable, that's on the {otherFS} coders, not MS.
I have to comment. Somewhere around here, I have a letter from Virginia Heinlein.
Sometime in the mid-90's I managed to lay hands on one of Mr. Heinlein's few nonfiction works, a book called 'Take Back Your Government!' in which he talks quite a bit about how to be politically active in your community. He drew on a lot of his own activist experiences. Anyhow, I wanted to be able to quote passages from the book in an online forum (WWIV BBS network, if anyone cares).
Virginia wrote back, graciously granting permission, but also taking a bit of time to note that Mr. Heinlein's views had changed significantly since that book, and that 'The Moon Is A Harsh Mistress' (Prof's views there especially) probably best captures the political sentiments Mr. Heinlein had in his latter years.
I would characterize that book as leaning hard libertarian, with the realistic understanding that the masses prefer bureaucratised bread and circuses.
No. That's the whole point of software-defined radio (SDR)... you have a general purpose receiver or transmitter/receiver, and the software defines what frequency it listens and/or radiates at.
Slashdot Mirror
I was hoping you might be able to give examples. You know ... name a few of them.
The current system of more and more data collecting isn't for more security. That's just how it's sold. It is, bluntly, control. Over your data and you. It is easier to pinpoint and neutralize "troublemakers" before they start gaining a lot of support.
Who are these troublemakers being pinpointed and neutralized? What sort of trouble are they making? How is the pinpointing and neutralizing being accomplished? Ever known anyone who was pinpointed and neutralized? And why haven't they (the ominous 'they') pinpointed and neutralized you yet - for posting things like this?
The article is 48 paragraphs long. MS/Gates are mentioned in 3 paragraphs. Intel is mentioned in more than 15 paragraphs, and their bad effect on OLPC is singled out by Negroponte multiple times. Yet most of the /. replies single out MS as the bad guy here - for cutting price on their software. Hardly a mention of Intel, who have done far more to damage OLPCs chances in the market.
Hmm. Slashdot reading comprehension in top form, eh?
Way to go on that security model thing, Microsoft!
That security model thing doesn't work any better when you bypass it by being Administrator on Windows, than when you bypass it by being root on Linux.
Please stop trying to fit me into any of the pro- or anti-MS pigeonholes.
None of my post cared specifically who has the data. Microsoft, Canonical, God, the Devil - that wasn't the point. The point was, whomever you get your updates from can know what software and hardware you have. The question isn't what they can know, it's whether they bother to do anything with the knowledge.
You've misinterpreted my response. I have no desire to FUD for MS, only to clarify the technical picture. A more succinct version of what I said:
- Windows Update works in a similar way to the Ubuntu update you are championing. It downloads a list of all updates along with some XML data about what they apply to.
- But according to the linked article WU does send MS server a list of installed hardware; from this the server sends WU a list of applicable driver updates for that hardware.
- In either case, if the update server wanted to track what software you have installed, it's still a trivial matter to parse the server-side logs of what was downloaded. Even though the initial apt-get update only pulls a list, after that you still have to download the actual updates your system needs. If someone downloads the update for SuperFoobar2001.3, it's a good deduction that this someone (at IP address 123.4.5.6 or whatever) has SuperFoobar2001.3 installed on his computer. Drivers too - a user is unlikely to download drivers via apt-get for hardware that's not installed.
What's the difference here? Either update mechanism can learn what you have on the PC. The WU setup does give MS one bit of info that the APT system doesn't pass: the unique Product Information number (PID) for that installation of Windows. ow that we know this data is collectable (but not necessarily saved or analyzed), the real question is: what are they doing with that data?Well, I did read the linked article. They claim that Windows Update (WU) uploads a complete list of installed hardware to the MS server; and the server then sends WU a list of applicable updates for that hardware. They also claim (with less certainty) that the product identification key and a signed hash of that key are sent to Microsoft as a way of potentially denying updates to pirated copies of Windows.
These are possibly reasons for concern, but just to be clear they are a far cry from the upload everything!!!oneoneone!!! approach claimed by the grandparent post. Keep in mind that at the end of the day, any automatic update server (Windows, Ubuntu, insert your OS) can learn a lot about what's installed in the system being updated, if only by analyzing what gets downloaded. Or would we all be better served by an automatic update system which always downloaded every available update whether it was needed or not?
It's not actually 'right to make money'. It's 'right to set their own price and terms'.
... they could write the contract foolishly, leaving you a way to keep the phone without their being able to recoup costs. But they'd soon be out of business if they did, so they learn not to write foolish contracts.
In the case where they are giving you a $200 phone for "free", they get you to sign a contract for a number of months or years so that they can earn back their cash outlay for the "free" phone. They don't have a 'right to make money'
They certainly do have a right to make their delivery of phone + service conditional on your entrance into a contract with them (as long as nothing in the contract is illegal). Just as you have a right to say no to the whole deal.
You've just rewritten the news. MS did not laugh; indeed they were the first to report that one of their employees had gone out of bounds with email promises. Literally within hours they contacted the two partners who had been sent the improper emails, saying basically 'disregared that, it's wrong'. If the Evil MS Borg had these companies so completely enthralled as some have suggested, it would have ended there and no one would be the wiser. But no ... it was Microsoft who contacted the Swedish Institute of Standards to explain that an impropriety had occurred. SIS then abstained from the vote.
Through MS's own, conscious actions, they ended up losing Sweden's vote. Hardly something to laugh openly about.
It's sad that Gateway provides no cheap media replacement procedure - *against* MS recommended policy. Were Gateway to go out of business, MS would replace the disk for $30: http://support.microsoft.com/kb/326246#OEM
If I were you, I'd go ahead and install it, then. If it takes you to phone activation, tell 'em your computer was stolen, feeling free to provide the case number your local law enforcement gave you. betcha a buck they allow the activation.
If you bought the retail version of XP, you can install on another system, then. OEM copies are 'locked' to the first computer they are installed on.
... knew that, right? Or were you trying to mislead people here?
You
As for file associations, there's no reason for this to be inaccessible by users. If I want to open .jpg images with "mirage" instead of "kview" by default, why should I not be able to set that? This is an issue purely about user preferences, just like what I want my screen saver and desktop background to be. How would "security vulnerabilities" have anything to do with this?
Because (in Windows at least) the user is changing system-wide file associations, not just his own. So, if the next user doubleclicks on foo.doc and it is passed as an argument to filedeleter.exe (just to make up a hideous but possible example), that does have security and usability repercussions. Screensaver and desktop backround are per-user changes, so in that case it's not an issue. However the sysadmin should still have the ability to override user choice of screensaver/desktop backround, for whatever reasons the organization deems appropriate.
"When I first started my experiment I was trying to keep it a secret out of fear of attacks from angry Microsoft worshipers (especially from the admins and desktop support). What I am finding out is that most of the folks that I was hiding from are sick and tired of supporting Windows and are proponents of Linux."
Granted my own experience is anecdotal, but this completely unsurprising. I wouldn't go so far as to say that most Windows-using folk are tired of their OS and actively seeking or evangelizing another, though. They mostly just don't really have a big interest in what the other guy is using.
Within the Linux communities it is pretty hard to miss the hatred and derision of Microsoft - it's everywhere. And so the Linux folk seem to assume that there will be a corresponding hatred of all things Linux in the MS-using communities. But by and large, that Linux-hatred just doesn't exist in the MS-using communities. Oh, you can dig up your 10 year old Halloween Documents, or you can violently twist the wording for a 'Get The Facts' whitepaper or something from the ODF-OOXML skirmishes. And once in awhile you'll find an anti-Linux zealot who loves Microsoft - but these guys are few and far between. What you won't find is a plethora of BadLinux sites to mirror BadVista.org, or an anti-Slashdot, where hundreds of thousands of Windows users go to regularly slag on Linux. Or a gazillion "Linux sucks and here is why!" posts from the MS-lovers in response to any blog post which even faintly praises anything from the F/OSS world.
By and large most Windows users have a pretty mild reaction to F/OSS software in general, or Linux in particular. Mention you're using Linux, and the average Windows-using person will shrug, or express mild interesst ... then turn back to whatever they were doing before. It's just not a big deal. These guys aren't about to bend over backwards to learn Linux, but by the same token, they aren't interested enough to need to convince anyone of anything. Even the sysadmin who may reimage your work system back to Windows (from the Linux you had installed) isn't doing so from personal conviction; he's doing it because that's company policy.
In my experience most Windows-using folk are a little nonplussed by the defensiveness and anti-MS rants/jokes/propaganda coming from the Linux camp. They don't see any reason to get emotional over some software, unless the software isn't performing as expected.
One word: WINE.
Hear hear - I hope this comment gets modded up.
I don't mind the element of DRM which tries to enforce payment for copyrighted content. I'm happy to look at the price of whatever content I am interested in, and make the decision to buy or not buy. If the price is too high, I won't buy and I won't even attempt to circumvent.
But, the element of DRM which makes it difficult or impossible for me to move that content onto other devices I own (now or in the future) - THAT is the problem these days. I want my digital copy of the content to be at least as durable and portable as an analog tape or record used to be.
Now, please note that I said *move* that content from device to device, not *copy* it. The difference is that a copy leaves behind an instance of the content, and a move does not. The copying is the problem for the folks who need to make enough money to support the work they put into making this (and future) content - if I can simply copy stuff, then I can give it to friends who didn't pay for it (And still enjoy it myself). But if I have to move it from device to device, such that I can only experience it from a device which is physically near me, then the 'giving a copy to my friend who didn't pay' problem is eliminated. I should still be able to give/loan it to my friend, but only on the condition that once I give it away, I don't have it anymore.
It feels a little silly writing this out, but the endless ranting over DRM convinces me that quite a few people have forgotten these fundamentals. But I would be satisfied with a universal DRM scheme that allowed me to easily *move* my paid content from device to device.
But things like indentation, bullets, numbered lists, etc ... these can change the reader's understanding of what was written. You can't just write that stuff off as if it has no value.
The article definitely notes the need for perfect emulation, because of the need to perfectly convert all old data into the new format, and do so in a batchprocessing way, without requiring humans to touch up the results. But there's no reason you couldn't add more stuff into the format once it was capable of full emulation and automated translation from other formats.
The citizen does not have to pay an MS tax. As noted over and over in these debates, there are free-to-download viewers for Word, Excel, Powerpoint, Visio files (and probably others; these are just the ones I have used). Admittedly they only work on Windows; the Linux solution is to open those files in OO, though you will sometimes get misformatted data as a result.
So far I have gotten through about 75% of the comments, and it seems that everyone is willing to vent their personal experiences and vitriol, but no one is addressing the main point of the LinuxWorld article: the fact that if an entity wishes to change its data formats today, it will need to convert all of its legacy data into the new format. And it will need to do this in an automated way, with zero transcription errors. Rip and replace isn't going to get it done.
Goddamn, will people stop saying this?!? I've formatted a 200GB hard drive as FAT with the Windows XP installer. There is no 32GB limit.
OK, here is the Real Deal:
Other references: Limitations of the FAT32 File System, Raymond Chen, NTFS vs FAT.
There are a few ways to attack this, but the most common is always going to be security and honestly, the biggest flaw for Windows (and most apps that use it) are the requirement for Administrator access for users.
Except that's not true; it never really was. I have been running Windows and all my daily use applications as nonadmin since the mid-90's on NT4. I also ran as nonadmin on Windows 2000 and XP; now I run as nonadmin on Vista. Granted, it hasn't always been as easy as falling off a log, since some apps are poorly written and tested, not following the MS guidelines which were there all along. Those guidelines are/were easy to summarize:
1) Apps should install to Program Files (Admin privs needed for install).
2) once the app is installed, it should do file writes to the current user's Documents and Settings directory only; not to Program Files
3) During normal program use, registry writes to the HKEY_CURRENT_USER hive (not HKEY_LOCAL_MACHINE).
Some programs (and not as many of them as you might think) broke these guidelines. The easy way to solve the problem was to just run as Admin; the harder way was to break out your copies of FileMon and RegMon, find out which specific folders/regkeys the program was writing to in violation of guideline, and fix the permissions. Yes this is geeky and painful, but once you've done it a few times, it gets easy, and rarely takes more than 20 minutes per program. Whenever I did it, I'd take a few more minutes to document the permissions changes I had to make, and send these off to the program's authors along with a copy of the guidelines. This accomplished two things for me: first I'd have some documentation of the issue in case I needed to do it for someone else, and there was a (slim) chance the coders would fix it in their next revision.
The root of all this traces back to a fairly simple and stupid thing. NT4 Workstation didn't sell that well. Users graduating to NT4 from the Win9x series found that a secure OS was very painful! So some bright bulb over at MS implemented a quick fix in Windows 2000: during installation (and only during installation) any new user you created would automtatically be made a member of the Administrators group. Worse, this happened without any sort of notification to the user. He never knew he'd just bypassed all security features of the OS. So, a lot of coding and testing took place in the Administrator context - bypassing all security. Those programs now depended on having Admin privs. Then more and more people went to 'always on' internet connections while running as Admin all the time, and Windows' reputation as an insecure platform was inevitable.
Vista attempts to make up for all this lost ground with several tricks, but really the big one is UAC. The point of UAC is simply to make it easier for people to run nonadmin all the time, only elevating privs when they have to. Sadly, the bad habit of invisibly making Admins out of all users created during installation has still not been kicked, but I have my hopes up for Seven. (sigh)
Awesome response. I hope more people mod this up.
Negativity sucks.Try to run something that uses 100% CPU and then try to do anything else while that happens. What a great scheduler...
Works fine here. In basic concept, Windows' scheduler (priority values) isn't so different from the Linux scheduler (nice values). Obviously any system starts to bog when it has a high Load Average (Linux) or CPU Queue Length (Windows).
Also, try to fill up your RAM. Kind of hard, isn't it? Windows doesn't seem to think you have as much RAM as you do and starts to swap far too early to be considered useful. This is why people complain about Firefox using $x amount of RAM; Windows starts to swap way too early and causes slowdowns all around.
This comment shows an almost total lack of understanding of the Windows memory model. See any of the Inside Windows or Windows Internals books by Russinovich and Solomon for definitive reference; for a shorter slam-bang course, have a look at this Understanding Virtual Memory article. Pay close attention to the concept of the backing store. I took a quick look at an XP system with 2GB of RAM: 92% of memory in use. And a Vista system with 1GB RAM: 100% memory in use.
Try to delete a file that's in use (something you can do in any Unix-like system). File in use? Whoops, can't do that.
Agreed; this sucks. You can mitigate somewhat with utilities like MoveFile or Process Explorer, but again, I agree. Tracking down the process that locked your file, or scheduling a reboot for the rename or delete operation, is a little too baroque for my taste!
Also, Windows has jack shit support for more filesystems than their own FAT and NTFS families (both of which get fragmented; modern filesystems prevent that on the fly). Sure, you can get more support via plugins (I believe there are two different ways to make a filesystem plugin for Windows: kernel and shell), but that isn't as reliable as having native support for them.
You've noted that plugins exist. Which is how most new filesystem support for Linux originally evolved - either you had to manually compile it into the kernel (until Linus decided to just do that for you) or run a FUSE FS. MS doesn't build {otherFS} support into any Windows 'distro', but that's the only step missing. If you want {otherFS} support in Windows, you find or write it, then install that to your system. With IFS builtin, this can be just as reliable as native. In fact, NTFS/FAT are loaded via IFS. So if your {otherFS} via IFS is less reliable, that's on the {otherFS} coders, not MS.
I have to comment. Somewhere around here, I have a letter from Virginia Heinlein.
Sometime in the mid-90's I managed to lay hands on one of Mr. Heinlein's few nonfiction works, a book called 'Take Back Your Government!' in which he talks quite a bit about how to be politically active in your community. He drew on a lot of his own activist experiences. Anyhow, I wanted to be able to quote passages from the book in an online forum (WWIV BBS network, if anyone cares).
Virginia wrote back, graciously granting permission, but also taking a bit of time to note that Mr. Heinlein's views had changed significantly since that book, and that 'The Moon Is A Harsh Mistress' (Prof's views there especially) probably best captures the political sentiments Mr. Heinlein had in his latter years.
I would characterize that book as leaning hard libertarian, with the realistic understanding that the masses prefer bureaucratised bread and circuses.No. That's the whole point of software-defined radio (SDR) ... you have a general purpose receiver or transmitter/receiver, and the software defines what frequency it listens and/or radiates at.