LOL, god that's typical. Someone drank the Kool-Aid.:)
Seriously, Linux in those days WAS amazingly robust, for the most part. I bet, once you got it installed, it was reliable as a tank. The kernels up through the 2.2 series were some of the finest pieces of software I've ever run. Things have been going downhill ever since 2.4. That series didn't stabilize until about 2.4.11, and finally got GOOD by.20 or so. 2.6 has never been that stable...they let ridiculously bad bugs out now. (2.6.14 broke traceroute... and 2.6.15 won't stay up an hour on my Intel 865 systems. 2.6.16 seems to work fine -- finally -- but 2.6.15 was garbage.)
It's a good things filesystems have improved so much. I have been testing their crash recovery a great deal over the last year or so.
Yeah, I guess I just expect better from Linux people. They're smart enough to write an amazing system and a huge universe of free software. I suppose the hypersensitive ones may not be the same crowd that's actually writing the software... it's easy to lump disparate groups of people using Linux into the same 'community' and treat them as a whole.
There are lots of great people in the Linux community, but the 'blame the user!' contingent really sucks.
I don't have to admit any such thing. We were a complete NT 4.0 shop. NTFS wasn't as good then as it is now, but it was still a zillion times as robust as ext2. It had to be, because the damn boxes bluescreened all the frickin' time. I don't think I ever lost data from a power failure or system crash. We lost data from developer mistakes, but that's what the backups were for.
EXT2 SUCKED. Pure and simple. Lose power, and there was a high chance you'd lose files. It was eggshell-fragile to hardware failure. Linux could get away with that, because the OS was so incredibly robust at the time that it _never_ crashed. Nowadays, the filesystems are great... the OS itself... not so much.
You realize... you are doing the exact same thing, right? You're blaming the user again. A good operating system survives a power loss. I have taken significant filesystem damage due to OS crashes/power failures from only two sources: one with the game Darklands under DOS and QEMM (wiped my drive), and several times from Linux running ext2. Hint: that means ext2 sucks. Since the advent of journaled filesystems, if I've lost anything, I haven't noticed. Hint: that means ext2 sucks.
And as long as we're genitalia-waving, I was using Linux sometime in late '92 or early '93... my first install was SLS. Mind you, I didn't have a fucking clue what to DO with it, being a DOS guy with no Unix people around to learn from, but under Slackware I managed to figure out modelines and get X going. And I only blew up one monitor doing it.:)
It was many years before Linux was really business-suitable... mostly because it took businesses awhile to catch up. I had to push really hard to get those DNS boxes in that early. Nowadays, it's an easy decision, but at the time it felt very risky and weird to be running free software in the workplace. It WAS nine years ago... a lot has changed.
This has been true for as long as I can remember. If the software is inadequate or confusing, blame the user. It's happened to me, even here on Slashdot.
Way back in the stone age.... sometime in 1997, maybe? Maybe 1998. Not sure. Anyway... Linux was _just_ starting to get deployed occasionally in business. I had a couple of DNS servers up on an early RedHat box. The box lost power... these were just desktop machines sitting in someone's cubicle. (We hadn't grown to the point of needing 'real' servers quite yet, and an actual server room was a year off.) My primary box took a LOT of filesystem damage, and it took me ages to fix it. So I commented in a slashdot thread that ext2 was very fragile, and that it was one of Linux's real weak points.
You just wouldn't believe the crap I got. Slashdot doesn't seem to archive that far back, so I can't give you links, but _most_ of the replies I got blamed me for being stupid. I "should have used a UPS"... ok, I could grant that, but remember we were a shoestring outfit, and we didn't need those on Windows servers. A couple people went off on me for, get this, not knowing how to use a disk editor to find my secondary superblocks and repair with those. I kid you not. Linux was perfect, and ANYTHING that went wrong was obviously the user's fault... to the point that I should know how to manually repair my filesystem. Instead of admitting that the filesystem should survive a power failure, it was my fault for breaking it.
Several years later, after Reiserfs and ext3 came out, we had a similar conversation, also here on Slashdot. Suddenly everyone is all about how great the journaling filesystems are, and how bad ext2 sucks. It was probably even some of the same people, but the original conversation had already been lost, so I couldn't prove it.
People just will NOT criticize software they're emotionally involved with. It's the most ridiculous thing I've seen... these theoretically intelligent, rational software designers that become absolutely insane when you suggest their software is imperfect. Blame the user! "You're just too stupid to use our software. Go away."
Fortunately, there's enough people in the Linux community now that the lunatic fringe doesn't dominate quite like it did, but these people are still out there.
It was ridiculous then, and it's just as ridiculous now. It doesn't happen as much, but it still sucks.
Hardware virtualization is not a 'bandaid'. In fact, it's the only proper way to do it. Software virtualization is the band-aid.
I suspect you're going the wrong way; with the hardware support in VT and Pacifica, Xen and VMWare are going to get a lot faster. It won't be that hard for them to add in features like what you have (particularly for Xen, which is integrated in the Linux kernel), but it will be very hard for you to do what they're doing.
Specialized hardware support is always faster than general-purpose software. While it may not be all the way there yet in the first generation of hardware, ultimately, solutions that take advantage of hardware to do virtualization will run faster and be much more secure than any software-level solution.
While what you're doing is very useful in the short run, I think you will eventually need to adapt to the hardware model, or die.
The submitter didn't point out that if AT&T is claiming the documents are trade secrets, that means they're accurate. Made-up documents wouldn't be trade secrets.
In other words, AT&T has just admitted that they are spying on you.
Yeah, I was thinking the same thing. This article would have been interesting, say, 18 months ago... but with VT and Pacifica, things are different now. Without at least mentioning those, it's not very useful.
Anyone have a pointer to a good writeup on the differences between VT, Pacifica, and regular old software virtualization?
See, you're not the problem; you're a computer professional and, at least in theory, you should be highly expert at using a PC. The problem is Tracy in Accounting and Bob the Receptionist, who haven't a clue what's going on with their machines, and who happily install spyware if it promises something slightly better than a sharp stick in the eye.
Think of it as the "OMG Ponies!" crowd, writ large. You just have no idea how freaking stupid these people can be.
Even in the best and brightest companies I've worked in, there have always been a few that got hired that knew a lot less about their PCs than they thought. In particular, they do not appear to hire salespeople for raw brainpower. The clueless users, especially the ones that don't realize (and never will) that they ARE clueless, cause enormous trouble. Unless the network is internally firewalled (which is getting to be a better and better idea, these days), they're often the vectors for network-wide infection.
The draconian policies of some admins may seem stupid, but remember that admins run on fear. They are, by and large, only noticed when things break, and then everyone is mad at them. When a single user can potentially bring a virus into the network that can stop the entire company dead in its tracks, well... it's a heck of a lot safer and easier to just lock EVERYTHING down and then install what people need, as they ask for it.
They do it that way, and then charge weird amounts of points for things in an attempt to either:
A) Have the use of your money for free as long as you have unspent points, or; B) Sell you something completely useless at an insane profit margin, so that you've spent all your points.
Both benefit Microsoft enormously. I'm sure I will buy very, very few things off XBL.
With the six degrees of separation thing, and careful choices of 'suspects', they can probably get a 90% surveillance rate by declaring only a couple of thousand primary targets.
In other words, as far as I can see, the Australian Parliament has just decreed that the government can read all the email it likes, whenever it likes.
If I were in charge, and unscrupulous, the first person I'd declare a suspect would be the chief of the opposition party.
My apologies, Slashdot editors. I must have missed it. It looks like they ran it even before I got the announcement email from VMWare, so of course they'd have refused my submission. Doh.
Anyway, it IS good software. Wish it were truly free, and I'm looking forward to what happens with Xen, but free-as-in-beer will do for now.:)
Note that VMWare is also giving away their Server product for free. For some reason, Slashdot hasn't been willing to run this story, even though it's important.
It's a new product, still in beta... about equivalent to the GSX Server. They just released Beta 2 either today or yesterday. It's a _really_ good product. The current keys they're giving away expire, but they say the final version will also be free-as-in-beer.
Basically, it'll do everything Workstation will, plus it allows you to see the consoles of virtual machines that are on another computer. It also gives you a fairly rudimentary web-based control panel, wherein you can start, stop, or restart particular VMs. You can also set up user accounts, and restrict access to particular machines appropriately. It's not ISP-class, but it'd be damn useful for QA teams or suchlike.
It'll never happen, stop asking. CmdrTaco has explicitly said, very recently, that he won't even fix spelling mistakes. He's afraid, get this, that Slashdot will look too professional.
The definition of professional is 'not amateur'... ie, you're getting paid for it. So they want to gete paid for what they're doing, but not do the actual work of doing it right.
I haven't had this place as my homepage in years, and that kind of comment just confirms the decision. I much prefer Ars Technica; they don't have the volume that Slashdot does, but their articles are _really_ good. A mix of digg and Ars does a better job than Slashdot... digg handles the "cursory overview by twits" part, and then there's Ars for in-depth, actual reporting. They even use good grammar, spell things correctly, and get their facts right. It's quite refreshing.
The Slashdot approach worked when they were all 18 and fresh out of school, but I expect a bit more out of adults.
I know gcc itself improved a very great deal over the same time period, and I have always assumed that the speed gains were (largely? mostly?) due to that, rather than wondrous new algorithms on Apple's part.
Linux and KDE sped up a lot too, over the same timeframe.
Slashdot Mirror
LOL, god that's typical. Someone drank the Kool-Aid. :)
.20 or so. 2.6 has never been that stable...they let ridiculously bad bugs out now. (2.6.14 broke traceroute... and 2.6.15 won't stay up an hour on my Intel 865 systems. 2.6.16 seems to work fine -- finally -- but 2.6.15 was garbage.)
Seriously, Linux in those days WAS amazingly robust, for the most part. I bet, once you got it installed, it was reliable as a tank. The kernels up through the 2.2 series were some of the finest pieces of software I've ever run. Things have been going downhill ever since 2.4. That series didn't stabilize until about 2.4.11, and finally got GOOD by
It's a good things filesystems have improved so much. I have been testing their crash recovery a great deal over the last year or so.
Yeah, I guess I just expect better from Linux people. They're smart enough to write an amazing system and a huge universe of free software. I suppose the hypersensitive ones may not be the same crowd that's actually writing the software... it's easy to lump disparate groups of people using Linux into the same 'community' and treat them as a whole.
There are lots of great people in the Linux community, but the 'blame the user!' contingent really sucks.
I don't have to admit any such thing. We were a complete NT 4.0 shop. NTFS wasn't as good then as it is now, but it was still a zillion times as robust as ext2. It had to be, because the damn boxes bluescreened all the frickin' time. I don't think I ever lost data from a power failure or system crash. We lost data from developer mistakes, but that's what the backups were for.
:)
EXT2 SUCKED. Pure and simple. Lose power, and there was a high chance you'd lose files. It was eggshell-fragile to hardware failure. Linux could get away with that, because the OS was so incredibly robust at the time that it _never_ crashed. Nowadays, the filesystems are great... the OS itself... not so much.
You realize... you are doing the exact same thing, right? You're blaming the user again. A good operating system survives a power loss. I have taken significant filesystem damage due to OS crashes/power failures from only two sources: one with the game Darklands under DOS and QEMM (wiped my drive), and several times from Linux running ext2. Hint: that means ext2 sucks. Since the advent of journaled filesystems, if I've lost anything, I haven't noticed. Hint: that means ext2 sucks.
And as long as we're genitalia-waving, I was using Linux sometime in late '92 or early '93... my first install was SLS. Mind you, I didn't have a fucking clue what to DO with it, being a DOS guy with no Unix people around to learn from, but under Slackware I managed to figure out modelines and get X going. And I only blew up one monitor doing it.
It was many years before Linux was really business-suitable... mostly because it took businesses awhile to catch up. I had to push really hard to get those DNS boxes in that early. Nowadays, it's an easy decision, but at the time it felt very risky and weird to be running free software in the workplace. It WAS nine years ago... a lot has changed.
This has been true for as long as I can remember. If the software is inadequate or confusing, blame the user. It's happened to me, even here on Slashdot.
Way back in the stone age.... sometime in 1997, maybe? Maybe 1998. Not sure. Anyway... Linux was _just_ starting to get deployed occasionally in business. I had a couple of DNS servers up on an early RedHat box. The box lost power... these were just desktop machines sitting in someone's cubicle. (We hadn't grown to the point of needing 'real' servers quite yet, and an actual server room was a year off.) My primary box took a LOT of filesystem damage, and it took me ages to fix it. So I commented in a slashdot thread that ext2 was very fragile, and that it was one of Linux's real weak points.
You just wouldn't believe the crap I got. Slashdot doesn't seem to archive that far back, so I can't give you links, but _most_ of the replies I got blamed me for being stupid. I "should have used a UPS"... ok, I could grant that, but remember we were a shoestring outfit, and we didn't need those on Windows servers. A couple people went off on me for, get this, not knowing how to use a disk editor to find my secondary superblocks and repair with those. I kid you not. Linux was perfect, and ANYTHING that went wrong was obviously the user's fault... to the point that I should know how to manually repair my filesystem. Instead of admitting that the filesystem should survive a power failure, it was my fault for breaking it.
Several years later, after Reiserfs and ext3 came out, we had a similar conversation, also here on Slashdot. Suddenly everyone is all about how great the journaling filesystems are, and how bad ext2 sucks. It was probably even some of the same people, but the original conversation had already been lost, so I couldn't prove it.
People just will NOT criticize software they're emotionally involved with. It's the most ridiculous thing I've seen... these theoretically intelligent, rational software designers that become absolutely insane when you suggest their software is imperfect. Blame the user! "You're just too stupid to use our software. Go away."
Fortunately, there's enough people in the Linux community now that the lunatic fringe doesn't dominate quite like it did, but these people are still out there.
It was ridiculous then, and it's just as ridiculous now. It doesn't happen as much, but it still sucks.
Hardware virtualization is not a 'bandaid'. In fact, it's the only proper way to do it. Software virtualization is the band-aid.
I suspect you're going the wrong way; with the hardware support in VT and Pacifica, Xen and VMWare are going to get a lot faster. It won't be that hard for them to add in features like what you have (particularly for Xen, which is integrated in the Linux kernel), but it will be very hard for you to do what they're doing.
Specialized hardware support is always faster than general-purpose software. While it may not be all the way there yet in the first generation of hardware, ultimately, solutions that take advantage of hardware to do virtualization will run faster and be much more secure than any software-level solution.
While what you're doing is very useful in the short run, I think you will eventually need to adapt to the hardware model, or die.
Phew, good thing there was room. It would have been awful if all the fish drowned.
You don't get it, do you? They wouldn't object if the documents were fake. They would simply say, "those aren't real documents" and fight them.
The fact that they are claiming that the documents are private means they're correct.
The submitter didn't point out that if AT&T is claiming the documents are trade secrets, that means they're accurate. Made-up documents wouldn't be trade secrets.
In other words, AT&T has just admitted that they are spying on you.
Yeah, I was thinking the same thing. This article would have been interesting, say, 18 months ago... but with VT and Pacifica, things are different now. Without at least mentioning those, it's not very useful.
Anyone have a pointer to a good writeup on the differences between VT, Pacifica, and regular old software virtualization?
*waves hand*. Seven people.
Slashdot links to Groklaw all the time. I count over 120 mentions of Groklaw in Old Stories.... there's probably tons more, but I got bored looking.
Put down the crack pipe, and step away from the keyboard slowly.
A more accurate label would be "Five Day Script Kiddie Class".
See, you're not the problem; you're a computer professional and, at least in theory, you should be highly expert at using a PC. The problem is Tracy in Accounting and Bob the Receptionist, who haven't a clue what's going on with their machines, and who happily install spyware if it promises something slightly better than a sharp stick in the eye.
Think of it as the "OMG Ponies!" crowd, writ large. You just have no idea how freaking stupid these people can be.
Even in the best and brightest companies I've worked in, there have always been a few that got hired that knew a lot less about their PCs than they thought. In particular, they do not appear to hire salespeople for raw brainpower. The clueless users, especially the ones that don't realize (and never will) that they ARE clueless, cause enormous trouble. Unless the network is internally firewalled (which is getting to be a better and better idea, these days), they're often the vectors for network-wide infection.
The draconian policies of some admins may seem stupid, but remember that admins run on fear. They are, by and large, only noticed when things break, and then everyone is mad at them. When a single user can potentially bring a virus into the network that can stop the entire company dead in its tracks, well... it's a heck of a lot safer and easier to just lock EVERYTHING down and then install what people need, as they ask for it.
Think of it as a default-deny firewall.
They do it that way, and then charge weird amounts of points for things in an attempt to either:
A) Have the use of your money for free as long as you have unspent points, or;
B) Sell you something completely useless at an insane profit margin, so that you've spent all your points.
Both benefit Microsoft enormously. I'm sure I will buy very, very few things off XBL.
With the six degrees of separation thing, and careful choices of 'suspects', they can probably get a 90% surveillance rate by declaring only a couple of thousand primary targets.
In other words, as far as I can see, the Australian Parliament has just decreed that the government can read all the email it likes, whenever it likes.
If I were in charge, and unscrupulous, the first person I'd declare a suspect would be the chief of the opposition party.
That's a good observation. It'll be interesting to see if the trend continues.
My apologies, Slashdot editors. I must have missed it. It looks like they ran it even before I got the announcement email from VMWare, so of course they'd have refused my submission. Doh.
:)
Anyway, it IS good software. Wish it were truly free, and I'm looking forward to what happens with Xen, but free-as-in-beer will do for now.
Note that VMWare is also giving away their Server product for free. For some reason, Slashdot hasn't been willing to run this story, even though it's important.
It's a new product, still in beta... about equivalent to the GSX Server. They just released Beta 2 either today or yesterday. It's a _really_ good product. The current keys they're giving away expire, but they say the final version will also be free-as-in-beer.
Basically, it'll do everything Workstation will, plus it allows you to see the consoles of virtual machines that are on another computer. It also gives you a fairly rudimentary web-based control panel, wherein you can start, stop, or restart particular VMs. You can also set up user accounts, and restrict access to particular machines appropriately. It's not ISP-class, but it'd be damn useful for QA teams or suchlike.
Eh, who you going to believe... a publicly traded corporation, or a random Slashdot poster?
You can have my slightly-beat-up 40g iPod, with a tired battery, for TWICE that price.
:)
There.... 'most expensive' didn't last that long.
He should have bought the GPL instead.
Because the data was gathered before the news about the rootkit hit the mainstream media.
It'll never happen, stop asking. CmdrTaco has explicitly said, very recently, that he won't even fix spelling mistakes. He's afraid, get this, that Slashdot will look too professional.
The definition of professional is 'not amateur'... ie, you're getting paid for it. So they want to gete paid for what they're doing, but not do the actual work of doing it right.
I haven't had this place as my homepage in years, and that kind of comment just confirms the decision. I much prefer Ars Technica; they don't have the volume that Slashdot does, but their articles are _really_ good. A mix of digg and Ars does a better job than Slashdot... digg handles the "cursory overview by twits" part, and then there's Ars for in-depth, actual reporting. They even use good grammar, spell things correctly, and get their facts right. It's quite refreshing.
The Slashdot approach worked when they were all 18 and fresh out of school, but I expect a bit more out of adults.
Doesn't Apple use gcc?
I know gcc itself improved a very great deal over the same time period, and I have always assumed that the speed gains were (largely? mostly?) due to that, rather than wondrous new algorithms on Apple's part.
Linux and KDE sped up a lot too, over the same timeframe.