I think you're exceptionally skilled - the type of person to whom generalizations like that do not apply.
Imaging does take a lot of the grind out of refreshing a system. Once you get the habit of it, you can do some cool stuff. It takes a good process though, and a lot of storage to make waypoint snapshots.
You may find it amusing that the post at the top of the thread was moderated 23 times so far - including every moderation key except funny or redundant. A personal best for me, I think.
Using clonezilla, which is free, you can make an image or restore one in 10-20 minutes. It is a good idea to get several snapshots when building the system. And occasionally for backups.
And I didn't call people who don't reimage every year idiots. Those folks are just making a choice that is not what I would recommend. I called people who upgrade their OS in place idiots. That at least we should be able to get consensus on.
As much as I would like to find fault with Microsoft here...
Anybody that "upgrades" a Windows operating system in place from one version to another is an idiot.
People should reinstall their Windows from scratch at least once a year. Any less frequent than that and the successive patches to patches to patches become too much for the system to bear. The successive software installs and uninstalls leave hanging dependencies that slow the system to even worse of a crawl than it was at first install. An "upgraded" system drags with it the legacy rootkits previously installed, and those cause issues even in the best case. In the worst case the malware and crudware bog down the system so much you're lucky to get any work done at all.
A fresh install of XP on modern equipment is almost as snappy as Linux. After a year you're powering up and going for coffee while it "wakes up". After an "OS Upgrade" you don't dare power the thing off unless you're going on vacation for a week. Patch Tuesday has spawned "Team Building Wednesday".
The engineering solution to the psychology problem is to put some mutually agreeable and compatible people in the capsule and launch it. People are mission critical component and a good engineer doesn't fight the properties of his materials.
More people can and should do this. 27C is plenty cool enough for servers. It annoys me to go into a nipple crinkling datacenter knowing they're burning more juice cooling the darned thing than they are crunching the numbers. A simple exhaust fan and some air filters would be fine almost all of the time, and would be less prone to failure.
So just because something has a lot of CPUs and can crunch a lot of numbers, doesn't make it a supercomputer.
Actually yes it does. The type of computer you're talking about is a "general purpose supercomputer". Boinc, Folding@home and Conficker are all "special purpose supercomputers". Any cluster of computers that can act on a single dataset, perform a quadrillion calculations per second, and give coherent results is a "super" computer - at least today. In two decades it might be a cellular phone, but that's a different question.
And because you asked, Mr. AC, the majority (282) of the top500 use gigabit ethernet for interconnect. Given the high computation:data ratio of these special problems, latency is more of an issue than bandwidth. That's great because over the Internet latency has improved quite a bit as more people become aware of what latency and jitter do to their streaming media and gaming experience.
Since in these huge compute:data ratio problems data is passed for a second or two every hour, Internet latency is just not that big of a deal. As latency improves, the compute:data ratio will come down and more and more finely grained problems will be solved.
Sounds to me like a bad idea. Surely there are some tradeoffs between speed and some other also important metric. Ex, faster might mean larger erase block sizes.
Progress is not made by people who concoct imaginary dangers over yonder hill, and then use them as an excuse not to set out at all.
Lot's of corporations still have IE 6 as their "corporate IT approved" browser. I know we do because all our corporate web apps are such shit that they don't work in anything else.
And when the server side software update project gets going, one bullet item on the "requirements" list will be "Cross Platform support".
"Cross platform support: Service must support access from, and usability for, users of AT LEAST TWO of: Internet Explorer 6, Internet Explorer 7, Internet Explorer 8, Firefox 4.2, gopher."
I have nothing to add. Instead, I'll just quote you.
Being an expert in one programming language is fine if you just want to be a programmer. If you want to be a computer scientist, it's neither necessary nor sufficient.
I know more people who "downgraded" from Vista to XP than I ever knew installed ME. I guess that's a major accomplishment for Microsoft's marketing - that they actually got as many people to try this dog as they did.
Strangely enough, I still use ME for some rare needs. There are peculiar things about its software stack that make it useful as a boot OS for odd jobs. But yeah, a terror that one.
On what basis would you expect it to have marketshare even remotely close to XP's ?
The pirates would have moved to Vista right away if it was worth a damn. That's half of the market right there. By Microsoft's own licensing numbers, Vista should have passed XP sometime last year. Apparently a whole lot of people bought Vista who didn't want it. Why is that?
If even the people who steal their software won't use it, that's a damning condemnation right there.
BTW Vista has a larger OS market share than linux + apple combined
And it'll never have more market share than it has today, so let's have a look, shall we? 24.35% according to hitslink for May (The June numbers are "under review" reportedly because they contain unexpected data). That's not a lot to peak at for a Windows OS. It's better than ME, but nowhere near XP.
No, we on the Internets damaged their credibility strongly, and perhaps went a little too far bagging vista.
If we had gone too far we would have damaged our own credibility strongly, not theirs. (checks slashdot karma) Nope. We're good.
It's a difficult position to take that the upswelling rejection of a bunch of blog nerds could overcome the billions of dollars Microsoft drenched their marketing effort with; the brutal armtwisting with OEMs to deliver validated product and offer Vista exclusively. You seem to bear that burden well. Don't you believe in the efficacy of professional marketing to drive a product's image from neutral to positive? Why do you hate America?
The main change is that Microsoft goes back to marketing a product people actually want. From what I can see, pushing Vista damaged their credibility pretty strongly, but with 7 they'll likely regain much of that trust, and in fact already have with the open beta/RC.
Let's fix that - a product people might actually want. It's well established that Vista is a product that people don't want. Whether or not Vista 7 is a product people actually want will depend on what's in the RTM version: whether it's more useful than XP, if it's not more painful to use, if it supports enough hardware and software, if it includes enough new functionality to replace the utility of the inevitable incompatibilities, if it's secure enough to get through the first six months without a major worm.
Since we don't have it yet, we don't know yet how it weighs in the balance.
They are a blessing and a curse. They can be very amusing as they attempt to argue their points from the shallow depth of their experience. I've seen them attribute all manner of things to Microsoft: multitasking, BSD, heirarchical filesystems, multithreading, shared libraries, and on and on. But they can be bullheaded in their attempts to keep their post count up and their scripts are necessarily shallow as there can only be so many response tabs in a talking points folder. Their command of American idiom can be amusing too.
I dread the inevitable "my favorite feature is more important than your favorite feature because..." threads. May FSM protect us from any of that garbage being posted as an actual article. Bangalore must be a dismal place that hundreds of people will line up for a job blogging in a cube 3' square for 12 hours a day for a couple bucks a day.
They're not the doom of Web 2.0, but they can be a nuisance. We're at the same point we were at before the launch of Window Vista - attempting to build momentum and energy for a huge launch day where people stand in line to get the product at midnight. I don't see it happening yet, but they could get that going.
I propose that we call them MATT: Microsoft AstroTurf Team.
What does Windows do wrong, that your "proper" OS does right?
A proper OS adheres to decades of best practice on security without regard to whether discarding the basic principles of security enables "popular" features.
Interoperability with past and future versions of the same operating environment and other operating environments - both with applications and data - is a core property of the thing called an "operating system". Without this property it's not an operating system, it's an operating environment. The difference is in whether it's suitable a suitable tool for solving a transient problem, or a platform upon which considerable business intelligence can be invested, and whether or not it's a foundation for progress. An operating system doesn't require an army of lawyers and deliberate engineering to prevent compatibility to defend its market share because it's not a tool for the vendor, it's a tool for the user.
A proper operating system comes with a toolchain to migrate it to another hardware platform, because hardware changes and a core principle of operating systems is not to trap its user into using an ephemeral hardware system.
There are more core principles of operating systems, but these three should be enough to illustrate that Windows never has been and never will be a "proper" operating system.
When people create innovation and make it common and then other people build upon that platform to drive ever higher advancements by also making them commonly available, that's called progress. It's the advantage that free software has over the commercial variety.
If after a while you haven't discovered that all languages are essentially the same you should find another career path. But don't program in COBOL if you can avoid it.
If you're using hard drives for tape-like tasks, for example backup, you get the same advantages of bandwidth multiplication and storage capacity as tape. Also, the random-access nature of hard drives offers the facility to do differential backups that can function as live snapshots. Current virtual tape libraries emulate all of the advantages of tape except one: archive longevity. Tape is still the standout winner there. Hardware compression is overrated in the modern era, as the task is just not complex and modern processors already have hardware assist for typical compression and encryption operations at the machine language level, so these tasks are no sweat for your ipod, let alone server class equipment. This is even before we consider data deduplication.
A single 2TB SATA drive with typical 4x compression can store 8TB of data, writing at 320MB/s and reading at 640MB/s. If someone would bother to make a 10-drive cartridge for backup we could bump that figure to 64TB capacity with raid-6 dual redundancy and triple the bandwidth. I don't know, but there may be NAS devices that offer this with a 10Gbps Ethernet interface in the market now. To fully exploit the bandwidth of such a device you would need external PCIe, hot-pluggable, which is indeed part of the PCIe 2.0 standard but not yet implemented that I know of. For your $/GB metric you're considering only the cost of media, not the horrendous cost of proprietary equipment and software to implement a tape solution. SATA drives are bog-standard equipment and no special software is needed, and they don't tie you to a particular vendor.
I would say that the days of tape are done for everything except the legal requirement for long-duration archival storage or to satisfy written policies that take many years to turn. In technology though with time all things change.
The current problem with enterprise backup is that the communications monopolies have bound up all that dark fiber to drive up the costs. Otherwise we could all just use continuous data replication to multiple geographically diverse sites for primary storage, with each site having its own offline storage for truly redundant unkillable data storage. It doesn't even have to cost a lot. Those 2TB drives are $250, and servers start at around $700 now.
Slashdot Mirror
I think you're exceptionally skilled - the type of person to whom generalizations like that do not apply.
Imaging does take a lot of the grind out of refreshing a system. Once you get the habit of it, you can do some cool stuff. It takes a good process though, and a lot of storage to make waypoint snapshots.
You may find it amusing that the post at the top of the thread was moderated 23 times so far - including every moderation key except funny or redundant. A personal best for me, I think.
Using clonezilla, which is free, you can make an image or restore one in 10-20 minutes. It is a good idea to get several snapshots when building the system. And occasionally for backups.
And I didn't call people who don't reimage every year idiots. Those folks are just making a choice that is not what I would recommend. I called people who upgrade their OS in place idiots. That at least we should be able to get consensus on.
As much as I would like to find fault with Microsoft here...
Anybody that "upgrades" a Windows operating system in place from one version to another is an idiot.
People should reinstall their Windows from scratch at least once a year. Any less frequent than that and the successive patches to patches to patches become too much for the system to bear. The successive software installs and uninstalls leave hanging dependencies that slow the system to even worse of a crawl than it was at first install. An "upgraded" system drags with it the legacy rootkits previously installed, and those cause issues even in the best case. In the worst case the malware and crudware bog down the system so much you're lucky to get any work done at all.
A fresh install of XP on modern equipment is almost as snappy as Linux. After a year you're powering up and going for coffee while it "wakes up". After an "OS Upgrade" you don't dare power the thing off unless you're going on vacation for a week. Patch Tuesday has spawned "Team Building Wednesday".
You're kidding, right?
The engineering solution to the psychology problem is to put some mutually agreeable and compatible people in the capsule and launch it. People are mission critical component and a good engineer doesn't fight the properties of his materials.
If you're Google? Apparently the answer is "yes."
More people can and should do this. 27C is plenty cool enough for servers. It annoys me to go into a nipple crinkling datacenter knowing they're burning more juice cooling the darned thing than they are crunching the numbers. A simple exhaust fan and some air filters would be fine almost all of the time, and would be less prone to failure.
How did you get that ghostly omage of Steve Ballmer in there?
I'm sure the process creates works of intrinsic art of themselves, new works.
So just because something has a lot of CPUs and can crunch a lot of numbers, doesn't make it a supercomputer.
Actually yes it does. The type of computer you're talking about is a "general purpose supercomputer". Boinc, Folding@home and Conficker are all "special purpose supercomputers". Any cluster of computers that can act on a single dataset, perform a quadrillion calculations per second, and give coherent results is a "super" computer - at least today. In two decades it might be a cellular phone, but that's a different question.
And because you asked, Mr. AC, the majority (282) of the top500 use gigabit ethernet for interconnect. Given the high computation:data ratio of these special problems, latency is more of an issue than bandwidth. That's great because over the Internet latency has improved quite a bit as more people become aware of what latency and jitter do to their streaming media and gaming experience.
Since in these huge compute:data ratio problems data is passed for a second or two every hour, Internet latency is just not that big of a deal. As latency improves, the compute:data ratio will come down and more and more finely grained problems will be solved.
Sounds to me like a bad idea. Surely there are some tradeoffs between speed and some other also important metric. Ex, faster might mean larger erase block sizes.
Progress is not made by people who concoct imaginary dangers over yonder hill, and then use them as an excuse not to set out at all.
Lot's of corporations still have IE 6 as their "corporate IT approved" browser. I know we do because all our corporate web apps are such shit that they don't work in anything else.
And when the server side software update project gets going, one bullet item on the "requirements" list will be "Cross Platform support".
"Cross platform support: Service must support access from, and usability for, users of AT LEAST TWO of: Internet Explorer 6, Internet Explorer 7, Internet Explorer 8, Firefox 4.2, gopher."
Strangely enough 410 of the supercomputers listed in the top500, or 82% are of architecture type "cluster".
Look back at the OLPC problems once Microsoft got to their customers.
Where do you want to go today?
I have nothing to add. Instead, I'll just quote you.
Being an expert in one programming language is fine if you just want to be a programmer. If you want to be a computer scientist, it's neither necessary nor sufficient.
That was beautiful.
I know more people who "downgraded" from Vista to XP than I ever knew installed ME. I guess that's a major accomplishment for Microsoft's marketing - that they actually got as many people to try this dog as they did.
Strangely enough, I still use ME for some rare needs. There are peculiar things about its software stack that make it useful as a boot OS for odd jobs. But yeah, a terror that one.
It's not learning about SQL I have an issue with, I just think CS grads should have more exposure and practice working with it.
Programming 252: Understanding SQL. ....
Lab objective 3: Port PostgreSQL source code to one of the following languages and pass regression testing.
On what basis would you expect it to have marketshare even remotely close to XP's ?
The pirates would have moved to Vista right away if it was worth a damn. That's half of the market right there. By Microsoft's own licensing numbers, Vista should have passed XP sometime last year. Apparently a whole lot of people bought Vista who didn't want it. Why is that?
If even the people who steal their software won't use it, that's a damning condemnation right there.
BTW Vista has a larger OS market share than linux + apple combined
And it'll never have more market share than it has today, so let's have a look, shall we? 24.35% according to hitslink for May (The June numbers are "under review" reportedly because they contain unexpected data). That's not a lot to peak at for a Windows OS. It's better than ME, but nowhere near XP.
No, we on the Internets damaged their credibility strongly, and perhaps went a little too far bagging vista.
If we had gone too far we would have damaged our own credibility strongly, not theirs. (checks slashdot karma) Nope. We're good.
It's a difficult position to take that the upswelling rejection of a bunch of blog nerds could overcome the billions of dollars Microsoft drenched their marketing effort with; the brutal armtwisting with OEMs to deliver validated product and offer Vista exclusively. You seem to bear that burden well. Don't you believe in the efficacy of professional marketing to drive a product's image from neutral to positive? Why do you hate America?
The main change is that Microsoft goes back to marketing a product people actually want. From what I can see, pushing Vista damaged their credibility pretty strongly, but with 7 they'll likely regain much of that trust, and in fact already have with the open beta/RC.
Let's fix that - a product people might actually want. It's well established that Vista is a product that people don't want. Whether or not Vista 7 is a product people actually want will depend on what's in the RTM version: whether it's more useful than XP, if it's not more painful to use, if it supports enough hardware and software, if it includes enough new functionality to replace the utility of the inevitable incompatibilities, if it's secure enough to get through the first six months without a major worm.
Since we don't have it yet, we don't know yet how it weighs in the balance.
They are a blessing and a curse. They can be very amusing as they attempt to argue their points from the shallow depth of their experience. I've seen them attribute all manner of things to Microsoft: multitasking, BSD, heirarchical filesystems, multithreading, shared libraries, and on and on. But they can be bullheaded in their attempts to keep their post count up and their scripts are necessarily shallow as there can only be so many response tabs in a talking points folder. Their command of American idiom can be amusing too.
I dread the inevitable "my favorite feature is more important than your favorite feature because..." threads. May FSM protect us from any of that garbage being posted as an actual article. Bangalore must be a dismal place that hundreds of people will line up for a job blogging in a cube 3' square for 12 hours a day for a couple bucks a day.
They're not the doom of Web 2.0, but they can be a nuisance. We're at the same point we were at before the launch of Window Vista - attempting to build momentum and energy for a huge launch day where people stand in line to get the product at midnight. I don't see it happening yet, but they could get that going.
I propose that we call them MATT: Microsoft AstroTurf Team.
And those Seinfeld ads were brilliant!
What does Windows do wrong, that your "proper" OS does right?
A proper OS adheres to decades of best practice on security without regard to whether discarding the basic principles of security enables "popular" features.
Interoperability with past and future versions of the same operating environment and other operating environments - both with applications and data - is a core property of the thing called an "operating system". Without this property it's not an operating system, it's an operating environment. The difference is in whether it's suitable a suitable tool for solving a transient problem, or a platform upon which considerable business intelligence can be invested, and whether or not it's a foundation for progress. An operating system doesn't require an army of lawyers and deliberate engineering to prevent compatibility to defend its market share because it's not a tool for the vendor, it's a tool for the user.
A proper operating system comes with a toolchain to migrate it to another hardware platform, because hardware changes and a core principle of operating systems is not to trap its user into using an ephemeral hardware system.
There are more core principles of operating systems, but these three should be enough to illustrate that Windows never has been and never will be a "proper" operating system.
When people create innovation and make it common and then other people build upon that platform to drive ever higher advancements by also making them commonly available, that's called progress. It's the advantage that free software has over the commercial variety.
If after a while you haven't discovered that all languages are essentially the same you should find another career path. But don't program in COBOL if you can avoid it.
If you're using hard drives for tape-like tasks, for example backup, you get the same advantages of bandwidth multiplication and storage capacity as tape. Also, the random-access nature of hard drives offers the facility to do differential backups that can function as live snapshots. Current virtual tape libraries emulate all of the advantages of tape except one: archive longevity. Tape is still the standout winner there. Hardware compression is overrated in the modern era, as the task is just not complex and modern processors already have hardware assist for typical compression and encryption operations at the machine language level, so these tasks are no sweat for your ipod, let alone server class equipment. This is even before we consider data deduplication.
A single 2TB SATA drive with typical 4x compression can store 8TB of data, writing at 320MB/s and reading at 640MB/s. If someone would bother to make a 10-drive cartridge for backup we could bump that figure to 64TB capacity with raid-6 dual redundancy and triple the bandwidth. I don't know, but there may be NAS devices that offer this with a 10Gbps Ethernet interface in the market now. To fully exploit the bandwidth of such a device you would need external PCIe, hot-pluggable, which is indeed part of the PCIe 2.0 standard but not yet implemented that I know of. For your $/GB metric you're considering only the cost of media, not the horrendous cost of proprietary equipment and software to implement a tape solution. SATA drives are bog-standard equipment and no special software is needed, and they don't tie you to a particular vendor.
I would say that the days of tape are done for everything except the legal requirement for long-duration archival storage or to satisfy written policies that take many years to turn. In technology though with time all things change.
The current problem with enterprise backup is that the communications monopolies have bound up all that dark fiber to drive up the costs. Otherwise we could all just use continuous data replication to multiple geographically diverse sites for primary storage, with each site having its own offline storage for truly redundant unkillable data storage. It doesn't even have to cost a lot. Those 2TB drives are $250, and servers start at around $700 now.