A lot of programs can run from USB memory sticks, so you would get the best of all worlds. A clean install to get the machine running, and the users could use Thunderbird etc. without messing up the machine.
I said "everything you're likely to care about". Why does the IT department hate people installing software? Because it messes up the machine. If they're just running Thunderbird from a USB stick, it doesn't cause support burdens for the IT department, so they don't care about it.
The important thing is that users don't go installing spyware crud and other run-at-boot stuff every day.
You forget that the whole thing happened months ago. The exchange would go more like this:
Muckety Muck: Last quarter your unit had profits of $1.5mil. But this quarter you have a loss of $.5mil. Care to explain? Weasely Sony Music Exec: My predecessor was trying to prevent piracy, but chose a flawed product to do it. It could have cost us tens of millions, but I managed to get it down to $2m. Also, if you give us another 10 million dollars, we have this surefire thing that is guaranteed to work! Muckety Muck: You sound full of confidence, so you must be right. Here's another 2 million dollars, and a bonus for saving the company money.
These kind of decisions are always made by people who intend to have been promoted into another division by the time the lawsuits roll around.
Cut the distro down to what will fit on one CD (two max). That will reduce a lot of Debian's headaches. Less for them to maintain and less to test between releases.
That will accomplish nothing. You clearly don't understand how Debian is developed. Each package is maintained by people who care about getting that particular package in the next release. If it's working at that point, it goes in. If it isn't, it gets thrown out, and nobody else wastes any time on it. There is no conceivable reason why throwing out the stuff that already works would make things any easier - and the stuff that doesn't work is already thrown out. Size is not the problem.
Don't be so anal and patch-happy with mainstream packages. Big projects like Gnome and KDE already do extensive testing upstream. Those packages should be able to move more quickly through the unstable-testing-stable cycle without sacrificing stability or extensive patching.
You have clearly never attempted to maintain them downstream. All that 'extensive testing' they do upstream? It's on Redhat, or SuSE, with their own extensively patched versions, tested for the purposes which their paying customers just happened to be using. Fine if you're doing the same thing as them. Useless if you aren't. Does not even attempt to fix the huge numbers of bugs introduced with each new upstream release of these projects.
Redhat and SuSE find loads of bugs with their testing, and send the fixes back upstream.... to be included in the next release. Which has had more new bugs added. Nobody serious ships the unmodified upstream code, it's just used as a common base for patching and propagating patches between distributions.
How much of the debian patching on these type of big projects is *really* functionally necessary versus "I 'm the debian package mantainer and I want to put my mark on it".
Almost every single patch applied to a Debian package is made in response to bug reports filed by Debian users (most of the remaining handful is for policy compliance, portability, or licensing issues). Debian maintainers are far too lazy to go inventing new work to do when there are thousands of outstanding bug reports against these 'extensively tested' packages, listing all the ways in which they suck and need to be fixed.
(I'm an ex-Debian developer who quit for personal reasons)
Nothing catches everythhing. Only clueless CIOs and non-technical middle IT managers think that happens. Security is a state of mind - not a reality. There will always be someone smarter with more time or more resources that can beat your "best practices".
The 'Sisyphus' method catches everything you're likely to care about:
Every desktop should be created via disk image, as a clean install with all the applications that the employees are supposed to be using.
Every night, when everybody's gone home, the image is reinstalled automatically. Wake-on-lan network cards take care of boxes which are switched off. On the rare occasions that a system fails to reinstall, a helpdesk goon is sent over the following morning to swap the box for one of the hot spares (no investigation on site, so the users experience no disruption).
It's that simple. Anything the user does to the desktop is gone the next day. Sure, they could in theory install their pet toys every day... but after they've done it a couple dozen times, they're going to get bored and quit bothering. At the same time, you've eliminated a large range of issues that previously would waste support time (such as bitrot on Windows boxes), and created a system for you to deploy new versions of applications across the entire company easily (just update the image and they'll all get it the next day).
In any case, the quality of Fedora is significant because it determines the first impression of Linux on many people. Even though I have switched distributions, it it possible that I may have stopped using Linux if I had come to the conclusion that Fedora was of too poor quality to use on a daily basis.
If your life is significantly affected by the actions of stupid people, you're screwed already. If it isn't, you aren't going to care about this. While you might possibly be able to justify calling this point 'important', it either doesn't matter or there's nothing you can do about it, so it doesn't merit any attention.
Firefox has not succeeded based on technical merit, it's succeeded based on being the only damn thing out there, unless you use KDE or proprietary software. On Windows, it's been driven by marketing (largely by Google).
If it was a choice based on technical merit, we wouldn't be going with the memory-leaking hog.
What you are proposing is to bury Congress in bureaucracy until they can no longer do anything, on the basis that anything they do is always bad for people. While both the basis and the objective are laudable, this method is needlessly complicated and expensive. I find this to be a simpler and more effective method:
Split the legislature in half. Give one half the ability to grant laws, requiring a 2/3 majority, and the other half the ability to repeal laws, requiring only a 1/3 vote to do so (so in the two-party system, if *anybody* thinks this is a bad law, it's gone). They'll waste most of their time fighting each other, and in the process the number of bad laws on the books will gradually decrease. Any new nonsense that gets passed will be thrown out so many times that it'll take them years to polish it enough, by which point any short-term interests will have expired - effectively, a law can only be written for the long term. Lobbying now costs a fortune and takes forever, so it's not really practical any more (and as soon as you *stop* paying for it, the law gets thrown out again). Note that the repealing house will have to repeal stuff all the time in order to have something to brag about next election.
Yes, this effectively castrates the legislature. It's supposed to. That's got to be an improvement over the current plutocracy.
The trick to dealing with politicians is always to exploit the fact that there's nothing a politician would rather do than fight another politician, given a reason to do so. Most of the time you just need to give them a reason.
PDF is a form of programming system, like postscript. It's possible to write both portable and non-portable PDFs. If you're getting variations across platforms, either your renderer or your generator is broken - and it's usually the generator. Sadly, many versions of Acrobat itself are indeed broken.
It's about three lines of shell script longer to do it that way (basically, you shove an inotifywatch call in front of it - only works on linux, but other platforms have equivalents). I prefer not to do it incrementally like that, because it disrupts normal operation by spending CPU and IO time on running backups - I schedule my backups for when I am sleeping to avoid this - but it's really really easy to do if you want to.
You have failed to read the rsync manpage. If you wanted an atomic copy, evms can give you that on any filesystem you like (there is absolutely no sanity in implementing atomic copy at the filesystem level, it's naturally a device-level concept).
for business I use Word running under Crossover, because exact formating is crucial for me
If exact formatting is crucial, why on earth are you using Word? It's really not very good at precisely reproducing formatting. It only works reliably if both systems have the same *printer drivers* installed (yeah, wtf?) - the rest of the time, it's pot luck whether things go where you want them, or get moved by half a millimetre, knocking all your carefully arranged lines out of position...
If you want exact reproduction of formatting, use PDF. Or latex.
NTFS shares a lot of common structures with IBM's HPFS, from OS/2. This isn't very surprising, because Microsoft have full access to the HPFS source. The exact geneology of this misbegotten filesystem is murky, but it would probably not be too inaccurate to say that it's HPFS, chewed a bit by the DEC crew, and then with ACL support crudely nailed to it.
So... Apple needs a new filesystem to accomplish something that I have been doing for years on ext3?
Here's a one-line implementation of "Time Machine", works on any UNIX filesystem: rsync -a --link-dest=/mnt/backup/yesterday/home/mnt/backup/today
Or there's about a dozen alternative implementations that do basically the same thing, with more or less degrees of sophistication. The one I use is about 20 lines of shell script and shuffles the backups to keep 35 days worth, plus the first day of each month until it runs out of space. You don't need filesystem support for this. It's what hardlinks are for.
Marine-grade engineering is difficult, which means expensive. Very few things will last in a salt water environment without constant maintainance. Furthermore, the structures would have to be extremely large.
Basically, tidal power on a large scale is usually so much more expensive than other methods of power generation that it just isn't worth bothering. Exceptions occur when the tide/coast are abnormally convenient (large, long tidal rivers are good - a huge volume of water passes for a comparatively small power plant), or alternative power sources are not available, for whatever reason.
Second is preventing crime. The theory is if you take a mentally unstable person and bathe him in child porn, virtual or not, he's more likely to actually commit a crime acting out what he's been exposed to. So, by removing the stimulus, you prevent the crime.
I don't know how much data there is on that, but the hypothesis at least has merit for study.
As a subject for research, perhaps. As a legal theory, not at all. Paedophilia is legally considered to be a specific mental disorder. It cannot be caused by looking at pictures - you either are or you aren't. They use this in court all the time (to get higher and more punitive sentences), so they don't get to turn around and introduce a contradictory principle just because it's politically convenient. I don't even care whether this is a medically sound theory - it is inexcusable to set up a contradiction like this. Even a bad, wrongheaded law is better than a law which says different things depending on what the politicians want it to say today.
Bottom line, if you value your data then use backup tapes! That's what that technology was invented for!
Tapes were invented as a short-term mostly-portable bulk storage medium - they were the first such medium invented.
Nowadays, you can actually buy archival backup tapes - but the tapes that you bought for your general purpose backups today are not archival-quality. They will not last 20 years, even on a dark dry shelf. Do not expect them to. If you want to keep them, they need to be regularly rewritten every few years. Even archival tapes need to be rewritten every few decades (all magnetic storage media fades eventually).
If you want your data to survive, don't use magnetic media. The most reliable long-term bulk storage we currently have is actually a chemical/optical technology - microfiche (estimated lifespan ~500 years under the right conditions). That's what people use for things that really, really need to survive (ask your librarian).
I would expect military grade source code to be a very closely guarded, and heavily tested secret !
Why? If we've learned *anything* from free software, it's: secret code is insecure code.
If people having access to the code would be a problem, then you're already screwed - sooner or later, they're going to find out anyway, and once the secret is out then it will never be secret again, and suddenly you've got a billion dollars of useless junk instead of an aircraft.
Trident's a white elephant though - it has never been used and never will be used. It exists for political reasons, not military ones. It's not important in the way that operational aircraft are.
Which wally thought that the primary motivation for programmers was making money?
An editor of a financial magazine. Most people have a bad habit of assuming that all other people in the world are like them. Very few people ever get past it.
They are consuming a luxury (which means "wealth"), while not producing anything or otherwise being of benefit to society (by paying taxes). Yes, a well-run economy would discourage this. If they want to live in a "luxury" house (above the cutoff limit, whatever that is), they'll just have to get enough money to pay the taxes on it - otherwise, they have to go and live in a house which is merely "adequate". Whether this particular proposal for the *degree* of discouragement is sound is a subtle and complicated issue (which houses should be considered "luxury"?), which may make it too difficult to implement in practice - but there's nothing wrong with the *concept* of doing this.
Beware the erroneous implication -- that because wealth is concentrated, the people at the bottom are in worse shape than they were when wealth is not so concentrated.
Wealth has always been so concentrated. It's invalid to track individuals in the population like you did - a person with a job should naturally increase their wealth as they get older, so all you're saying is that older people tend to have more money than young people. If you watch the population as a whole over time, most of the wealth is concentrated in the top few percent. The identities of those people change (all else aside, people die, so it's inevitable), but the distribution doesn't. That's the whole point.
I know people like to rant about the "spam problem" a lot, but for all practical purposes, the problem has been largely solved for several years now.
I wish my purposes were practical. I have somewhere between 99% and 99.9% accuracy in my spam filters (no false positives with the proviso that I'm not counting marketing spiel from companies that I've done business with before - I don't mind losing that), but I still get about 50 slipping through per day. If I want to get rid of the most of those, I'm going to have to sit down and write some new filtering software - nothing that's currently out there will catch it (much of it is megahal-esque gibberish that no naive bayes filter is going to be able to trap - a smarter language model could possibly figure it out, but that's computationally infeasible at present).
Since I don't have 5+ years of experience, excellent "soft skills" (PHB corporate-speak if I've ever heard it), and I don't want to sell anything, I'm apparently unemployable, no matter what school I went to or how well I did.
What you have to realise is that IT-related jobs are often seen as a fast track to middle- or upper-management (common in companies whose main product area is technical). As such, they're more interested in hiring people with management "skills" than people with any level of real technical ability. They will not test your real abilities at either management or technical work before deciding whether to hire you, so those abilities are worthless to you; hiring decisions will be made based on their impressions of whatever they happen to think the relevant attributes are. If you aren't inclined towards sales/management, forget about any large or established company. People will tell you all sorts of things that you "should" be doing, but what it comes down to is that they think you should be doing sales/management instead.
Realistically, you have two options:
- start your own company, or get on board with somebody who is starting one (or just recently started and needs to expand fast)
- forget IT, get a job in another field - your degree is worth pretty much the same regardless of what field you're working in
Slashdot Mirror
I said "everything you're likely to care about". Why does the IT department hate people installing software? Because it messes up the machine. If they're just running Thunderbird from a USB stick, it doesn't cause support burdens for the IT department, so they don't care about it.
The important thing is that users don't go installing spyware crud and other run-at-boot stuff every day.
You forget that the whole thing happened months ago. The exchange would go more like this:
Muckety Muck: Last quarter your unit had profits of $1.5mil. But this quarter you have a loss of $.5mil. Care to explain?
Weasely Sony Music Exec: My predecessor was trying to prevent piracy, but chose a flawed product to do it. It could have cost us tens of millions, but I managed to get it down to $2m. Also, if you give us another 10 million dollars, we have this surefire thing that is guaranteed to work!
Muckety Muck: You sound full of confidence, so you must be right. Here's another 2 million dollars, and a bonus for saving the company money.
These kind of decisions are always made by people who intend to have been promoted into another division by the time the lawsuits roll around.
That will accomplish nothing. You clearly don't understand how Debian is developed. Each package is maintained by people who care about getting that particular package in the next release. If it's working at that point, it goes in. If it isn't, it gets thrown out, and nobody else wastes any time on it. There is no conceivable reason why throwing out the stuff that already works would make things any easier - and the stuff that doesn't work is already thrown out. Size is not the problem.
You have clearly never attempted to maintain them downstream. All that 'extensive testing' they do upstream? It's on Redhat, or SuSE, with their own extensively patched versions, tested for the purposes which their paying customers just happened to be using. Fine if you're doing the same thing as them. Useless if you aren't. Does not even attempt to fix the huge numbers of bugs introduced with each new upstream release of these projects.
Redhat and SuSE find loads of bugs with their testing, and send the fixes back upstream.... to be included in the next release. Which has had more new bugs added. Nobody serious ships the unmodified upstream code, it's just used as a common base for patching and propagating patches between distributions.
Almost every single patch applied to a Debian package is made in response to bug reports filed by Debian users (most of the remaining handful is for policy compliance, portability, or licensing issues). Debian maintainers are far too lazy to go inventing new work to do when there are thousands of outstanding bug reports against these 'extensively tested' packages, listing all the ways in which they suck and need to be fixed.
(I'm an ex-Debian developer who quit for personal reasons)
The 'Sisyphus' method catches everything you're likely to care about:
Every desktop should be created via disk image, as a clean install with all the applications that the employees are supposed to be using.
Every night, when everybody's gone home, the image is reinstalled automatically. Wake-on-lan network cards take care of boxes which are switched off. On the rare occasions that a system fails to reinstall, a helpdesk goon is sent over the following morning to swap the box for one of the hot spares (no investigation on site, so the users experience no disruption).
It's that simple. Anything the user does to the desktop is gone the next day. Sure, they could in theory install their pet toys every day... but after they've done it a couple dozen times, they're going to get bored and quit bothering. At the same time, you've eliminated a large range of issues that previously would waste support time (such as bitrot on Windows boxes), and created a system for you to deploy new versions of applications across the entire company easily (just update the image and they'll all get it the next day).
If your life is significantly affected by the actions of stupid people, you're screwed already. If it isn't, you aren't going to care about this. While you might possibly be able to justify calling this point 'important', it either doesn't matter or there's nothing you can do about it, so it doesn't merit any attention.
Firefox has not succeeded based on technical merit, it's succeeded based on being the only damn thing out there, unless you use KDE or proprietary software. On Windows, it's been driven by marketing (largely by Google).
If it was a choice based on technical merit, we wouldn't be going with the memory-leaking hog.
What you are proposing is to bury Congress in bureaucracy until they can no longer do anything, on the basis that anything they do is always bad for people. While both the basis and the objective are laudable, this method is needlessly complicated and expensive. I find this to be a simpler and more effective method:
Split the legislature in half. Give one half the ability to grant laws, requiring a 2/3 majority, and the other half the ability to repeal laws, requiring only a 1/3 vote to do so (so in the two-party system, if *anybody* thinks this is a bad law, it's gone). They'll waste most of their time fighting each other, and in the process the number of bad laws on the books will gradually decrease. Any new nonsense that gets passed will be thrown out so many times that it'll take them years to polish it enough, by which point any short-term interests will have expired - effectively, a law can only be written for the long term. Lobbying now costs a fortune and takes forever, so it's not really practical any more (and as soon as you *stop* paying for it, the law gets thrown out again). Note that the repealing house will have to repeal stuff all the time in order to have something to brag about next election.
Yes, this effectively castrates the legislature. It's supposed to. That's got to be an improvement over the current plutocracy.
The trick to dealing with politicians is always to exploit the fact that there's nothing a politician would rather do than fight another politician, given a reason to do so. Most of the time you just need to give them a reason.
PDF is a form of programming system, like postscript. It's possible to write both portable and non-portable PDFs. If you're getting variations across platforms, either your renderer or your generator is broken - and it's usually the generator. Sadly, many versions of Acrobat itself are indeed broken.
It's about three lines of shell script longer to do it that way (basically, you shove an inotifywatch call in front of it - only works on linux, but other platforms have equivalents). I prefer not to do it incrementally like that, because it disrupts normal operation by spending CPU and IO time on running backups - I schedule my backups for when I am sleeping to avoid this - but it's really really easy to do if you want to.
You have failed to read the rsync manpage. If you wanted an atomic copy, evms can give you that on any filesystem you like (there is absolutely no sanity in implementing atomic copy at the filesystem level, it's naturally a device-level concept).
If exact formatting is crucial, why on earth are you using Word? It's really not very good at precisely reproducing formatting. It only works reliably if both systems have the same *printer drivers* installed (yeah, wtf?) - the rest of the time, it's pot luck whether things go where you want them, or get moved by half a millimetre, knocking all your carefully arranged lines out of position...
If you want exact reproduction of formatting, use PDF. Or latex.
NTFS shares a lot of common structures with IBM's HPFS, from OS/2. This isn't very surprising, because Microsoft have full access to the HPFS source. The exact geneology of this misbegotten filesystem is murky, but it would probably not be too inaccurate to say that it's HPFS, chewed a bit by the DEC crew, and then with ACL support crudely nailed to it.
So... Apple needs a new filesystem to accomplish something that I have been doing for years on ext3?
/home /mnt/backup/today
Here's a one-line implementation of "Time Machine", works on any UNIX filesystem: rsync -a --link-dest=/mnt/backup/yesterday
Or there's about a dozen alternative implementations that do basically the same thing, with more or less degrees of sophistication. The one I use is about 20 lines of shell script and shuffles the backups to keep 35 days worth, plus the first day of each month until it runs out of space. You don't need filesystem support for this. It's what hardlinks are for.
Marine-grade engineering is difficult, which means expensive. Very few things will last in a salt water environment without constant maintainance. Furthermore, the structures would have to be extremely large.
Basically, tidal power on a large scale is usually so much more expensive than other methods of power generation that it just isn't worth bothering. Exceptions occur when the tide/coast are abnormally convenient (large, long tidal rivers are good - a huge volume of water passes for a comparatively small power plant), or alternative power sources are not available, for whatever reason.
El Reg has already cornered the market on inserting pure fiction into news feeds.
As a subject for research, perhaps. As a legal theory, not at all. Paedophilia is legally considered to be a specific mental disorder. It cannot be caused by looking at pictures - you either are or you aren't. They use this in court all the time (to get higher and more punitive sentences), so they don't get to turn around and introduce a contradictory principle just because it's politically convenient. I don't even care whether this is a medically sound theory - it is inexcusable to set up a contradiction like this. Even a bad, wrongheaded law is better than a law which says different things depending on what the politicians want it to say today.
Why would you go to all that effort just to recreate perl?
Tapes were invented as a short-term mostly-portable bulk storage medium - they were the first such medium invented.
Nowadays, you can actually buy archival backup tapes - but the tapes that you bought for your general purpose backups today are not archival-quality. They will not last 20 years, even on a dark dry shelf. Do not expect them to. If you want to keep them, they need to be regularly rewritten every few years. Even archival tapes need to be rewritten every few decades (all magnetic storage media fades eventually).
If you want your data to survive, don't use magnetic media. The most reliable long-term bulk storage we currently have is actually a chemical/optical technology - microfiche (estimated lifespan ~500 years under the right conditions). That's what people use for things that really, really need to survive (ask your librarian).
Why? If we've learned *anything* from free software, it's: secret code is insecure code.
If people having access to the code would be a problem, then you're already screwed - sooner or later, they're going to find out anyway, and once the secret is out then it will never be secret again, and suddenly you've got a billion dollars of useless junk instead of an aircraft.
Trident's a white elephant though - it has never been used and never will be used. It exists for political reasons, not military ones. It's not important in the way that operational aircraft are.
An editor of a financial magazine. Most people have a bad habit of assuming that all other people in the world are like them. Very few people ever get past it.
They are consuming a luxury (which means "wealth"), while not producing anything or otherwise being of benefit to society (by paying taxes). Yes, a well-run economy would discourage this. If they want to live in a "luxury" house (above the cutoff limit, whatever that is), they'll just have to get enough money to pay the taxes on it - otherwise, they have to go and live in a house which is merely "adequate". Whether this particular proposal for the *degree* of discouragement is sound is a subtle and complicated issue (which houses should be considered "luxury"?), which may make it too difficult to implement in practice - but there's nothing wrong with the *concept* of doing this.
Wealth has always been so concentrated. It's invalid to track individuals in the population like you did - a person with a job should naturally increase their wealth as they get older, so all you're saying is that older people tend to have more money than young people. If you watch the population as a whole over time, most of the wealth is concentrated in the top few percent. The identities of those people change (all else aside, people die, so it's inevitable), but the distribution doesn't. That's the whole point.
I wish my purposes were practical. I have somewhere between 99% and 99.9% accuracy in my spam filters (no false positives with the proviso that I'm not counting marketing spiel from companies that I've done business with before - I don't mind losing that), but I still get about 50 slipping through per day. If I want to get rid of the most of those, I'm going to have to sit down and write some new filtering software - nothing that's currently out there will catch it (much of it is megahal-esque gibberish that no naive bayes filter is going to be able to trap - a smarter language model could possibly figure it out, but that's computationally infeasible at present).
What you have to realise is that IT-related jobs are often seen as a fast track to middle- or upper-management (common in companies whose main product area is technical). As such, they're more interested in hiring people with management "skills" than people with any level of real technical ability. They will not test your real abilities at either management or technical work before deciding whether to hire you, so those abilities are worthless to you; hiring decisions will be made based on their impressions of whatever they happen to think the relevant attributes are. If you aren't inclined towards sales/management, forget about any large or established company. People will tell you all sorts of things that you "should" be doing, but what it comes down to is that they think you should be doing sales/management instead.
Realistically, you have two options:
- start your own company, or get on board with somebody who is starting one (or just recently started and needs to expand fast)
- forget IT, get a job in another field - your degree is worth pretty much the same regardless of what field you're working in