My only worry is that... with 4.4 out, are we going to be subjected to KDE5.0 soon?
Highly unlikely unless Nokia decides to release Qt 5.0, which isn't on any roadmap yet. Everything that's happened with KDE can be traced back to the release of Qt 4.0 in June 2005, which was a major breaking change in the underlying toolkit. The result was that porting something from Qt3 to Qt4 was fairly big rewrite of the UI parts, and of course once you're rewriting everyone wants to add those compatibility-breaking changes they've wanted for so long. No doubt the new way is superior to the old, but the changes were extremely basic. Quoting the WP page:
Trolltech released Qt 4.0 on June 28, 2005 and introduced five new technologies in the framework:
Tulip - A set of template container classes.
Interview - A model/view architecture for item views.
Arthur - A 2D painting framework.
Scribe - A Unicode text renderer with a public API for performing low-level text layout.
MainWindow - A modern action-based main window, toolbar, menu, and docking architecture.
To me those basics look very solid, so I think the 4.x series will be a long lasting one. That's not to say KDE might want to do changes on their own, but there's nothing forcing them from the bottom up to rewrite things that are essentially already working. And particularly not all at once, because clearly 4.0 was too much to swallow all at once.
I think we need to get over this misstep. I totally agree that they played the version number badly, but they also released plenty of warnings about what 4.0 meant and that it was different than a traditional point-oh release. I read these warnings and knew not to take 4.0 seriously. Why didn't other people?
For one, because the distros didn't seem to hear or pass on those warnings. The KDE-centric distros pretty much all went "and now we're upgrading you to 4.0" as if it was the most natural upgrade path in the world. And I dare you to find any place in the release announcement that gave you any hint that's it's not for everyone. You hear "Wait for x.1" about every x.0 release, so you expect the general warnings of "this is a major new release, expect bugs" but still have certain expectations. They would have to come with much, much more explicit warnings that said "This is NOT what you normally expect from a x.0 release, it's much, much more incomplete and buggy than that. Maybe they did but it was a whisper compared to the fanfare it was introduced with.
But if you can't rely on the comments because they're often bogus, and you know that even if you spend time writing good comments the next guy won't believe the comments either and go check the source anyway, then what? One man cant turn the comments from a timesink into an asset, there must be a shared confidence in them. That's where all the bad training comes from.
I've found the following to work well: Place a one-liner comment per method in the header, that's for the public interface giving a clue what it's about, more only if it's really not obvious. This should only explain it as a blackbox function, what it does not how it does it.
In the implementation, add pseudocode comments every 5-15 lines depending on complexity. I usually keep them brief like "// load from database", "// loop to find items", "// display items". The good part is that you can just skim the code this way, and yet the comments are so close to the code it's easy to keep them updated.
Plus, another pattern that I'm very fond of is the exit early pattern, though I know it's deeply hated. To me it keeps functions far more readable with less "active" branches you have to keep in mind, if it failed that test it existed, so assuming this then... I'd much rather see:
Trouble is when you can substitute "IE5 or IE6" with "The legacy codebase that noone ever really wanted to take the cost of cleaning up made by contractors or developers that have quit or retired, ancient technologies and most of it written to pass only the exact use case the business wanted when it was written." Disturbingly enough, I don't think IE would win any real WTF-contests.
Overall I am actually quite impressed (gasp! shoot me now).
I think you're in good company. I just recently saw a poll on a site I visit with about 3600 votes on what OS they were running:
Win 7: 47% XP: 23% Vista: 11% Mac: 10% Linux: 8%
Yeah I know not exactly representative... but at least among Windows users I'd say early adopters, and clearly Windows 7 is a hit. It's completely killed Vista, and even those coming from XP seem happy. I think you can push the "Year of the Linux desktop" back another few years, I'm happy on Linux but any window Vista gave it has closed.
Many objects deal with so small number of records it doesn't matter. I do understand O() notation, but honestly I've never needed to go beyond this page which tells me QList is almost always what I want unless I insert a lot in the middle, then it's QLinkedList. For key lookups QHash is right, if there's an order from a scoring function then QMap or QMultiMap for single/multivalue.
That's served me well, I think the biggest case of n I've used is a file match by hash with n = ~400,000 with completely acceptable performance. Sure it's nothing for cases where you're dealing with millions and billions of records, but if you got 1000 employee records there's no way to sort them "wrong" in my opinion. More often than not the problem is algorithms of O(moron) complexity like fetching all the data and discarding the 99.9% not matching the filter.
Well, probably because computer science is one of the few places where you really go from build to design. Sure it happens that a construction worker becomes a civil engineer or architect, but it's not something that happens by itself. In most lines of work you'll often end up with people doing it some weird way because they've never learned that sort of thing, you can see it in computers too with people that never learned any design patterns and decided to invent their own - mostly poorly. Sure, proven experience beats all but if I was choosing between someone that's learned the theory and has a little experience versus someone that's been busy writing low level procedures all that time it'd be a tough call. If I could have both I'd probably ask the guy with the academic background to draft it and ask the other to sanity check it. Code can be "ugly but works" and it's not really important, people don't touch it much unless they're changing functionality. There's no such as "ugly but works" design, then it IS an ugly design that'll come back to haunt you again and again.
Stardate 5191.5. While googling frantically for pornographic footage of green three breasted alien women
Clearly this is TOS, in TNG you'd just say "Computer, list all pornographic footage of green three breasted alien women." Not surprising that it's Google delivering that amazingly context-sensitive half mind-reading search assistant though, the way they're going it's either that or Skynet.
No, astromatt's understanding of the rules of grammar is correct. Because "White Dwarf" refers to an entity unrelated to a "dwarf", it's pluralised as a proper noun (which generally ignores suffix manipulation). If you knew two people named "Dwarf", you wouldn't say "the two dwarves" (which would only be valid if Dwarf and Dwarf were very short and you were intentionally pointing it out), you'd say "the two Dwarfs".
I'm not a master of the finer nitpicks of the English language, but "white dwarf" is not a name - it's a class like being a star, pulsar, planet, asteroid, meteorite, comet or whatever. Try reading something like a Star Trek science report to yourself "Three pulars, two supernovas, two neutron stars and five white dwarves." No way are those "White Dwarf", "White Dwarf", "White Dwarf", "White Dwarf" and "White Dwarf" as in the "five White Dwarfs".
If the choice (and I know it isn't this simple) was between the military dropping a bomb on a terrorist training camp killing 100 terrorists and those terrorists killing 100 civilians I'd say drop the bomb. There's a difference between you kidnapping me and locking me in the basement and the government, after all due process and all that, putting me in prison. Why and how matters, it is often the difference between good and evil.
I will argue, and I guess you can disagree with me on that, that there are people that should never be allowed to return to society. That it is not worth the risk of more destroyed or lost lives to put murderers and rapists and terrorists - particularly the serial kind, and those we have clear signs have not changed at all in prison - back on the streets again, just to see if they'll do it again. Those people are beyond b) and d), and for a) both are just as effective. I guess it depends if you consider a quick death more of a punishment than living out the rest of your life in a cage. To be both sound about equally grim.
This statement seems silly on the face of it, and would benefit from some, you know, supporting evidence.
My experience every time: I pick an Apple model I could use, and compare to a similar PC - no big difference. I pick the PC I want, try to find something Apple has matching that - Apple loses bigtime.
The only explanation I have for Apple's success is that people don't feel the difference... computers like all have a gigahertz multicore processors, gigs of memory, terabytes of disk hundreds of GPU shaders, all sorts of ports and wireless and whatever. That only form factor and software really matters, the limited and expensive selection of everything else does not. I'm guessing that's true for many people, honestly I've been a power user almost all my life but lately even the most bloated abominations of software malpractice run decently. More speed is starting to be "nice to have", nothing more.
What happens if somebody is released for sexual predation of children? Should their names be stricken from any record of the crime? Does the young mother living next door to this released predator have no right to know of a potential danger?
Welcome to the world that is not just the United States. For example here in Norway, there is in general no public record of any criminal's name. Criminal records are sealed and not even potential employees can in general see them, with some exceptions like working with children, and even then only relevant parts are released so they will see any molestation convictions but not of tax embezzlement. There is no public "sex offender register" or any other form of register to serve as a modern form of the stocks, either you're free to return to society or you aren't. Of course, everyone will notice the 20 years missing in your CV, but they have no right to see the criminal record all the same. Nor is there, as far as I know, any restrictions how far you can live away from anything though you can have restraining orders with regards to your victims.
It's an imperfect solution for an imperfect world, but it's the one we've chosen. I think it's better than the US system, where you release people but give them so many kinds of hell that they have to sleep under bridges and can't lead any form of normal life. If I had to live like that I'd probably kill myself, which I assume is the general idea. It's not going to fix everything, the state isn't going to fix up your relation, friends, family, work relations and work history. But you're not forced to have your crimes branded on your forehead by the government. Somehow the US Supreme Court managed to twist that into not being an additional punishment, which is about the same level as "We'll have this guy with a huge sign following you around, distributing rotten eggs and tomatoes. As long as we're not doing the throwing, we're not punishing you." Nooooooooo, sure you're not. Maybe they earned it but the leaps of logic involved makes it smell dead rotten.
Rude would be if I called you ein verdammter Idiot, it could be seen as either justice or cruelty but never rudeness I think. Personally, I think there are people that are beyond redemption and should be given life without the possibility for release, but I'm against the death penalty on the principle that courts are fallible. We know there has been cases where people convicted to death later have turned out to be innocent, of course nothing can really give them back the 20 years in jail either but then you can at least do something to correct your mistakes.
Instead, this user is interested in finding "that god-damn file" that he saved somewhere yesterday morning and has no idea where it is. He doesn't organize his files, he doesn't care about file hierarchies, he just wants his file. He also wants to easily find that openoffice window that got lost in the 20 windows he opened and never closed in the last hour. Believe it or not, no desktop environment makes it really easy to do such basic stuff.
I just have to ask, who are all these people? That really have no concepts of folders, piles, boxes or whatever metaphor you want to use. I've met people that put everything on the desktop and they're the same people that have all their papers in a big pile, it's not computers they don't understand it's organization. Many, many otherwise quite uneducated computer users get it quite fine. You can't organize unorganized people because you have no idea what belongs together in their mind and trying to divine metadata out of the files themselves is useless except in extremely few limited circumstances where others have filled in the metadata like mp3 id tags. You'll never know that the mp3s, the cover image, the lyric text and their homepage bookmark somehow belongs together unless someone tags it, and it's exactly what those people won't do. They won't even put it into different piles!
If the first one is somewhat common, I've never met anyone with the second problem. Even preschoolers and people well into retirement age don't seem to have a problem with this concept. It's like having a book shelf and a desk, if you keep opening up books and putting them on your desk you'll have a mess, you should close those you are done with by clicking the X in the top right corner. It's a common element you'll find in browsers and tabs, chat messages, email messages in outlook, any of a million normal places in all common applications. If they really are this unfamiliar with computers, they should take some remedial classes at the local community college. This is like trying to design some incredibly complex (and epic fail) system to create a car without a brake pedal on the assumption the driver won't know when to brake.
For example, one source of endless confusion I've seen is launching multiple instances of one application. That one does not translate well from the real world, it's likc taking a book off the shelf multiple times. It makes sense in some contexts, but often what they really want is to bring the current application to the front - or they've forgotten launching it in the first place. Being able to configure that, without any application support, would be solving a real problem. Just configure any application to run in single instance mode and they'd never to get confused again, either click it from taskbar or launch it again and the same instance will be brought to front. Just one example of things I think they could be solving.
If we think of them a little bit as a company and not the anti-RMS, why would they bother now? They're complying, the tool works.... what's the business case? Big companies like to use open source when it's to their advantage. like one vendor I'm familiar with, they are now pushing Linux, Tomcat and many open standards but they're very clearly closed source on top.
More likely you'll get a "Huh? We never paid that much for software, WTF are you talking about? And there's no way we're paying that, we're sticking with what we have. Renewing? We're not renting any software, so why would we need to renew what works?"
As long as the boss doesn't see the difference between doing it legally and doing it illegally, you'll always fail against the old IT person who kept the company running on a shoestring budget. It's the old "I don't care how you do it, bu the last guy did and you either manage with what you got or I'll find someone competent that will." tactic.
The closest thing he could do is probably to get some form of CYA, stating that the IT department has too little information to guarantee that it is license compliant and that proof of existing licenses must be provided otherwise the responsibility must lie with Purchasing/Legal. And keep track of all software he's asked to install, using which keys/licenses.
In all likelihood he can't stop getting fired if shit hits the fan. What he does need is to come out of it with a normal termination because he's got enough proof that the management knew and/or was responsible for the compliance failure, instead of a termination with cause for installing pirated software. From what I gather in the US is pretty common that the only thing companies will confirm is the start date, end date and position of employment so that would be enough.
Is that really interesting outside of the post-apocalyptic scenarios? I'm thinking the point here is to have something you can throw in a vault and actually pick up in a few centuries and use. Unlike pretty much all things magnetic or solid state based, this is more a competitior to digital microfilm or something. For data that's constantly changing this it's easier to just migrate it to new HDDs, but there's a helluva cost to that over a 1000 year perspective. Perhaps the rapid improvements in technology make it cheaper and easier to go with the flow of terabyte hard disks, but it's also worth pursuing what we can find of really long-term storage so we don't get killed by upkeep.
Not to mention a part of the culprit is the fact that Ubuntu keeps shipping completely broken KDE packages. If they didn't ship broken packages, these crashes wouldn't have occured.
So your solution to a file system that goes to hell when machines crash is to make machines that don't crash? Your excuses are terrible.
After a clean reboot pretty much any file written to by any application (during the previous boot) was 0 bytes. For example Plasma and some of the KDE core config files were reset. Also some of my MySQL databases were killed...
-- Bogdan Gribincea
The files that were zeroed when my machine hardlocked I'd imagine were the ones that were in use; my desktop env is Gnome and I was running a game in Wine. Wine's reg files which it would have had open were wiped and also my Gnome terminal settings were wiped.
-- Ben Hodgetts
I'm using 2.6.28-8-generic and a crash just zeroed out a _load_ of important files in my git repository which I'd recently rebased a patch series in.
-- Peter Cliffton
Ack... had a power outage and ran into this one today too. Several configuration files from programs I was running ended up trashed. This also explains the corruption I've seen of my BOINC/SETI files when hard-rebooting in past weeks.
-- 3vi1
I did mix up 30 minutes and 30 seconds. But that's just an example of tons of different applications, databases, source files, Gnome settings and whatever cleaned out by this BUG. Why you keep denying it I don't know, but at least you earned youself a foe rating for it.
Isolated, learning one new way to do something is an acceptable cost. But Linux is to many death by a thousand cuts when it comes to that. I dabbled with Linux for ~7 years before I finally decided to switch and there was many reasons for that. But one of them was the total lack of any familiar application, even though I was fairly convinced Linux probably had some sort of application like that (in retrospect, sometimes a doubtful conclusion) but it was always too much new. I can set off some hours to learn something new, I can't set off a month all at once to learn everything new. It was really only the early experiences with Vista that pushed me over the edge, thinking this wasn't the way forward. Now that I know how to use Linux properly I wouldn't switch back, but that wall is too tall. Not one brick individually but all of the piled on top of each other.
Actually those "Ext4" data corruption issues that set the Internet all ablaze (including Slashdot) were mainly due to KDE 4 not handling metadata correctly. In the end, it wasn't an Ext4 issue. However, feel free to spread FUD.
Bullshit. The error was very generic and not limited to KDE, any system crash could lead to zero byte files in almost any application. It broke an extremely standard way of writing a new version of a file by writing it to a new file then renaming it in place of the old.
You have a file foo.txt You write foo2.txt You rename foo2.txt to foo.txt
Now your machine could crash for up to half an hour (I think, a long time), and you'd have neither the old or the new file. Unless you didn't explicitly fsync() the write, it could have renamed foo2.txt into place without having written the contents of f002.txt, and gone was all your data. It just became painfully obvious in KDE because it broke vital configuration files but it broke thousands of applications. It's a horrible case of the denial you see in Linux land, hiding behind the POSIX specification to defend all the real world damage.
I have no faith in the people developing ext4 after that stupidity, and wouldn't use it on any of my machines. One thing would be if they handled it responsibly, but the childish approach they took was embarrassing. Any person that breaks convention wtih so little regard for the consequences, then tries to defend it so vigorously will not come near my systems. How knows how much else he's willing to break unless there's a standard that says he can't? Because there's sure as hell are many other Linuxisms that aren't POSIX in there as well, fortunately Linus has never been this kind of reckless idiot.
Slashdot Mirror
My only worry is that... with 4.4 out, are we going to be subjected to KDE5.0 soon?
Highly unlikely unless Nokia decides to release Qt 5.0, which isn't on any roadmap yet. Everything that's happened with KDE can be traced back to the release of Qt 4.0 in June 2005, which was a major breaking change in the underlying toolkit. The result was that porting something from Qt3 to Qt4 was fairly big rewrite of the UI parts, and of course once you're rewriting everyone wants to add those compatibility-breaking changes they've wanted for so long. No doubt the new way is superior to the old, but the changes were extremely basic. Quoting the WP page:
Trolltech released Qt 4.0 on June 28, 2005 and introduced five new technologies in the framework:
To me those basics look very solid, so I think the 4.x series will be a long lasting one. That's not to say KDE might want to do changes on their own, but there's nothing forcing them from the bottom up to rewrite things that are essentially already working. And particularly not all at once, because clearly 4.0 was too much to swallow all at once.
I think we need to get over this misstep. I totally agree that they played the version number badly, but they also released plenty of warnings about what 4.0 meant and that it was different than a traditional point-oh release. I read these warnings and knew not to take 4.0 seriously. Why didn't other people?
For one, because the distros didn't seem to hear or pass on those warnings. The KDE-centric distros pretty much all went "and now we're upgrading you to 4.0" as if it was the most natural upgrade path in the world. And I dare you to find any place in the release announcement that gave you any hint that's it's not for everyone. You hear "Wait for x.1" about every x.0 release, so you expect the general warnings of "this is a major new release, expect bugs" but still have certain expectations. They would have to come with much, much more explicit warnings that said "This is NOT what you normally expect from a x.0 release, it's much, much more incomplete and buggy than that. Maybe they did but it was a whisper compared to the fanfare it was introduced with.
But if you can't rely on the comments because they're often bogus, and you know that even if you spend time writing good comments the next guy won't believe the comments either and go check the source anyway, then what? One man cant turn the comments from a timesink into an asset, there must be a shared confidence in them. That's where all the bad training comes from.
I've found the following to work well: Place a one-liner comment per method in the header, that's for the public interface giving a clue what it's about, more only if it's really not obvious. This should only explain it as a blackbox function, what it does not how it does it.
In the implementation, add pseudocode comments every 5-15 lines depending on complexity. I usually keep them brief like "// load from database", "// loop to find items", "// display items". The good part is that you can just skim the code this way, and yet the comments are so close to the code it's easy to keep them updated.
Plus, another pattern that I'm very fond of is the exit early pattern, though I know it's deeply hated. To me it keeps functions far more readable with less "active" branches you have to keep in mind, if it failed that test it existed, so assuming this then... I'd much rather see:
If ( !foo ) {
log( "bad stuff happened" );
return false;
}
than five levels of
if ( foo ) { ...
if ( bar ) {
if ( whatever ) {
Trouble is when you can substitute "IE5 or IE6" with "The legacy codebase that noone ever really wanted to take the cost of cleaning up made by contractors or developers that have quit or retired, ancient technologies and most of it written to pass only the exact use case the business wanted when it was written." Disturbingly enough, I don't think IE would win any real WTF-contests.
Overall I am actually quite impressed (gasp! shoot me now).
I think you're in good company. I just recently saw a poll on a site I visit with about 3600 votes on what OS they were running:
Win 7: 47%
XP: 23%
Vista: 11%
Mac: 10%
Linux: 8%
Yeah I know not exactly representative... but at least among Windows users I'd say early adopters, and clearly Windows 7 is a hit. It's completely killed Vista, and even those coming from XP seem happy. I think you can push the "Year of the Linux desktop" back another few years, I'm happy on Linux but any window Vista gave it has closed.
Many objects deal with so small number of records it doesn't matter. I do understand O() notation, but honestly I've never needed to go beyond this page which tells me QList is almost always what I want unless I insert a lot in the middle, then it's QLinkedList. For key lookups QHash is right, if there's an order from a scoring function then QMap or QMultiMap for single/multivalue.
That's served me well, I think the biggest case of n I've used is a file match by hash with n = ~400,000 with completely acceptable performance. Sure it's nothing for cases where you're dealing with millions and billions of records, but if you got 1000 employee records there's no way to sort them "wrong" in my opinion. More often than not the problem is algorithms of O(moron) complexity like fetching all the data and discarding the 99.9% not matching the filter.
Well, probably because computer science is one of the few places where you really go from build to design. Sure it happens that a construction worker becomes a civil engineer or architect, but it's not something that happens by itself. In most lines of work you'll often end up with people doing it some weird way because they've never learned that sort of thing, you can see it in computers too with people that never learned any design patterns and decided to invent their own - mostly poorly. Sure, proven experience beats all but if I was choosing between someone that's learned the theory and has a little experience versus someone that's been busy writing low level procedures all that time it'd be a tough call. If I could have both I'd probably ask the guy with the academic background to draft it and ask the other to sanity check it. Code can be "ugly but works" and it's not really important, people don't touch it much unless they're changing functionality. There's no such as "ugly but works" design, then it IS an ugly design that'll come back to haunt you again and again.
Sooooooo... they should ignore that about not listening to religious leaders, and follow them blindly? I think they got that covered already.
Stardate 5191.5. While googling frantically for pornographic footage of green three breasted alien women
Clearly this is TOS, in TNG you'd just say "Computer, list all pornographic footage of green three breasted alien women." Not surprising that it's Google delivering that amazingly context-sensitive half mind-reading search assistant though, the way they're going it's either that or Skynet.
No, astromatt's understanding of the rules of grammar is correct. Because "White Dwarf" refers to an entity unrelated to a "dwarf", it's pluralised as a proper noun (which generally ignores suffix manipulation). If you knew two people named "Dwarf", you wouldn't say "the two dwarves" (which would only be valid if Dwarf and Dwarf were very short and you were intentionally pointing it out), you'd say "the two Dwarfs".
I'm not a master of the finer nitpicks of the English language, but "white dwarf" is not a name - it's a class like being a star, pulsar, planet, asteroid, meteorite, comet or whatever. Try reading something like a Star Trek science report to yourself "Three pulars, two supernovas, two neutron stars and five white dwarves." No way are those "White Dwarf", "White Dwarf", "White Dwarf", "White Dwarf" and "White Dwarf" as in the "five White Dwarfs".
If the choice (and I know it isn't this simple) was between the military dropping a bomb on a terrorist training camp killing 100 terrorists and those terrorists killing 100 civilians I'd say drop the bomb. There's a difference between you kidnapping me and locking me in the basement and the government, after all due process and all that, putting me in prison. Why and how matters, it is often the difference between good and evil.
I will argue, and I guess you can disagree with me on that, that there are people that should never be allowed to return to society. That it is not worth the risk of more destroyed or lost lives to put murderers and rapists and terrorists - particularly the serial kind, and those we have clear signs have not changed at all in prison - back on the streets again, just to see if they'll do it again. Those people are beyond b) and d), and for a) both are just as effective. I guess it depends if you consider a quick death more of a punishment than living out the rest of your life in a cage. To be both sound about equally grim.
This statement seems silly on the face of it, and would benefit from some, you know, supporting evidence.
My experience every time:
I pick an Apple model I could use, and compare to a similar PC - no big difference.
I pick the PC I want, try to find something Apple has matching that - Apple loses bigtime.
The only explanation I have for Apple's success is that people don't feel the difference... computers like all have a gigahertz multicore processors, gigs of memory, terabytes of disk hundreds of GPU shaders, all sorts of ports and wireless and whatever. That only form factor and software really matters, the limited and expensive selection of everything else does not. I'm guessing that's true for many people, honestly I've been a power user almost all my life but lately even the most bloated abominations of software malpractice run decently. More speed is starting to be "nice to have", nothing more.
What happens if somebody is released for sexual predation of children? Should their names be stricken from any record of the crime? Does the young mother living next door to this released predator have no right to know of a potential danger?
Welcome to the world that is not just the United States. For example here in Norway, there is in general no public record of any criminal's name. Criminal records are sealed and not even potential employees can in general see them, with some exceptions like working with children, and even then only relevant parts are released so they will see any molestation convictions but not of tax embezzlement. There is no public "sex offender register" or any other form of register to serve as a modern form of the stocks, either you're free to return to society or you aren't. Of course, everyone will notice the 20 years missing in your CV, but they have no right to see the criminal record all the same. Nor is there, as far as I know, any restrictions how far you can live away from anything though you can have restraining orders with regards to your victims.
It's an imperfect solution for an imperfect world, but it's the one we've chosen. I think it's better than the US system, where you release people but give them so many kinds of hell that they have to sleep under bridges and can't lead any form of normal life. If I had to live like that I'd probably kill myself, which I assume is the general idea. It's not going to fix everything, the state isn't going to fix up your relation, friends, family, work relations and work history. But you're not forced to have your crimes branded on your forehead by the government. Somehow the US Supreme Court managed to twist that into not being an additional punishment, which is about the same level as "We'll have this guy with a huge sign following you around, distributing rotten eggs and tomatoes. As long as we're not doing the throwing, we're not punishing you." Nooooooooo, sure you're not. Maybe they earned it but the leaps of logic involved makes it smell dead rotten.
Rude would be if I called you ein verdammter Idiot, it could be seen as either justice or cruelty but never rudeness I think. Personally, I think there are people that are beyond redemption and should be given life without the possibility for release, but I'm against the death penalty on the principle that courts are fallible. We know there has been cases where people convicted to death later have turned out to be innocent, of course nothing can really give them back the 20 years in jail either but then you can at least do something to correct your mistakes.
Instead, this user is interested in finding "that god-damn file" that he saved somewhere yesterday morning and has no idea where it is. He doesn't organize his files, he doesn't care about file hierarchies, he just wants his file. He also wants to easily find that openoffice window that got lost in the 20 windows he opened and never closed in the last hour. Believe it or not, no desktop environment makes it really easy to do such basic stuff.
I just have to ask, who are all these people? That really have no concepts of folders, piles, boxes or whatever metaphor you want to use. I've met people that put everything on the desktop and they're the same people that have all their papers in a big pile, it's not computers they don't understand it's organization. Many, many otherwise quite uneducated computer users get it quite fine. You can't organize unorganized people because you have no idea what belongs together in their mind and trying to divine metadata out of the files themselves is useless except in extremely few limited circumstances where others have filled in the metadata like mp3 id tags. You'll never know that the mp3s, the cover image, the lyric text and their homepage bookmark somehow belongs together unless someone tags it, and it's exactly what those people won't do. They won't even put it into different piles!
If the first one is somewhat common, I've never met anyone with the second problem. Even preschoolers and people well into retirement age don't seem to have a problem with this concept. It's like having a book shelf and a desk, if you keep opening up books and putting them on your desk you'll have a mess, you should close those you are done with by clicking the X in the top right corner. It's a common element you'll find in browsers and tabs, chat messages, email messages in outlook, any of a million normal places in all common applications. If they really are this unfamiliar with computers, they should take some remedial classes at the local community college. This is like trying to design some incredibly complex (and epic fail) system to create a car without a brake pedal on the assumption the driver won't know when to brake.
For example, one source of endless confusion I've seen is launching multiple instances of one application. That one does not translate well from the real world, it's likc taking a book off the shelf multiple times. It makes sense in some contexts, but often what they really want is to bring the current application to the front - or they've forgotten launching it in the first place. Being able to configure that, without any application support, would be solving a real problem. Just configure any application to run in single instance mode and they'd never to get confused again, either click it from taskbar or launch it again and the same instance will be brought to front. Just one example of things I think they could be solving.
If we think of them a little bit as a company and not the anti-RMS, why would they bother now? They're complying, the tool works.... what's the business case? Big companies like to use open source when it's to their advantage. like one vendor I'm familiar with, they are now pushing Linux, Tomcat and many open standards but they're very clearly closed source on top.
More likely you'll get a "Huh? We never paid that much for software, WTF are you talking about? And there's no way we're paying that, we're sticking with what we have. Renewing? We're not renting any software, so why would we need to renew what works?"
As long as the boss doesn't see the difference between doing it legally and doing it illegally, you'll always fail against the old IT person who kept the company running on a shoestring budget. It's the old "I don't care how you do it, bu the last guy did and you either manage with what you got or I'll find someone competent that will." tactic.
The closest thing he could do is probably to get some form of CYA, stating that the IT department has too little information to guarantee that it is license compliant and that proof of existing licenses must be provided otherwise the responsibility must lie with Purchasing/Legal. And keep track of all software he's asked to install, using which keys/licenses.
In all likelihood he can't stop getting fired if shit hits the fan. What he does need is to come out of it with a normal termination because he's got enough proof that the management knew and/or was responsible for the compliance failure, instead of a termination with cause for installing pirated software. From what I gather in the US is pretty common that the only thing companies will confirm is the start date, end date and position of employment so that would be enough.
She also brought the overall issue up with the CEO, suggesting that the company should pay for its payware, or switch to FOSS.
Needless to say, not long afterwards, she was terminated with some lame excuse
That's when you call the BSA and take their reward money with glee. That's at least what I would have done.
Is that really interesting outside of the post-apocalyptic scenarios? I'm thinking the point here is to have something you can throw in a vault and actually pick up in a few centuries and use. Unlike pretty much all things magnetic or solid state based, this is more a competitior to digital microfilm or something. For data that's constantly changing this it's easier to just migrate it to new HDDs, but there's a helluva cost to that over a 1000 year perspective. Perhaps the rapid improvements in technology make it cheaper and easier to go with the flow of terabyte hard disks, but it's also worth pursuing what we can find of really long-term storage so we don't get killed by upkeep.
Not to mention a part of the culprit is the fact that Ubuntu keeps shipping completely broken KDE packages. If they didn't ship broken packages, these crashes wouldn't have occured.
So your solution to a file system that goes to hell when machines crash is to make machines that don't crash? Your excuses are terrible.
KDE 4 wasn't following POSIX standards for writing to a HDD.
You sir, are an idiot.
Read the bug report.
I'll give you some quotes:
After a clean reboot pretty much any file written to by any application (during the previous boot) was 0 bytes.
For example Plasma and some of the KDE core config files were reset. Also some of my MySQL databases were killed...
-- Bogdan Gribincea
The files that were zeroed when my machine hardlocked I'd imagine were the ones that were in use; my desktop env is Gnome and I was running a game in Wine. Wine's reg files which it would have had open were wiped and also my Gnome terminal settings were wiped.
-- Ben Hodgetts
I'm using 2.6.28-8-generic and a crash just zeroed out a _load_ of important files in my git repository which I'd recently rebased a patch series in.
-- Peter Cliffton
Ack... had a power outage and ran into this one today too. Several configuration files from programs I was running ended up trashed. This also explains the corruption I've seen of my BOINC/SETI files when hard-rebooting in past weeks.
-- 3vi1
I did mix up 30 minutes and 30 seconds. But that's just an example of tons of different applications, databases, source files, Gnome settings and whatever cleaned out by this BUG. Why you keep denying it I don't know, but at least you earned youself a foe rating for it.
You are saying as if it is a bad thing.
Isolated, learning one new way to do something is an acceptable cost. But Linux is to many death by a thousand cuts when it comes to that. I dabbled with Linux for ~7 years before I finally decided to switch and there was many reasons for that. But one of them was the total lack of any familiar application, even though I was fairly convinced Linux probably had some sort of application like that (in retrospect, sometimes a doubtful conclusion) but it was always too much new. I can set off some hours to learn something new, I can't set off a month all at once to learn everything new. It was really only the early experiences with Vista that pushed me over the edge, thinking this wasn't the way forward. Now that I know how to use Linux properly I wouldn't switch back, but that wall is too tall. Not one brick individually but all of the piled on top of each other.
Actually those "Ext4" data corruption issues that set the Internet all ablaze (including Slashdot) were mainly due to KDE 4 not handling metadata correctly. In the end, it wasn't an Ext4 issue. However, feel free to spread FUD.
Bullshit. The error was very generic and not limited to KDE, any system crash could lead to zero byte files in almost any application. It broke an extremely standard way of writing a new version of a file by writing it to a new file then renaming it in place of the old.
You have a file foo.txt
You write foo2.txt
You rename foo2.txt to foo.txt
Now your machine could crash for up to half an hour (I think, a long time), and you'd have neither the old or the new file. Unless you didn't explicitly fsync() the write, it could have renamed foo2.txt into place without having written the contents of f002.txt, and gone was all your data. It just became painfully obvious in KDE because it broke vital configuration files but it broke thousands of applications. It's a horrible case of the denial you see in Linux land, hiding behind the POSIX specification to defend all the real world damage.
I have no faith in the people developing ext4 after that stupidity, and wouldn't use it on any of my machines. One thing would be if they handled it responsibly, but the childish approach they took was embarrassing. Any person that breaks convention wtih so little regard for the consequences, then tries to defend it so vigorously will not come near my systems. How knows how much else he's willing to break unless there's a standard that says he can't? Because there's sure as hell are many other Linuxisms that aren't POSIX in there as well, fortunately Linus has never been this kind of reckless idiot.
RTFA. Seriously, you're making a fool of yourself.