Honestly, Windows just doesn't feel that heavy to me - read my other post about my experiences with XP and KDE on a 256MB laptop. I forgot to mention some other things in that post, so here they are now:
Linux takes aeons to boot to a usable desktop, whereas I find Windows much faster.
Certain apps (notably Firefox and Thunderbird, which both use XUL) visibly redraw when you switch to them under Linux, but I see nothing like this under Windows. You can actually see them filling themselves in!
Most apps start up slower under Linux, although this is mainly due to the need to dynamically link apps as they are started. Nevertheless, even with preloading on, launching Explorer in Windows is faster than launching Konqueror.
Dragging a gaim Window over Firefox leaves half-drawn copies of gaim all across my screen. Bigger windows are even worse - I can fill the entire screen with half-drawn copies of windows just by waggling the mouse!.
I hope you're not trying to paint me as "an MS shill"; check through my posting history and you'll see that this is anything but the case.
Anyway, point by point:
Well, unfortunately for them, 2000/XP, though better than Win 9x, are still far from reality. XP still does not run and run, 2000 does not either. They slowly come to a halt, XP faster than 2000. These OS can run and run, if you put one service only on them, and tweak them for two plain days until they ressemble nothing you could work with, which is a process I call tedious, not a "miracle".
I never saw one of these OS run and run, but I suppose it is acheivable. Compared to a Linux that run and run without any tweak, there is still a chasm between the two OS.
On this, we'll have to agree to disagree; I've seen XP crash a handful of times: a few times when it turned out my graphics card was faulty (this crashed Linux, too) and once when I was copying a file from my Linux comp to a friend of mine's XP comp via samba. The latter is inexcusable (but funny side-note: my friend blamed "Linux's crappy Samba implementation", even though it was his computer that crashed, not mine!). The rest of the time, my XP computer at work truly does run and run, requiring a reboot only when a critical Windows Update is required. Your experiences are apparently different.
When KDE, despite being stuffed with features at every new version, becomes more and more fast with each release.
This is true - KDE has been getting faster and more memory-friendly - check out the "top" output for KWrite (or is it Kate) under the KDE4 prototype code. I'm very pleased with the progress of KDE, and a recent talk by Robert Love (on optmising GNOME) shows that the Desktop Linux developers are very committed to reducing bloat, which I couldn't be happier about. However, the fact that KDE is getting faster either tells us that something that's always been good is getting better, or something that was slow and memory hungry before is getting better - much like Mac OS X, which started out dog-slow but which has been improving in speed with each successive release. Respectfully, I'd have to say it was the latter: on my 256MB laptop, KDE starts to swap more and much sooner than XP does (i.e. with fewer apps open). Firefox consumes far more resources than IE (although admittedly it also accomplishes far more). Having said that, the focus on the Linux side is on getting faster (I'm drooling with anticipation at XGL), so on the speed/ memory consumption side I see Linux ultimately winning out.
Now, one thing I wonder : how MSH will come superior to bash, when bash is cross-platform, and msh is not ?
It destroys one of the most important features of shells...
Cross-platform-ness is admirable, but a shell where objects are first-class citizens sounds pretty good to me. I determine the power of a shell by how much easier it will make my life.
It's pretty weird, yeah, and very obscure - I found it by sheer fluke. Of course, it's only fair to point out that CTRL+C is also expected to shut down the current foreground process in a CLI environment, so I can see why CTRL+C does not act as "Copy".
Because there are a surprisingly large contingent of people who define themselves by the operating system they use, and whose self-esteem is directly linked to the perceived superiority of this operating system over Microsoft Windows. During the late nineties, when Windows was truly a buggy, crashy, piece of shit, these people positively basked in the glee that came from the vindication of their chosen OS - back then, Linux truly was light-years ahead of Windows in terms of speed and stability, and geeks rejoiced in the streets.
Flash-forward to Windows 2000/ XP, and Microsoft apparently accomplished a miracle, producing a version of Windows that would literally run and run, and was still fairly nippy. Meanwhile, the writers of Linux Desktop Environments were discovering that it's very easy to be fast and light when you don't do much, or aren't particularly user-friendly, and that increased functionality almost always comes at the price of bloat.
So these people saw two pillars of the superiority of Linux (speed and stability) snatched away from them. The truly curious thing is what happened next: instead of being spurred into action by this new competition and addressing these concerns on the Linux side, these people instead simply went into a state of denial, and refused to let go of these cherished (and rapidly shrinking) areas where Linux once scored over Windows. Read through any anti-MS slashdot article on any given day and count the number of horribly outdated criticisms of Microsoft you see (BSOD's; bloat; Clippy(!)) - as a passionate believer in F/OSS, it really grieves me to see people behaving like this, rather than aiming to improve Linux to the state where it once again has many advantages over Windows.
Flash-forward to now, as one of the other areas in which Linux scores over Windows (a UNIX command-line is an awesome and enjoyable tool to use; the Windows command line, by contrast, is a rubber hammer with nails in the handle:)) may well be snatched away, and we see the same thing: people are hoping against hope that Microsoft foul it up, because if they don't another area of Linux superiority disappears, along with another shred of their self-esteem. This, I think, is why people care, and why they do not wish Microsoft well in this project, however helpful it may be to the common good.
You can already copy from/ paste to Window's default CLI (highlight text and press ENTER to copy; right-click to paste - same with cygwin). I don't know about cutting, though.
Ah, fair enough. I think this is almost certainly covered by the inotify mechanism and, if not, could be added in relatively short order. Linux is usually pretty agile that way:)
I've not used it myself, but look into Beagle. With Linux, the area corrsponding to "user space programs that are outside of the kernel developer's control" is the province of the distros maintainers. I have absolutely no idea how this ties in with cp, though - why does cp need to change? The Linux inotify Kernel patches (which Beagle uses to monitor for changes) work at a deeper layer than cp et al.
Thankfully, both KDE and GNOME are currently shifting their efforts to optimisation. Software development proceeds sporadically, and spans many phases (often several at the same time!). The three major phases are probably a) Massive Feature Addition, b) Bug-Fixing and c) Optimisation, corrsponding roughly to Making it Work, Making it Work Well, and Making it Work Fast. As I stated in a post in another thread, I don't really consider Desktop Linux to have started sincere development until a year and a half - two years ago, at which point Desktop Linux was horribly behind the times.
Since then, though, there has been a *massive* amount of work done, mostly in the areas of adding new features or increasing the power of the libraries. This has left Desktop Linux in a rather buggy, bloated state, but I think it is beginning to settle down slightly and tend towards b) and c) (the kernel itself seems to have a recent flare-up of new stuff being added, to the detriment of its speed and stability; doubtless, the latter two qualities will reassert their dominance once the frenzy wears off). Robert Love gave a very interesting talk on optimising GNOME (most of the ideas apply to KDE, also), and it seems that a lot of developers have a keen interest in getting Linux to run on old hardware - particularly the Ubuntu team, who have the resources and know-how to get the job done.
OO.o is a lost cause, alas, although I have heard good things about KOffice. Some have stated that it is approaching the functionality of OO.o, but is still fairly lightweight - can anyone who has used the two for "proper" work possibly comment on this?
Oh, and while I love Firefox, I find it odd that you were hail this as a prime example of "smaller and faster, not bigger and slower.":)
This causes two problems: 1) Distros take longer to release (uhh, sarge?) as they're maintaning not just the OS but all the applications that go with it too
Yes, as demonstrated by the glacial release cycle of Ubuntu, versus the incredibly rapid turnover of Windows and OS X. Or rather, not.
you install something that isn't packaged with your distros packaging system you break the package management completely
I've never actually seen this happen myself, but I'll take your word for it:). Anyone have any horror stories to share? Does it happen with actual packages, or only if you install from source?
I personally don't even notice it any more (probably because the only GTK app I use is gaim, and very occasioanlly, GIMP:)), but I guess it's confusing for most new users. Has anyone tried out gtk-qt? Does it work at all well? Obviously it can't handle entirely different UI conventions, but I guess it could possibly make things at least look the same.
why haven't any distros latched onto it? Yes, it means that the OS must promise a base set of shared libraries
You just answered your own question:) Seriously, the need to guarantee the presence of a basic toolkit is the only real obstacle here - the rest is pretty much cake - just add a GUI wrapper around dpkg/ whatever such that the act of dragging an app somewhere installs it at the given place; create some kind of meta-deb/ RPM that contains its dependencies as sub-debs that are installed as a shared library if they are not already present (no real need for big, statically linked binaries, I don't think...?), etc. Personally, I haven't had any trouble with file associations, so I'm not sure if this is a hurdle - I think the.desktop spec solved this a while ago. Most of the basic tools are already in place, except for this "meta-debs", but these would be easy to implement, if the community showed interest, and given automated tools would spread across pretty much every active project in the OSS world.
So the major problems here would be settling on some kind of Linux Standard Desktop Base of libraries, and persuading the community that "meta-debs" are a good idea.
My apologies for continuing the off-topicness, but I'd like to address this:
While I agree that these problems need to be tackled, you are flat-out wrong if you think that people aren't already working very hard on the problem. The fact is that (in my humble opinion), in spite of the fact that GNOME and KDE started up way back in '98 and '99, Desktop Linux has been almost entirely ignored prior to, say, a year and a half ago. Back then, if you asked on a forum about how to get your USB pen to automount when you plug it in, you would be told to write your own script for monitoring the tail of dmesg and respond to USB events my parsing the output, as if this were a completely acceptable solution.
Nowadays, though, with iniatives such as Ubuntu, this kind of attitude doesn't fly: sure, if the problem you have can't currently be solved in a neat, elegant way then they'll give you a Grandma-unfriendly piece of advice, but you can bet that someone will begin looking at a way to fix this specific problem elegantly for the next release. In short, there has been either a shift in attitude (or perhaps just an infusion of new blood) whose mantra is "If it doesn't work out of the box, it is a bug; report it to us, and we will fix it if we can" rather than the "If it doesn't work out of the box, fix it yourself" of yesteryear.
This development is a recent phenomenon, however, and Desktop Linux has a *huge* amount of catching up to do (both MS and Apple have a very big headstart in terms of time and resources that they have already thrown into the problem). Adding the polish and design cohesiveness required for "usability" is a very, very hard problem that takes a big investment of time and resources to address, but these resources are being levelled at the problem right now, and I tire of the constant barrage of posts who seem to think that the Linux community consists entirely of developers who know and care nothing for end-user usability and who seem to think that sitting back and saying "why don't you just make it usable?" is some kind of fantastic idea that had simply never occurred to the myopic developers, rather like someone watching a medical programme and shouting out "Hey, why don't you Doctors and Scientists try and cure cancer!"
They understand perfectly. They also understand that, thanks to bullshit laws like the DMCA, they can threaten to sue the (usually penniless) authors of the "cracks" and they will instantly cave, as has happened here. If, for whatever bizarre reason, the "offender" did not cave, they understand that they could financially drag the guy over the coals and ruin his life, making an effective example for anyone else who has any bright ideas about breaking their DRM schemes.
They understand perfectly that technical solutions on their own aren't always tenable; they also understand that technical solution + threat of lawsuit == "teh win".
It's nuts, isn't it? Especially with something as utterly generic (and probably free of "valuable" IP) as a sodding wireless card where I can pick up one just as good for $20 on eBay. I mean, I can understand high-end graphics cards manufacturers not giving out their specs as there probably is some very interesting stuff in there that has taken them a long while to develop, but a wireless card or a WinModem? Stupid, self-defeating and petty. Thank God for enlightened manufacturers like ralink - as a Linux user, I'm practically falling over myself to buy one of their nifty USB wireless-g pens:)
It most definitely affects Firefox on Linux, too. It depends a lot on your browser habits; the leak
seems to get worse every time you open a tab i.e. closing a tab does not appear to free all of the memory that was allocated when it was opened, and this makes using Firefox with my personal usage pattern a nightmare:) I used Firefox fairly intensely under Linux last weekend, and eventually my 1/2GB of RAM and my 1/2GB of swap were all used up. Firefox had just ten tabs open at the time. The time taken to close Firefox, from clicking OK on the "Close all tabs?" prompt to the ceasation of the flurry of disk activity was at least one and a half-minutes (7200RPM SATA, 8MB cache). At this point, the memory and swap consumption reported by top both instantly halved:)
Memory fragmentation should, but there is always the risk of an errant part of your code holding references to an object which is no longer needed, possibly until the program is closed. The situation is much-improved with languages that have garbage collection as it is more automated and remove some of the burden from the programmer, but the problem of determining when an object will definitely never be used again by a program and thus making it eligible for garbage collection is in general only solvable by strong AI (having no references to an object reachable by a chain of references stemming from the program root is a good heuristic for deciding that an object can be disposed of, but there are always cases it will miss).
There were many, many leaks in Firefox, and many have been fixed for 1.1 (do a search on their Bugzilla for "memory leak"). Hopefully, the situation is now much-improved, but I suspect it will be the case that long periods of heavy-browsing will require you to to restart Firefox for quite a while yet. For this reason, I always recommend the Session Saver extension - makes closing and restarting Firefox less painful.
Memory fragmentation is a big issue for modern desktop systems as the heap used by programs written in C/C++ can't be compacted, and most memory allocation systems weren't necessarily designed to support programs that would be continually allocating and deallocating memory for days on end. Robert Love gave a (fairly detailed and technical) talk on it at while back, with some suggestions for combating it on the Linux desktop, which I recommend to anyone who is interested. It's about 126MB, Ogg format.
They have acknowledged the other sides point of view, and even used some of it. Slashdot often cries that Microsoft doesn't look at the open source crowd's point of view, but they can be the same exact way when the positions are reversed. (And for you nitpickers out there, the keyword is 'can'.) Microsoft is a business, a relatively new one at that. If a new business has a way to keep the money flowing, then they will do it!
The thing of it is that Microsoft has deemed Linux internally to be Public Enemy Number 1 (see: The Halloween Documents) and have a track record of a) not suffering competition to exist, unless it benefits them some way; and b) fighting dirty. In brief, they have lied so many times and built up such a great deal of bad blood with the F/OSS community that, understandably, no one believes a single word they say. As for becoming more "open" - witness the current EU situation, where the EU states that their use of proprietary protocols (e.g. Samba) is anti-competitive, and Microsoft's resulting proposal - effectively, following the letter of the law but willfully acting in direct opposition to the spirit of the law (which Microsoft understood full well; Microsoft knew exactly what the EU requested of them, and made very sure that they did precisely the opposite).
So on the one hand, Microsoft loudly proclaim that they would seek an audience with representatives of the F/OSS community in order to "better understand how Microsoft will serve their needs" or somesuch (all the time knowing, having had it spelt out for them countless times, exactly what the F/OSS community need from them - tell us how to talk to your software; promise to compete on merit, not through patents), and on the other, they openly flout openness, and thumb their noses at the entire EU at the same time.
Have you noticed that they are becoming more 'open' lately, whether it be in ideas or practice? They already have the market share, they already have a remarkable stream of money coming in, and it will be that way for a long time. So, with money taken care of, they can become more open and reap in the benefits of that. News sources will carry the story and they will be seen as heroes of the people. The attitude toward Microsoft by geeks will also soften (not by much, but still...).
No, but I have noticed how they have made a big show and dance about how "open" they are becoming, without yet really delivering on it, apart from a few minor projects donated to open source. I'm perfectly willing to believe an old dog can learn new tricks, but I'll believe it when I see it. If a human being lied as much as Microsoft had, he'd be referred to a psychiatrist - honestly, it's like something out of comic book! Unbelieveable. But I digress:)
Your point about revenue streams is perhaps more important to this discussion than you realise. Microsoft make substantial profits on only two products: Windows itself, and Office. On every other line they either make a marginal profit, barely break even, or make a huge loss. Let's put that into perspective here: Microsoft has dozens (hundreds?) of product lines. They employ thousands of staff, and their costs are very large. And this is all supported by just two products. Both are under attack: Windows by Linux and Mac OS X (the latter of which is alledged to be superior to Windows in many respects, although I cannot judge for myself as I have never even seen it in action); and Office by OpenOffice.
Linux on the desktop has a huge load of problems, and it will be at least two years until it can rival Windows for ease of use, if ever - the list of complaints from new users is endless. However, the list of complaints from would-be convertees to OpenOffice is more interesting. In brief, very few people complain about the feature-set of OO.o when compared to Office. Do you know what the No. 1 complaint about OO.o i
Microsoft announces that they are going to do something that Slashdot has wanted them to do for ages, and Slashdot proceeds to find faults with it?
More like "Microsoft states that it is doing something Slashdot has wanted them to do for ages [open up their formats in such a way that they compete on a level playing-field and compete on merit, rather than lock-in plus the network effect], and slashdot finds the faults that show that Microsoft are in fact doing no such thing".
It's actually memory leaks; there's a whole bunch of the buggers. Lots of leaks have been found and fixed, so I'm betting it won't be as bad as it used to be, but I doubt they've got them all, alas.
You're right to be dubious in general as there is a depressing trend of governments using Linux to threaten Microsoft into reducing their prices (but, hey, whatever puts the squeeze on Microsoft is fine by me;)), but the Munich thing is a done deal. Munich simply does not want Microsoft, and if they aborted their plans and went back to the bosom of MS, I would publicly eat my own head.
Slashdot Mirror
Linux takes aeons to boot to a usable desktop, whereas I find Windows much faster.
Certain apps (notably Firefox and Thunderbird, which both use XUL) visibly redraw when you switch to them under Linux, but I see nothing like this under Windows. You can actually see them filling themselves in!
Most apps start up slower under Linux, although this is mainly due to the need to dynamically link apps as they are started. Nevertheless, even with preloading on, launching Explorer in Windows is faster than launching Konqueror.
Dragging a gaim Window over Firefox leaves half-drawn copies of gaim all across my screen. Bigger windows are even worse - I can fill the entire screen with half-drawn copies of windows just by waggling the mouse!.
Again, YMMV, etc.
Anyway, point by point:
On this, we'll have to agree to disagree; I've seen XP crash a handful of times: a few times when it turned out my graphics card was faulty (this crashed Linux, too) and once when I was copying a file from my Linux comp to a friend of mine's XP comp via samba. The latter is inexcusable (but funny side-note: my friend blamed "Linux's crappy Samba implementation", even though it was his computer that crashed, not mine!). The rest of the time, my XP computer at work truly does run and run, requiring a reboot only when a critical Windows Update is required. Your experiences are apparently different. This is true - KDE has been getting faster and more memory-friendly - check out the "top" output for KWrite (or is it Kate) under the KDE4 prototype code. I'm very pleased with the progress of KDE, and a recent talk by Robert Love (on optmising GNOME) shows that the Desktop Linux developers are very committed to reducing bloat, which I couldn't be happier about. However, the fact that KDE is getting faster either tells us that something that's always been good is getting better, or something that was slow and memory hungry before is getting better - much like Mac OS X, which started out dog-slow but which has been improving in speed with each successive release. Respectfully, I'd have to say it was the latter: on my 256MB laptop, KDE starts to swap more and much sooner than XP does (i.e. with fewer apps open). Firefox consumes far more resources than IE (although admittedly it also accomplishes far more). Having said that, the focus on the Linux side is on getting faster (I'm drooling with anticipation at XGL), so on the speed/ memory consumption side I see Linux ultimately winning out. Cross-platform-ness is admirable, but a shell where objects are first-class citizens sounds pretty good to me. I determine the power of a shell by how much easier it will make my life.It's pretty weird, yeah, and very obscure - I found it by sheer fluke. Of course, it's only fair to point out that CTRL+C is also expected to shut down the current foreground process in a CLI environment, so I can see why CTRL+C does not act as "Copy".
Flash-forward to Windows 2000/ XP, and Microsoft apparently accomplished a miracle, producing a version of Windows that would literally run and run, and was still fairly nippy. Meanwhile, the writers of Linux Desktop Environments were discovering that it's very easy to be fast and light when you don't do much, or aren't particularly user-friendly, and that increased functionality almost always comes at the price of bloat.
So these people saw two pillars of the superiority of Linux (speed and stability) snatched away from them. The truly curious thing is what happened next: instead of being spurred into action by this new competition and addressing these concerns on the Linux side, these people instead simply went into a state of denial, and refused to let go of these cherished (and rapidly shrinking) areas where Linux once scored over Windows. Read through any anti-MS slashdot article on any given day and count the number of horribly outdated criticisms of Microsoft you see (BSOD's; bloat; Clippy(!)) - as a passionate believer in F/OSS, it really grieves me to see people behaving like this, rather than aiming to improve Linux to the state where it once again has many advantages over Windows.
Flash-forward to now, as one of the other areas in which Linux scores over Windows (a UNIX command-line is an awesome and enjoyable tool to use; the Windows command line, by contrast, is a rubber hammer with nails in the handle :)) may well be snatched away, and we see the same thing: people are hoping against hope that Microsoft foul it up, because if they don't another area of Linux superiority disappears, along with another shred of their self-esteem. This, I think, is why people care, and why they do not wish Microsoft well in this project, however helpful it may be to the common good.
You can already copy from/ paste to Window's default CLI (highlight text and press ENTER to copy; right-click to paste - same with cygwin). I don't know about cutting, though.
Ah, fair enough. I think this is almost certainly covered by the inotify mechanism and, if not, could be added in relatively short order. Linux is usually pretty agile that way :)
I've not used it myself, but look into Beagle. With Linux, the area corrsponding to "user space programs that are outside of the kernel developer's control" is the province of the distros maintainers. I have absolutely no idea how this ties in with cp, though - why does cp need to change? The Linux inotify Kernel patches (which Beagle uses to monitor for changes) work at a deeper layer than cp et al.
Since then, though, there has been a *massive* amount of work done, mostly in the areas of adding new features or increasing the power of the libraries. This has left Desktop Linux in a rather buggy, bloated state, but I think it is beginning to settle down slightly and tend towards b) and c) (the kernel itself seems to have a recent flare-up of new stuff being added, to the detriment of its speed and stability; doubtless, the latter two qualities will reassert their dominance once the frenzy wears off). Robert Love gave a very interesting talk on optimising GNOME (most of the ideas apply to KDE, also), and it seems that a lot of developers have a keen interest in getting Linux to run on old hardware - particularly the Ubuntu team, who have the resources and know-how to get the job done.
OO.o is a lost cause, alas, although I have heard good things about KOffice. Some have stated that it is approaching the functionality of OO.o, but is still fairly lightweight - can anyone who has used the two for "proper" work possibly comment on this?
Oh, and while I love Firefox, I find it odd that you were hail this as a prime example of "smaller and faster, not bigger and slower." :)
I personally don't even notice it any more (probably because the only GTK app I use is gaim, and very occasioanlly, GIMP :)), but I guess it's confusing for most new users. Has anyone tried out gtk-qt? Does it work at all well? Obviously it can't handle entirely different UI conventions, but I guess it could possibly make things at least look the same.
So the major problems here would be settling on some kind of Linux Standard Desktop Base of libraries, and persuading the community that "meta-debs" are a good idea.
While I agree that these problems need to be tackled, you are flat-out wrong if you think that people aren't already working very hard on the problem. The fact is that (in my humble opinion), in spite of the fact that GNOME and KDE started up way back in '98 and '99, Desktop Linux has been almost entirely ignored prior to, say, a year and a half ago. Back then, if you asked on a forum about how to get your USB pen to automount when you plug it in, you would be told to write your own script for monitoring the tail of dmesg and respond to USB events my parsing the output, as if this were a completely acceptable solution.
Nowadays, though, with iniatives such as Ubuntu, this kind of attitude doesn't fly: sure, if the problem you have can't currently be solved in a neat, elegant way then they'll give you a Grandma-unfriendly piece of advice, but you can bet that someone will begin looking at a way to fix this specific problem elegantly for the next release. In short, there has been either a shift in attitude (or perhaps just an infusion of new blood) whose mantra is "If it doesn't work out of the box, it is a bug; report it to us, and we will fix it if we can" rather than the "If it doesn't work out of the box, fix it yourself" of yesteryear.
This development is a recent phenomenon, however, and Desktop Linux has a *huge* amount of catching up to do (both MS and Apple have a very big headstart in terms of time and resources that they have already thrown into the problem). Adding the polish and design cohesiveness required for "usability" is a very, very hard problem that takes a big investment of time and resources to address, but these resources are being levelled at the problem right now, and I tire of the constant barrage of posts who seem to think that the Linux community consists entirely of developers who know and care nothing for end-user usability and who seem to think that sitting back and saying "why don't you just make it usable?" is some kind of fantastic idea that had simply never occurred to the myopic developers, rather like someone watching a medical programme and shouting out "Hey, why don't you Doctors and Scientists try and cure cancer!"
Sorry, just venting :)
They understand perfectly that technical solutions on their own aren't always tenable; they also understand that technical solution + threat of lawsuit == "teh win".
It's nuts, isn't it? Especially with something as utterly generic (and probably free of "valuable" IP) as a sodding wireless card where I can pick up one just as good for $20 on eBay. I mean, I can understand high-end graphics cards manufacturers not giving out their specs as there probably is some very interesting stuff in there that has taken them a long while to develop, but a wireless card or a WinModem? Stupid, self-defeating and petty. Thank God for enlightened manufacturers like ralink - as a Linux user, I'm practically falling over myself to buy one of their nifty USB wireless-g pens :)
I've seen one of mine pop up, reposted anonymously. I feel so violated :(
It most definitely affects Firefox on Linux, too. It depends a lot on your browser habits; the leak seems to get worse every time you open a tab i.e. closing a tab does not appear to free all of the memory that was allocated when it was opened, and this makes using Firefox with my personal usage pattern a nightmare :) I used Firefox fairly intensely under Linux last weekend, and eventually my 1/2GB of RAM and my 1/2GB of swap were all used up. Firefox had just ten tabs open at the time. The time taken to close Firefox, from clicking OK on the "Close all tabs?" prompt to the ceasation of the flurry of disk activity was at least one and a half-minutes (7200RPM SATA, 8MB cache). At this point, the memory and swap consumption reported by top both instantly halved :)
Memory fragmentation should, but there is always the risk of an errant part of your code holding references to an object which is no longer needed, possibly until the program is closed. The situation is much-improved with languages that have garbage collection as it is more automated and remove some of the burden from the programmer, but the problem of determining when an object will definitely never be used again by a program and thus making it eligible for garbage collection is in general only solvable by strong AI (having no references to an object reachable by a chain of references stemming from the program root is a good heuristic for deciding that an object can be disposed of, but there are always cases it will miss).
Memory fragmentation is a big issue for modern desktop systems as the heap used by programs written in C/C++ can't be compacted, and most memory allocation systems weren't necessarily designed to support programs that would be continually allocating and deallocating memory for days on end. Robert Love gave a (fairly detailed and technical) talk on it at while back, with some suggestions for combating it on the Linux desktop, which I recommend to anyone who is interested. It's about 126MB, Ogg format.
http://stream.fluendo.com/archive/6uadec/Robert_Lo ve_-_Optimizing_GNOME.ogg
The thing of it is that Microsoft has deemed Linux internally to be Public Enemy Number 1 (see: The Halloween Documents) and have a track record of a) not suffering competition to exist, unless it benefits them some way; and b) fighting dirty. In brief, they have lied so many times and built up such a great deal of bad blood with the F/OSS community that, understandably, no one believes a single word they say. As for becoming more "open" - witness the current EU situation, where the EU states that their use of proprietary protocols (e.g. Samba) is anti-competitive, and Microsoft's resulting proposal - effectively, following the letter of the law but willfully acting in direct opposition to the spirit of the law (which Microsoft understood full well; Microsoft knew exactly what the EU requested of them, and made very sure that they did precisely the opposite).
So on the one hand, Microsoft loudly proclaim that they would seek an audience with representatives of the F/OSS community in order to "better understand how Microsoft will serve their needs" or somesuch (all the time knowing, having had it spelt out for them countless times, exactly what the F/OSS community need from them - tell us how to talk to your software; promise to compete on merit, not through patents), and on the other, they openly flout openness, and thumb their noses at the entire EU at the same time.
No, but I have noticed how they have made a big show and dance about how "open" they are becoming, without yet really delivering on it, apart from a few minor projects donated to open source. I'm perfectly willing to believe an old dog can learn new tricks, but I'll believe it when I see it. If a human being lied as much as Microsoft had, he'd be referred to a psychiatrist - honestly, it's like something out of comic book! Unbelieveable. But I digress :)
Your point about revenue streams is perhaps more important to this discussion than you realise. Microsoft make substantial profits on only two products: Windows itself, and Office. On every other line they either make a marginal profit, barely break even, or make a huge loss. Let's put that into perspective here: Microsoft has dozens (hundreds?) of product lines. They employ thousands of staff, and their costs are very large. And this is all supported by just two products. Both are under attack: Windows by Linux and Mac OS X (the latter of which is alledged to be superior to Windows in many respects, although I cannot judge for myself as I have never even seen it in action); and Office by OpenOffice.
Linux on the desktop has a huge load of problems, and it will be at least two years until it can rival Windows for ease of use, if ever - the list of complaints from new users is endless. However, the list of complaints from would-be convertees to OpenOffice is more interesting. In brief, very few people complain about the feature-set of OO.o when compared to Office. Do you know what the No. 1 complaint about OO.o i
Also, OpenOffice itself uses zip files for most of its formats.
Forgot to add - check out the Session Saver extension. Makes closing and re-opening a tab-heavy session pretty much painless.
It's actually memory leaks; there's a whole bunch of the buggers. Lots of leaks have been found and fixed, so I'm betting it won't be as bad as it used to be, but I doubt they've got them all, alas.
You're right to be dubious in general as there is a depressing trend of governments using Linux to threaten Microsoft into reducing their prices (but, hey, whatever puts the squeeze on Microsoft is fine by me ;)), but the Munich thing is a done deal. Munich simply does not want Microsoft, and if they aborted their plans and went back to the bosom of MS, I would publicly eat my own head.