But I'd venture a guess that with OSX converting Linux users left and right that it'd be around 6-8% by now. Thoughts?
I think that's wrong - firstly, even if every desktop Linux user on the planet went to OS X (not going to happen), userbase would rise by about 2% probably, not 4-5%.
Secondly, I'd questions the assertion that Linux users are switching left and right. Out of all the Linux users I know, not one of them have switched to MacOS (although one or two have bought Macs and installed Linux on them).
But so what? There's no real way of measuring, so baseless assertions are pointless. Nobody knows, all you can do is guess, and I'd guess you're very wrong.
The public are mostly morons-- and since when have they ever done anything but look for the absolute cheapest of [product]?
Hmm, they don't agree with your purchasing decision - therefore they are morons. Did it ever occur to you that maybe they considered the cost/benefit ratio and decided to buy a PC based on sound consumerist reasoning? Clearly not.
They don't realize that while the Mac costs twice as much, it also remains a viable computer twice as long
They "don't realise it" because that statement could easily be bollocks. Evidence please.
I got more than six years out of the last brand new desktop Mac I bought (a Power Mac 7600, with a few modest upgrades sprinkled into it over the years to keep somewhat current), and could've gotten more but I wanted a machine that would run OS X
This has to be the weakest Mac argument I've ever heard. Just because Apple don't hike the performance as often as the PC makers do, does not mean they "last longer". There's nothing stopping you from never upgrading your computer whatever you use, but if you want the latest tech, you have to either upgrade your old one (which isn't an option with apple anyway) or buy something new.
Saying "my Mac lasted for years until I decided to upgrade" is silly, it just makes it look like Apple move far slower than the competition (which is sort of true, isn't it)
Is there a band that has artistic ingenuity or a political point? They won't get a contract, because the record company won't take a risk.
That's a gross exagguration and overly cynical. Of course there are such bands. Tatu have been number 1 in the charts here for three weeks, nothing remarkable there except they are from the Eastern Bloc and the song is about homosexuality.
And of course, if you like trance all this is largely irrelevant anyway, there's a constant stream of fantastic tracks coming out of names big and small. Those guys don't like the rock-star life style. They frequently change their own names in fact. It's purely about the music.
Where CD's only cost a few bucks instead of $13-15.
What, so you want music to be even more corporatised than it is now? The idea of corporations sponsoring concerts made me shudder, they already sponsor way too much as it is.
They won't all pay of course, but if you state that you need £1000 for the feature, then the people who really want it will pledge money for it
Damn. When I was writing that post, I meant to include this link , but got chatting to a friend and clean forgot.
Basically that bug shows Ben Bucksch being paid to hack on a feature for Mozilla. He hunted down the bug (Roaming Profiles), and found there were people wanting to scrap legacy NS4.7 networks but who were limited by the lack of roaming in Mozilla, and he managed to raise the funds and get paid to work on it.
We need to find a way so that people can make money producing (as opposed to "using") free software, without compromising the spirit of free software.
I don't think that's so hard.
Software isn't perfect. It's practically guaranteed that somebody, somewhere, is right now cursing some piece of softare. Maybe it's buggy, perhaps it's missing a feature they want. Maybe it costs so much they can't afford it. Maybe it's obsolete.
If all software in the world was under the (L)GPL, we could stil make money, by eliminating imperfection. Let's say I need some cash. OK, so I go bug hunting. A quick bugzilla query... what features have the most votes? Hmm, this one is pretty popular. It's a lot of work, and the maintainers are busy with other stuff. It's not got done. There are 200 votes. I think it'll take me a month. If each one of those people who voted chips in £5, that's £1000 for a months work, not bad at all.
They won't all pay of course, but if you state that you need £1000 for the feature, then the people who really want it will pledge money for it, and the ones who thought it'd be neat but don't care enough to pay will just wait it out. Eventually it'll get there.
Because, the model we use currently is actually very inefficient. Companies attempt to predict what all their customers want, and then write the code, and then sell it. What if really their customers wanted something different? Your shafted. Worse, because the culture is that you don't pay for bugfixes, new features are constantly introduced, with little thought given to whether they are actually useful or not.
So, freelance work on free software is more efficient. The example I gave above is less likely than a business saying "I need the software to do X, how much will it cost" - it's more efficient for them, because they only pay for what they need, and it's more efficient for us, because everybodies contributions are lumped together and we can all co-operate.
I call it the bounty hunter model, you hunt for bounty. Maybe a company wants you to port their apps to Linux using WineLib. Maybe a film company needs a new feature in the Gimp. Maybe an ISP is concerned about the security in the networking stacks and wants an audit.
And for new projects? Well, that's what volunteers are for:)
Whatever. Basically the 80% of people that work in software services becomes 100%. I think it's workable.
Re:Not All's Well that Ends Well ...
on
The Linux Uprising
·
· Score: 3, Insightful
Should Microsoft ever truly respond to the Linux threat, say by slashing their prices of Windows XP/Windows 2003/Windows Whatever in half, and slash the prices of Microsoft Office in half (much as they have already done in a recent promotion for Apple Macintosh users), it's game over for Linux on the desktop. Xandros is $100. LindowsOS is $130.
Well.... not really. One of the strongest desktops around is Redhat 8, which is free. You can't get cheaper than free. Also, remember the reason those distros are so expensive - proprietary NTFS resize code and CrossOver. As of about 4 days ago, we have stable open source NTFS resizing. That slashes quite a bit off the price. Xandros is already selling a version without CrossOver that comes in at a far more reasonable price.
Regardless, just because some companies charge a lot for Linux now, doesn't mean that this is what Linux costs. The fact is that Windows could be given away for free, and it wouldn't hurt Linux one bit in terms of development speed - how many free software hackers do you know who do it because they are too poor to buy Windows?
Re:The Romanticizing of "The Linux Uprising"
on
The Linux Uprising
·
· Score: 3, Insightful
I dunno. Big business has been using Linux for years, it's still a good story. I don't actually go out of my way to tell people about it, but sometimes people ask me stuff about it, and then they want to know the story behind such an unlikely thing.
They always find it interesting - despite the fact that it's about technology, the core story is a human one. It's about people who mostly have never met each other working together to achieve something totally amazing on a scale - nothing like it around.
So yes, some people use it because it's non mainstream. I expect when everybody uses Linux they'll go use some new ultra-cool OS with no apps or whatever. The story doesn't become any less interesting though.
That's cute, but I'm waaaaaaay more concerned about this part:
Before using open-source software, tech companies must sign a license in which they promise to give away innovations they build on top of it.
WTF? That has to be one of the more dangerous pieces of bad reporting I've seen lately. Not only is it utterly inaccurate (you don't have to sign anything to use open source software), it also hopelessly confuses "code" with "innovation".
Sure, point taken, but I know lots of good games that didn't follow formal creative design, then coding rules. Frozen Bubble? bzFlag? Even Tux Racer. Programmers are not magically exempt from being creative.
A game doesn't start with programmers. A game starts with an idea. A concept. A concept that is then fleshed out by writers, artists, etc...
Yep, absolutley. I mean, look at Tetris. That was started with a programmer first, they even completely forgot the plot writers, and look where that ended up!
I mean, damn, you wouldn't want this thing to turn into another Tetris would you. That'd be pretty sucky.
Considering that I can't really remember much about being 12, even though I'm only 18 now, I'd probably tell myself to go do something interesting.
Oh, and I'd probably tell myself to go on that bike ride with katie, she might be a bit wierd but she's also damn hot, and that kind of thing doesn't happen as often as TV makes you think it will.
I'm just repeating what I read. Yep, in a perfect world well documented interfaces would rule the day, and you'd never need to see the code. In practice, code is not perfect, sometimes there are bugs in the kernel, sometimes overriding a particular function and replacing it with a faster one gives you the performance boost that puts you ahead of the competition etc...
So we produce all this great stuff, but you don't think it's worth making it easy for other people to use it? Why shouldn't we share our toys, and make it pleasant to do so?
Well, maybe you get blown off because you go about things in the wrong way? Hint - constantly talking about forking things is the wrong way. Flaming people is another wrong way.
True, but the mac used the shift key for discontinuous
selection. I'll need to change ctrl to shift when I'm
get around to forking Nautilus. Just one more M$ism to
hunt down and kill (sigh)
... and another from google....
My choice after I quit film school was either to be a script writer for porno flicks or a linux UI designer. And to tell you the truth,
there's hardly any difference.
I'm sure you mean well, but having studied Psychology (which I have also), and read some materials makes you better informed about usability, not an expert.
Also, if you want to see the state of Linux usability improve, talking on mailing lists is fine, but there's talk and then there's action. Have you sat down and written written patches for the GNOME HIG (there are plenty of bugs against it in bugzilla, no coding needed)? Have you taken an application and gone through filing HIG bugs? Have you learnt basic coding and fixed some applications yourself?
In general, Linux usability won't improve by people debating it on Slashdot, it'll improve when people take the initiative.
At every step of the way the world of open source and free software has done everything possible to keep us out of the process.
Sorry, that's not true. Go look at the gnome usability lists. They have several professionals on board, and some of the core coders have read up on plenty of HCI material.
There are some interesting views on various Apple usability booboos here.
In particular the invisible font preview panel is pretty wierd, especially the way somebody who is apparently an Apple developer talks about it like it's a perfectly sane thing to do.
It always runs under single user mode. IRC somewhere, and you IRC as root. STUPID.
No, we're the stupid ones. Sorry, but the user/root distinction makes sense on a server but not on a single user desktop. It just gets in the way.
What you say? It's insecure? Perhaps the solution then is to write more secure apps, bearing in mind that if somebody cracks your IRC client they can still wipe out your personal data, execute programs at startup, turn your box into a warez box or whatever. Not being root doesn't stop any of that.
That move by Lindows was a wakeup call. Sorry, but it was. We need to make security less of a pain in the ass.
I've just read the rebuttal. To be honest, mosfet once again proves he couldn't argue his way out of a wet paper bag. Let's see, where shall we begin.
The first few paragraphs are contentless fluff. He talks about how KDE provides premade GUIs etc - so? Every other windowing system does this too. A choice quote:
Virtually all dialog boxes and keyboard bindings are identical because the application developer doesn't actually define them themselves - KDE provides them for you.
Perhaps his definition of dialog box is different to mine, but most useful apps provide at least some functionality via them. To claim that KDE provides every single dialog box is ridiculous, it does not, and even when apps have custom dialogs they still need to be internally consistant.
Most of the best parts of the KDE API were written by just a couple of people or an individual. A lot of free code that has been commercially influenced now is pretty unmaintainable and primary developers have left to work on other projects.
Assertions without supporting evidence do not make an argument. I don't know what that last sentance is meant to be, perhaps a jab at Nautilus, but I don't know of any "free code that's been commercially influenced" which is now "unmaintainable".
I'm constantly amazed at how people like Waldo Bastian, Dirk Mueller, David Faure, Lars, Simon, and many other large presences in KDE development volunteer to spend much of their time doing the "boring" stuff. As a matter of fact to say that volunteers don't do this, which basically is saying you need commercial companies to do it for you, is something I strongly disagree with.
Mosfet is either ignoring or ignorant of the fact that at least Waldo Bastian and David Faure have been employed to hack on KDE, and I recall somebody named Lars working for TrollTech on KDE a while ago, dunno if it's the same person. The last statement is far out - Havoc never even implied that, and even if he had, mosfet neatly demolished his own argument in the first sentance.
If there is a feature that enough people are willing to start a flamewar over, or even worse, switch applications, desktop environments, or OS'es over, you'd probably want to look at implementing that feature. Again it comes back to listening to your user base.
More distinctly dubious thinking. The people who make the most noise are not your "userbase". They are merely people who are making a lot of noise. You've always got to remember the silent majority who are just using the software, not bitching at the developers. That means you shouldn't add a feature just because a lot of noisy people want it, you should evaluate a feature based on the arguments for and against. Obviously, if it seems to be a popular feature take that into consideration, but the validity of the change should always come first.
His second point is rather amusing to me >:) In KDE you *can* autogenerate your GUI. I think you can in Gnome as well but in KDE it's done with QT designer and many of the dialogs in KDE are done in this manner.
"Auto-generating your gui" is not using a visual designer like QT Designer or Glade. Havoc was talking about people who thought you could for instance take a list of properties and dynamically generate the labels, layouts and edit boxes at runtime. Mosfet has totally misunderstood this throwaway statement, choosing instead to yet again pimp KDE and Qt, which isn't related to his argument and can therefore be safely ignored.
On the other hand hard coding something that people may want to modify and is relatively easy to make configurable lessens the value of your application. There is just no reason for it.
He apparently hasn't looked into the GNOME2 architecture in any depth. Most stuff isn't "hard coded", it's lifted from GConf. When Havoc talks about preferences bloat, it's mostly in relation to a good UI. Mosfet makes another comment earlier on about how having settings in config files upset users - odd, but I haven't heard of any users being upset about settings being in GConf, maybe because using a registry-style editor is actually easier than editing config files? Maybe it's to do with the choice of what to put in the UI as well.
I could go on, but most of the other counter-examples he gives can actually be turned around on him (kde does indeed have unbreak-me features). Although he makes a few fair points, the way he presents them makes his viewpoint far less compelling. Whenever Havoc makes a valid point, Mosfet responds by saying "I don't think this is true, and I think KDE does it well" - which is not a counter argument.
It is, though not impossible. Not every library on every linux system is GPLd, and you can always statically compile against the proprietary libraries (after you licensed them) that the GPL'd ones are based on. Ever notice commercial linux softwares tend to have gigantic executables?
That's not entirely accurate. In fact, virtually no system libraries are exclusively GPL, and in the rare cases that they are, they normally contain exception clauses that let you avoid using a free or open source license (or they are dual licensed).
The reason commercial programs are often so large is because we really suck at binary distribution. Making portable binaries is hard work, which we cunningly avoid by distributing software as source. Commercial software doesn't do that, nor can it make a new binary for every version of every distro, so they statically link stuff to keep themselves sane.
That's pretty broad. If I take that really literally, then if my networking program does DNS lookups in order to make it easier for users to specify what host they want to talk to, then my program is a "combined work" with a DNS server program?
IANAL, but... like most legal documents, there is a certain amount of slack involved, a certain fuzziness which is open to interpretation. Most of the time, common sense rules the day, and everybody lives happily spending their time Getting Things Done.
Occasionally there may be a dispute over the wording, or the interpretation. At that point, the two parties can normally resolve the issue between them. If they can't, then it goes to court, and the issue is decided there. If it's law, then it goes down as "case law" and in future judges will use it to aid them in their judgement. I don't know what the equivalent is for software licenses.
Basically the GPL is an agreement between two human parties. It implies a modicum of cooperation, of flexibility. If that cooperation doesn't exist, then the lawyers become useful:)
Slashdot Mirror
Well, according to OSNews, which is pretty neutral and doesn't tend to make up figures, it's about 2.9%, maybe a bit less.
But I'd venture a guess that with OSX converting Linux users left and right that it'd be around 6-8% by now. Thoughts?
I think that's wrong - firstly, even if every desktop Linux user on the planet went to OS X (not going to happen), userbase would rise by about 2% probably, not 4-5%.
Secondly, I'd questions the assertion that Linux users are switching left and right. Out of all the Linux users I know, not one of them have switched to MacOS (although one or two have bought Macs and installed Linux on them).
But so what? There's no real way of measuring, so baseless assertions are pointless. Nobody knows, all you can do is guess, and I'd guess you're very wrong.
Hmm, they don't agree with your purchasing decision - therefore they are morons. Did it ever occur to you that maybe they considered the cost/benefit ratio and decided to buy a PC based on sound consumerist reasoning? Clearly not.
They don't realize that while the Mac costs twice as much, it also remains a viable computer twice as long
They "don't realise it" because that statement could easily be bollocks. Evidence please.
I got more than six years out of the last brand new desktop Mac I bought (a Power Mac 7600, with a few modest upgrades sprinkled into it over the years to keep somewhat current), and could've gotten more but I wanted a machine that would run OS X
This has to be the weakest Mac argument I've ever heard. Just because Apple don't hike the performance as often as the PC makers do, does not mean they "last longer". There's nothing stopping you from never upgrading your computer whatever you use, but if you want the latest tech, you have to either upgrade your old one (which isn't an option with apple anyway) or buy something new.
Saying "my Mac lasted for years until I decided to upgrade" is silly, it just makes it look like Apple move far slower than the competition (which is sort of true, isn't it)
That's a gross exagguration and overly cynical. Of course there are such bands. Tatu have been number 1 in the charts here for three weeks, nothing remarkable there except they are from the Eastern Bloc and the song is about homosexuality.
And of course, if you like trance all this is largely irrelevant anyway, there's a constant stream of fantastic tracks coming out of names big and small. Those guys don't like the rock-star life style. They frequently change their own names in fact. It's purely about the music.
What, so you want music to be even more corporatised than it is now? The idea of corporations sponsoring concerts made me shudder, they already sponsor way too much as it is.
Damn. When I was writing that post, I meant to include this link , but got chatting to a friend and clean forgot.
Basically that bug shows Ben Bucksch being paid to hack on a feature for Mozilla. He hunted down the bug (Roaming Profiles), and found there were people wanting to scrap legacy NS4.7 networks but who were limited by the lack of roaming in Mozilla, and he managed to raise the funds and get paid to work on it.
That's the sort of thing I'm thinking of.
I don't think that's so hard.
Software isn't perfect. It's practically guaranteed that somebody, somewhere, is right now cursing some piece of softare. Maybe it's buggy, perhaps it's missing a feature they want. Maybe it costs so much they can't afford it. Maybe it's obsolete.
If all software in the world was under the (L)GPL, we could stil make money, by eliminating imperfection. Let's say I need some cash. OK, so I go bug hunting. A quick bugzilla query... what features have the most votes? Hmm, this one is pretty popular. It's a lot of work, and the maintainers are busy with other stuff. It's not got done. There are 200 votes. I think it'll take me a month. If each one of those people who voted chips in £5, that's £1000 for a months work, not bad at all.
They won't all pay of course, but if you state that you need £1000 for the feature, then the people who really want it will pledge money for it, and the ones who thought it'd be neat but don't care enough to pay will just wait it out. Eventually it'll get there.
Because, the model we use currently is actually very inefficient. Companies attempt to predict what all their customers want, and then write the code, and then sell it. What if really their customers wanted something different? Your shafted. Worse, because the culture is that you don't pay for bugfixes, new features are constantly introduced, with little thought given to whether they are actually useful or not.
So, freelance work on free software is more efficient. The example I gave above is less likely than a business saying "I need the software to do X, how much will it cost" - it's more efficient for them, because they only pay for what they need, and it's more efficient for us, because everybodies contributions are lumped together and we can all co-operate.
I call it the bounty hunter model, you hunt for bounty. Maybe a company wants you to port their apps to Linux using WineLib. Maybe a film company needs a new feature in the Gimp. Maybe an ISP is concerned about the security in the networking stacks and wants an audit.
And for new projects? Well, that's what volunteers are for :)
Whatever. Basically the 80% of people that work in software services becomes 100%. I think it's workable.
Well.... not really. One of the strongest desktops around is Redhat 8, which is free. You can't get cheaper than free. Also, remember the reason those distros are so expensive - proprietary NTFS resize code and CrossOver. As of about 4 days ago, we have stable open source NTFS resizing. That slashes quite a bit off the price. Xandros is already selling a version without CrossOver that comes in at a far more reasonable price.
Regardless, just because some companies charge a lot for Linux now, doesn't mean that this is what Linux costs. The fact is that Windows could be given away for free, and it wouldn't hurt Linux one bit in terms of development speed - how many free software hackers do you know who do it because they are too poor to buy Windows?
They always find it interesting - despite the fact that it's about technology, the core story is a human one. It's about people who mostly have never met each other working together to achieve something totally amazing on a scale - nothing like it around.
So yes, some people use it because it's non mainstream. I expect when everybody uses Linux they'll go use some new ultra-cool OS with no apps or whatever. The story doesn't become any less interesting though.
Before using open-source software, tech companies must sign a license in which they promise to give away innovations they build on top of it.
WTF? That has to be one of the more dangerous pieces of bad reporting I've seen lately. Not only is it utterly inaccurate (you don't have to sign anything to use open source software), it also hopelessly confuses "code" with "innovation".
Sure, point taken, but I know lots of good games that didn't follow formal creative design, then coding rules. Frozen Bubble? bzFlag? Even Tux Racer. Programmers are not magically exempt from being creative.
Rock on michael, I'm rooting for you :) Oh, and get Ballistics out, it's the one I'm most looking forward to, I'm a sucker for speed.
Yep, absolutley. I mean, look at Tetris. That was started with a programmer first, they even completely forgot the plot writers, and look where that ended up!
I mean, damn, you wouldn't want this thing to turn into another Tetris would you. That'd be pretty sucky.
Oh, and I'd probably tell myself to go on that bike ride with katie, she might be a bit wierd but she's also damn hot, and that kind of thing doesn't happen as often as TV makes you think it will.
I'm just repeating what I read. Yep, in a perfect world well documented interfaces would rule the day, and you'd never need to see the code. In practice, code is not perfect, sometimes there are bugs in the kernel, sometimes overriding a particular function and replacing it with a faster one gives you the performance boost that puts you ahead of the competition etc...
lol, very good :) I guess I was asking for that.... I still most of mosfets article was pretty lightweight though
So we produce all this great stuff, but you don't think it's worth making it easy for other people to use it? Why shouldn't we share our toys, and make it pleasant to do so?
That'd be the other 95% of the computer using populace then.
Here a fair point you make is masked by the way you express it....
One quote from a gnome list:
I'm sure you mean well, but having studied Psychology (which I have also), and read some materials makes you better informed about usability, not an expert.
Also, if you want to see the state of Linux usability improve, talking on mailing lists is fine, but there's talk and then there's action. Have you sat down and written written patches for the GNOME HIG (there are plenty of bugs against it in bugzilla, no coding needed)? Have you taken an application and gone through filing HIG bugs? Have you learnt basic coding and fixed some applications yourself?
In general, Linux usability won't improve by people debating it on Slashdot, it'll improve when people take the initiative.
At every step of the way the world of open source and free software has done everything possible to keep us out of the process.
Sorry, that's not true. Go look at the gnome usability lists. They have several professionals on board, and some of the core coders have read up on plenty of HCI material.
Here is a discussion of why user levels really suck.
In particular the invisible font preview panel is pretty wierd, especially the way somebody who is apparently an Apple developer talks about it like it's a perfectly sane thing to do.
No, we're the stupid ones. Sorry, but the user/root distinction makes sense on a server but not on a single user desktop. It just gets in the way.
What you say? It's insecure? Perhaps the solution then is to write more secure apps, bearing in mind that if somebody cracks your IRC client they can still wipe out your personal data, execute programs at startup, turn your box into a warez box or whatever. Not being root doesn't stop any of that.
That move by Lindows was a wakeup call. Sorry, but it was. We need to make security less of a pain in the ass.
I didn't even realise such a domain existed, but it does. It might be because "Great Britain" is technically different to the "United Kingdom".
The first few paragraphs are contentless fluff. He talks about how KDE provides premade GUIs etc - so? Every other windowing system does this too. A choice quote:
Perhaps his definition of dialog box is different to mine, but most useful apps provide at least some functionality via them. To claim that KDE provides every single dialog box is ridiculous, it does not, and even when apps have custom dialogs they still need to be internally consistant.
Assertions without supporting evidence do not make an argument. I don't know what that last sentance is meant to be, perhaps a jab at Nautilus, but I don't know of any "free code that's been commercially influenced" which is now "unmaintainable".
Mosfet is either ignoring or ignorant of the fact that at least Waldo Bastian and David Faure have been employed to hack on KDE, and I recall somebody named Lars working for TrollTech on KDE a while ago, dunno if it's the same person. The last statement is far out - Havoc never even implied that, and even if he had, mosfet neatly demolished his own argument in the first sentance.
More distinctly dubious thinking. The people who make the most noise are not your "userbase". They are merely people who are making a lot of noise. You've always got to remember the silent majority who are just using the software, not bitching at the developers. That means you shouldn't add a feature just because a lot of noisy people want it, you should evaluate a feature based on the arguments for and against. Obviously, if it seems to be a popular feature take that into consideration, but the validity of the change should always come first.
"Auto-generating your gui" is not using a visual designer like QT Designer or Glade. Havoc was talking about people who thought you could for instance take a list of properties and dynamically generate the labels, layouts and edit boxes at runtime. Mosfet has totally misunderstood this throwaway statement, choosing instead to yet again pimp KDE and Qt, which isn't related to his argument and can therefore be safely ignored.
He apparently hasn't looked into the GNOME2 architecture in any depth. Most stuff isn't "hard coded", it's lifted from GConf. When Havoc talks about preferences bloat, it's mostly in relation to a good UI. Mosfet makes another comment earlier on about how having settings in config files upset users - odd, but I haven't heard of any users being upset about settings being in GConf, maybe because using a registry-style editor is actually easier than editing config files? Maybe it's to do with the choice of what to put in the UI as well.
I could go on, but most of the other counter-examples he gives can actually be turned around on him (kde does indeed have unbreak-me features). Although he makes a few fair points, the way he presents them makes his viewpoint far less compelling. Whenever Havoc makes a valid point, Mosfet responds by saying "I don't think this is true, and I think KDE does it well" - which is not a counter argument.
That's not entirely accurate. In fact, virtually no system libraries are exclusively GPL, and in the rare cases that they are, they normally contain exception clauses that let you avoid using a free or open source license (or they are dual licensed).
The reason commercial programs are often so large is because we really suck at binary distribution. Making portable binaries is hard work, which we cunningly avoid by distributing software as source. Commercial software doesn't do that, nor can it make a new binary for every version of every distro, so they statically link stuff to keep themselves sane.
IANAL, but ... like most legal documents, there is a certain amount of slack involved, a certain fuzziness which is open to interpretation. Most of the time, common sense rules the day, and everybody lives happily spending their time Getting Things Done.
Occasionally there may be a dispute over the wording, or the interpretation. At that point, the two parties can normally resolve the issue between them. If they can't, then it goes to court, and the issue is decided there. If it's law, then it goes down as "case law" and in future judges will use it to aid them in their judgement. I don't know what the equivalent is for software licenses.
Basically the GPL is an agreement between two human parties. It implies a modicum of cooperation, of flexibility. If that cooperation doesn't exist, then the lawyers become useful :)