However, I feel that the issue here that Miguel brought up is that instead of working on all these side issues, Gnome should spend more time making the desktop better, as that's what users are going to see most often.
I think that's been the main target of GNOME for a while now, Havoc and a few otherse have been advocating such an approach for quite some time.
And I don't think the person who wrote the e-mail you quote was trying to start one either. But it seems like you are.
You do realise that "luge" is Luis Villa, Ximian bugmaster and somebody who has contributed greatly to Gnome? He's also one of the most persistantly calm and rational people working on the project. This guy is no troll, Miguel wrote some stuff, and most of his arguments were challenged. As the submitter of the story neglected to do so, Luis made sure the other side was shown.
I did a reply as well, here. It wasn't in the thread because I joined that list specifically to reply to Miguel.
This all scares me, it seems that the community has an important decision to make right now, between KDE and Gnome.
There is no decision, there is no spoon, there is no decision. Whoa.
If Linux's total victory over Microsoft meant Gnome winning over the Linux desktop with it's simplified interface, then I'd much rather a Linux with a small market share. Customizability, control, is something I love in Linux.
There are two separate issues here which you are conflating:
1) User interface.
2) Developer platform
They are separate. You can run GNOME apps in KDE and vice-versa.
GNOME has had its UI rewritten basically from the ground up to be in alignment with usability principles. That doesn't actually mean making everything non-configurable, see my desktop - I'm sure KDE can do panels like that, but I've never really seen it done. You see the line in the bottom right hand corner? That's a slide-out desktop switcher. I tweaked the timings in GConf so it pops out very fast, then there's a delay if the mouse moves out before it slides back in again, slower.
That's configurability. I had to use GConf to do it, but so what? GConf is just a generic preferences tool. Having all those timings in the GUI would have been braindead, I'm probably nearly alone in tweaking things in this way.
So, the idea that GNOME has no configurability is wrong IMO. It had a lot of garbage stripped out, and some useful features disappeared with it. If they are in fact useful, for most people, and they aren't working around brokenness, my experience is that they're accepted with minimal fuss.
In other words, I have no idea which should win the dektop, but it's a very important decision.
There is no winning, there is no war. There are only standards. If one company decides to use the KDE APIs, that's no skin off my nose, they clearly decided the cost was worth it. If they decide to use GTK/GNOME, that's just fine too. It's irrelevant to which desktop I use. They'll work fine in anything, even Waimea, Enlightenment, Window Maker, all those other desktops.
Someone help!
Reflect upon the situation. I think you're making a mountain out of a molehill;)
Well, it wouldn't be too expensive for me, being intro trance and dance, most of the good tracks are singles or compilations anyway.
Having said that, it WOULD be too expensive for me, because it's almost certainly heavily tied to the Mac as a platform. That's a bad idea. Even if they made it available on other platforms, QuickTime blows goats on anything other than the Mac, so it's still dead.
I think this is one of those things where Apple will have great influence but not much impact, like most things they seem doomed to be a minority player - famous, but still minority.
Well I don't use any KDE apps, but that's not a matter of principle, I simply don't use any. Back when I used KDE3 obviously I used some, but nowadays they're either GTK/GNOME/XUL/VCL (openoffice) or Wine. The only VFS these apps share is the kernel one. If I had a need that was solved best by a KDE app I'd use it, but I don't, and haven't for a long time now......
The main problem with LUFS is that you have to install it manually, it doesn't come with any distros, and it doesn't have GUI integration:(
However, the fact that Microsoft has withdrawn from the OpenGL ARB does not necessarily mean that they're going to stop supporting OpenGL on their OSes. As long as they allow OpenGL to run natively on future versions of Windows, I'm not too bothered by their withdrawal.
I was under the impression that they had stopped supporting it, isn't OpenGL on Windows frozen in time at about version 1.2? I don't think they're going to release updates for it.
The sucky ones don't stay big forever, because people will quit buying their products, even if they are better, if they don't like the company. Capitalism is a self correcting system. Its just sometimes we are too impatient and want the results unreasonably quick.
After over a decade of waiting for Microsofts monopoly to correct itself, I think it's safe to conclude that the system is in fact fundamentally broken. For some time now it's looked like the only thing that can move them from their position of power is Linux, which has to be one the biggest economic hacks in history.
Thanks to KDE's integration you can open kwrite (text editor) (or quanta, web stuff editor) and type in stfp://user@host/directory
Of course, unless you are a KDE fanboi who only ever uses KDE apps on a matter of principle, you might find LUFS more useful, as it allows you to mount remote systems via SSH directly into the filing system, making them seamlessly available to any app, regardless of whether it's linked against kde or not.
Or perhaps more helpfully, Jabber Inc, who provide all the support and stability you need for corporate IM.
Again, lots of different clients. I think Psi's GUI is nice. It certainly isn't as crufty as ICQ. But YMMV on this one.
Check out RhymBox. It's one of the coolest looking clients around, and Seb is willing to do corporate customisation - for a price of course;) Meanwhile I'm working on getting it to run on Linux via Wine.
Agreed. One stunningly poor man page I'm thinking of is that for "sudo", which attempts to give you a tutorial on BNF notation, and then right at the bottom shows you how to give somebody root access via sudo. D'oh.
The second worst is tar, which just says "man pages suck, use info", so you use info and find it wants to give you a flipping tutorial. Nowhere in that man page or info docs does it give you the command to extract a tarball, the most common operation. Double 'doh.
GTK on other platforms is/was dead, or kludgily ported at best.
GTK2 has great Windows and DirectFB ports. Nobody has bothered porting it to the Mac because that now has integrated X11. One thing you forgot to mention is that of course Qt is only GPLd on X11 platforms, so if the Gimp used Qt there probably wouldn't be a Windows port at all.
The idea that the Gimp should use C++/Qt/KDE is ludicrous. There have been talks about making it backend/frontend independant and sticking a KDE front end on it, but in general nobody could be bothered. It'd be a huge amount of work for..... for what? You can already theme GTK and KDE apps to look the same. The idea that Qt is light-years ahead of GTK is in my experience a complete myth, born more of people who take a quick glance at it and go "eurgh, objects in C", then write it off (as I did at first). Once you actually sit down and start working with the code, you begin to realise that first impressions were misguided
Winamp? Nope, the plugins are all configurable, plus Winamp itself. On top there are skins.
ICQ? Nope, configuration options are crawling everywhere and more are added with each version.
Over the last 3 or 4 years I've seen all but one of my friends leave ICQ, because the client took 20 minutes to load. So they released ICQ Lite. It still has skins. Now they all use MSN, because the client doesn't attempt to manage their life for them.
WinAmp 3 is not replacing WinAmp 2 at all - nobody I've met who's tried it thinks it's better than WinAmp 2.
In the Unix world there have always been two different sets of keybindings that people use, emacs keys and vi keys. I think that it is fair to say that the majority of unix users spend a lot of thier time in either emacs or vi. Gnome used to try to emulate some of the emacs default keybindings, but now they all seem to have been replaced with windows keybindings.
If you're going to go on a flamewar, please do some research. Go to Desktop Preferences, Keyboard Shortcuts and then choose "Emacs" from the text editing shortcuts dropdown box. There you go, emacs keybindings.
Another good example is the "too many clocks" problem.
That was a brilliant example of all that was wrong with GNOME. Who on Gods green earth needs 4 different type of clock?
So I was pissed off when I upgraded my version of gnome and half the applets I used had gone!
What, just like when you upgraded from KDE1 to KDE2, virtually all your programs had to be ported or rewritten? If you really care about those applets, port them!
Don't even get me started on window managers with maximise buttons!
Huh? What is wrong with a maximise button? Every UI I've ever seen has one of these, and no the Mac way of expanding to some apparently random size is not good, it's a pain in the ass.
Developers should remember who they are developing for, and give more precedence to unix traditions than to windows traditions.
Yes, and the GNOME developers are developing for non-geeks. If you have a problem with that, go fork the old gnome1.4 codebase and continue loading it up with more clocks or something.
Also, C++ is a much more natural fit for gui app development than C. Yeah, you can make a C library look somewhat OO, but if you've got C++ coding skills, why try to make a pig fly?
You're forgetting a few things, like C is far, far easier to use than C++. I must have learnt ten languages in my lifetime, most of them object oriented, but I still don't know C++. It scares me. I can normally read a language I don't know and figure out what's going on with a bit of effort. C++ eludes me. It's a mess. It appears to have several ways to typecast something. Multiple inheritance and generics makes any non-trivial codebase hard to understand. GLib C is a little easier, at least the base rules are more easily understood, but that's not so great either.
The good thing about GLib C is that it's very easily bound to other languages. The Python GNOME bindings for instance are a delight, and the GTKmm/GNOMEmm bindings are more C++ish than the KDE code is. For instance, you don't need a preprocessor to do signals, and you use the STL a lot more.
I know there are Python bindings for KDE. I've yet to find things that use them. On the other hand, for PyGnome there's Straw, the Redhat tools, for Perl::GTK there's the Mandrake setup tools and a photo gallery program I was working on a few months ago.... etc etc.
The language and framework baggage that the GNOME developers are saddled with make them work twice as hard to achieve the same thing.
GNOME has language baggage? You clearly haven't seen the ridiculous flamewars on kde-multimedia. Presented with a feature-rich, powerful multimedia framework that is light-years ahead of anything else (GStreamer) some of the KDE developers were objecting to it because despite the presence of KDE bindings to use it, you might need to write a GObject in C to extend it.
In fact with more complete bindings you wouldn't even need to do that. Nonetheless, it appeared to be their opinion that nobody could possibly want to use something based on the GObject object model, and made some very condescending remarks about it. It was even suggested that a new framework based on C++ was started, an utterly stupid idea, considering GStreamer hardly needs extending at the moment, and that it'd take 2+ years to get to where GStreamer is today. Nonetheless, that's how tied to C++ those guys were.
Unfortunately, KDE is still highly C++ centric. Perhaps this is no longer the case, but when I last looked KParts was simply a wrapper around the c++ classloader, you couldn't create a KPart in anything other than C++.
Thus, less time available to make things polished, bug-free, etc. Maybe this is why the GNOME leadership is pushing the super-stripped newbie desktop direction.
Instead of wildly speculating, why not go review the flamewars of the GNOME2 development era and you can find out why they simplified things right now. It's pretty simple - the majority of the developers felt there was a need for a clean desktop that had usability designed into the core. The fact that it's written in C had nothing to do with it.
Some registry upgrade strategy is required I believe...
They have one already though. Basically if you change the format of the keys (which only happens rarely), you have to migrate data across when you start, leaving the previous keys intact. Upgrades should in theory be painless, if they aren't it's a bug in the app.
Which means deletion of.gconf and.gnome* so defaults can be restored and things can run again.
Or you could just go and delete the individual config files, using the good old "cd" and "rm".
Don't give me "well, write it yourself," because I don't have the skill level, but more importantly, it's a ridiculous attitude to have to start with.
Why is it a ridiculous attitude? Tell you what, you go away and design this killer GUI, and come back when you've got it all figured out. If it's good, I'm sure somebody will be enthusiastic enough to help you.
Besides, if you want a graphic designer, I'm your man. Linux apps are always sorely lacking in the aesthetics department.
Sorry, but while you were whinging on Slashdot tigert, jimmac, everaldo and tackat were busy making GNOME and KDE pretty. By all means go and help them, but I find the latest GTK2/GNOME2 apps much better looking than their equivalents on Windows or the Mac.
Yeah right. 3-D on linux/Xfree SUCK ASS. Want compairsons? Go play X game (with port to linux) on windows and then play it on Linux. You get shit for framerates, and dont tell me you're different.
If you're seeing that, it's usually because either you have bad drivers, or because the game was originally developed for Direct3D and then an OpenGL renderer was added later. The OpenGL renderer usually isn't as well optimized in such games, so you get lower framerates.
In games that use OpenGL natively (ie have been well optimized) you normally get higher framerates than on Windows.
However, some of Petrely's remarks are just silly. For example, he thinks that KDE being "more feature rich" is a good thing. Sorry, but that's not true. Having lots of features and buttons and widgets may work for some users, others may prefer something simpler, and yet others may want a different set of complex features.
Agreed. He also seems to confuse "lots of stuff" with "features". Having 3 media players or whatever they're up to now is not feature rich, it's bloat. OK, OK, so I'm being picky, but I've yet to find features that I really miss from KDE. Sure, I liked the idea of having a mini-webserver in my panel, but I never actually used it.
And Gtk+ and Qt both make very inefficient use of the X11 APIs, giving X11 an undeserved reputation for being slow.
Ignoring the fact that both GTK and Qt have pluggable backends, which entails a very slight loss of tight integration, the real reasons that GTK2 is slower than GTK1.2 are
1) It is double buffered. This slows GTK down, but makes it feel faster and smoother to the user. Net win.
2) It uses anti-aliasing. XRender is still not finished, nor well optimized. Profiling the new gnome-terminal widget for instance showed that most of the time was spent inside the X server. That will speed up over time.
3) If you do opaque resizes of apps, you'll find the content area doesn't stick to the borders. That's due to a problem with the internal scheduler, not due to lack of speed on the part of GTK.
OK, that was an extreme example. Demurrage is where you have negative interest - savers are penalised by having some of their savings creamed off each month.
Obviously in any economic system there will be things you can't do, just like there are with capitalism (ie no abusive monopolies etc)
Just curiously... if all the linux users care about is open source, wouldn't the functionality of windows compared to linux IN SOME ASPECTS cause a flux of *nix users to use windows if they could fiddle with it as they liked? I mean besides server issues, windows is the way to go if your computer is really just a PC.
Before that would happen, Windows would have to be:
a) Free software and
b) No longer controlled by Microsoft.
That simply isn't going to happen, ever. Microsoft have no incentive to let go of Windows, and until that happens Linux will be as important as it always was, not because it's more stable or tweakable or whatever, but because it's owned by everybody.
They usually start with "if we just put these people in prison for doing things the way they want instead of the way we tell them, then there'll be XYZ benefit". Sometimes, they say "if we just steal XYZ from this group, we can use it for ABC purpose".
Well yes, obviously most suggestions are bunk. I didn't notice any stealing or prison sentances in my little suggestion before... obviously you have to separate the wheat from the chaff.
Here's a freedom-enhanced idea for everyone to get the new antibiotic for however cheap you can imagine: buy all the shares of the company at the market price and then give the drug away once you own it.
That might work once, but afterwards? It doesn't actually _change_ anything, it simply gets around it once, and you'll probably destroy the company at the same time.
See how everyone got the antibiotic they needed and no one's freedom was curtailed?
There are plenty of alternative systems that don't involve curtailing freedom - the idea of demurrage for instance.
Yeah yeah, I know it's bad form to reply to myself, but considering so far virtually all 15 replies are fairly similar it seems to make more sense than just picking one and replying to that.
A lot of peoples knee-jerk reactions seem to be "this guy is a communist!". What a pile of toss. I never even implied communism or socialism in the parent post, all I said was perhaps we should be trying to think of alternative systems that reward innovation but not by restricting its usage.
Come on people, use your imagination! I know none of us are economists, so what. OK, here's a random brainfart:
What if we made saving money illegal? Let's say that at the end of every month, any personal profit you had made you were forced to spend. It doesn't matter how you spend it, as long as you do. Clearly, everybody would have an average personal wealth of zero - if I earn £850/month and spend £600 on living, having fun etc that leaves me with £250. If I'm now compelled to spend that on something, I have the following choices:
a) Buy things with it. Nice food, clothes, games, computer upgrades, whatever. But, at the end of the day even the most material of us can't buy things forever - you'd simply run out of room. Obviously I wouldn't be able to buy anything with a price of over £250 unless I borrowed money off of other people...... which leads me onto
b) Give it to other people. Perhaps my friend rob wants to buy a new stereo system costing £500, he has £250 left over, so do I, so I give my surplus to him. After all, it's either that or I lose it. Maybe I give it to him on the condition that he repays me, but in practice that'd be quite hard to co-ordinate as if one month he dumped £250 on me, I'd now have £500 to get rid of, instead of only £250.
c) Invest it. This is the obvious thing to do. If somebody has a scheme and they want capital for it, they can go around and get surplus money off of other people, with a certain return. You'll have more money to get rid of next month, but hey, your lifestyle costs will probably go up to match it.
d) Buy insurance. Put some away for a rainy day right? Well, I can't save any more, but I can buy some insurance in case I lose my job, or suddenly I have to pay for something expensive. Insurance is clearly going to be a slop bucket of wealth - balancing the books so that in any one month enough wealth comes in to cover the usual payouts would be fun. Maybe governments or independant organisations would have a license to stockpile capital in case of huge unexpected payouts like in natural disasters or something.
e) Give it to society. I dunno what to do now. Another month, another paycheque. I don't feel like giving my money to a friend, nor investing it. Instead, why not make a proportion of my income a contribution to society? It could be spent on research, on traditional charity, on funding the education system.... whatever.
Companies obviously have to fit into this system. If they can't accumulate capital, how do they grow? Well, by borrowing from those with surplus, and repaying people in shares? Any month they have underspent, the money goes on dividends. Giving the director a nice fat $500,000 bonus isn't much use, nobody can spend that much in a month, not even on investments.
So, somebody disovers a new anti-biotic in some slime, and they want to turn it into a cure for a disease. Why? Altruism? Fame? The fear they or a loved one might get it someday? Why do people write free software? Who knows. Anyway, the investment needed to turn it into a working drug is huge. They go around, finding people who routinely make surplus capital, and ask for a proportion of it. After all, use it or lose it. Individuals who care for society (they do exist!) would donate, and their funding grows, and when the research is completed the drug is released to the world.
Well, it's almost certainly an incredibly stupid idea. It's probably riddled with holes, I can think of a few problems with it now. It might have been tried before. It'd probably be ridiculously unstable. The idea is 99% guaranteed to be junk.
BUT that doesn't matter! Don't focus on the idea. Ignore the idea. I don't want to see replies flaming me for what a stupid idea it is, I know that already. Focus on the meta-idea: we should be dreaming up alternatives to the current system.
Come on, give me some credit, I'm an incredibly poor economist but at least I'm trying. The current system IS a big hack, it DOES have big problems. The solution isn't to deride people who point this out by making smart-alec comments about socialism, the solution is to sit down and try and create something better. It doesn't have to be original or massively inventive, it just has to be better than what we currently have. Previous attempts failed - so what? BeOS and OS/2 died, yet we have Linux.
If so many people can sit down and bend the rules of the software economy by creating a complete OS from scratch, I think there is enough intelligence and willpower at least here to do the same for economics. It's worth a shot.
ps - I would never normally say this, but if you have some spare mod points please mod this up, I think it's very very important that people lose this lethargy over thinking about economics and society. I know we're mostly hackers, but all I see here is bitching about how evil the RIAA is, or software patents, or even microsoft when really they are simply products of the system. I want people to see this post and sit back and think about it, and maybe out of the discussion something cool will appear.
Slashdot Mirror
I think that's been the main target of GNOME for a while now, Havoc and a few otherse have been advocating such an approach for quite some time.
And I don't think the person who wrote the e-mail you quote was trying to start one either. But it seems like you are.
You do realise that "luge" is Luis Villa, Ximian bugmaster and somebody who has contributed greatly to Gnome? He's also one of the most persistantly calm and rational people working on the project. This guy is no troll, Miguel wrote some stuff, and most of his arguments were challenged. As the submitter of the story neglected to do so, Luis made sure the other side was shown.
I did a reply as well, here. It wasn't in the thread because I joined that list specifically to reply to Miguel.
There is no decision, there is no spoon, there is no decision. Whoa.
If Linux's total victory over Microsoft meant Gnome winning over the Linux desktop with it's simplified interface, then I'd much rather a Linux with a small market share. Customizability, control, is something I love in Linux.
There are two separate issues here which you are conflating:
1) User interface.
2) Developer platform
They are separate. You can run GNOME apps in KDE and vice-versa.
GNOME has had its UI rewritten basically from the ground up to be in alignment with usability principles. That doesn't actually mean making everything non-configurable, see my desktop - I'm sure KDE can do panels like that, but I've never really seen it done. You see the line in the bottom right hand corner? That's a slide-out desktop switcher. I tweaked the timings in GConf so it pops out very fast, then there's a delay if the mouse moves out before it slides back in again, slower.
That's configurability. I had to use GConf to do it, but so what? GConf is just a generic preferences tool. Having all those timings in the GUI would have been braindead, I'm probably nearly alone in tweaking things in this way.
So, the idea that GNOME has no configurability is wrong IMO. It had a lot of garbage stripped out, and some useful features disappeared with it. If they are in fact useful, for most people, and they aren't working around brokenness, my experience is that they're accepted with minimal fuss.
In other words, I have no idea which should win the dektop, but it's a very important decision.
There is no winning, there is no war. There are only standards. If one company decides to use the KDE APIs, that's no skin off my nose, they clearly decided the cost was worth it. If they decide to use GTK/GNOME, that's just fine too. It's irrelevant to which desktop I use. They'll work fine in anything, even Waimea, Enlightenment, Window Maker, all those other desktops.
Someone help!
Reflect upon the situation. I think you're making a mountain out of a molehill ;)
Of course you should note that this response was hardly unprovoked, a forum full of lamers flaming your distro would probably piss you off too.
Because according to a thread on the GNOME desktop-devel-list, they already have one!
Having said that, it WOULD be too expensive for me, because it's almost certainly heavily tied to the Mac as a platform. That's a bad idea. Even if they made it available on other platforms, QuickTime blows goats on anything other than the Mac, so it's still dead.
I think this is one of those things where Apple will have great influence but not much impact, like most things they seem doomed to be a minority player - famous, but still minority.
The main problem with LUFS is that you have to install it manually, it doesn't come with any distros, and it doesn't have GUI integration :(
I was under the impression that they had stopped supporting it, isn't OpenGL on Windows frozen in time at about version 1.2? I don't think they're going to release updates for it.
After over a decade of waiting for Microsofts monopoly to correct itself, I think it's safe to conclude that the system is in fact fundamentally broken. For some time now it's looked like the only thing that can move them from their position of power is Linux, which has to be one the biggest economic hacks in history.
Of course, unless you are a KDE fanboi who only ever uses KDE apps on a matter of principle, you might find LUFS more useful, as it allows you to mount remote systems via SSH directly into the filing system, making them seamlessly available to any app, regardless of whether it's linked against kde or not.
Or perhaps more helpfully, Jabber Inc, who provide all the support and stability you need for corporate IM.
Again, lots of different clients. I think Psi's GUI is nice. It certainly isn't as crufty as ICQ. But YMMV on this one.
Check out RhymBox. It's one of the coolest looking clients around, and Seb is willing to do corporate customisation - for a price of course ;) Meanwhile I'm working on getting it to run on Linux via Wine.
The second worst is tar, which just says "man pages suck, use info", so you use info and find it wants to give you a flipping tutorial. Nowhere in that man page or info docs does it give you the command to extract a tarball, the most common operation. Double 'doh.
GTK2 has great Windows and DirectFB ports. Nobody has bothered porting it to the Mac because that now has integrated X11. One thing you forgot to mention is that of course Qt is only GPLd on X11 platforms, so if the Gimp used Qt there probably wouldn't be a Windows port at all.
The idea that the Gimp should use C++/Qt/KDE is ludicrous. There have been talks about making it backend/frontend independant and sticking a KDE front end on it, but in general nobody could be bothered. It'd be a huge amount of work for ..... for what? You can already theme GTK and KDE apps to look the same. The idea that Qt is light-years ahead of GTK is in my experience a complete myth, born more of people who take a quick glance at it and go "eurgh, objects in C", then write it off (as I did at first). Once you actually sit down and start working with the code, you begin to realise that first impressions were misguided
Desktop Preferences -> Keyboard Shortcuts -> Text Editing shortcuts -> Emacs
woop woop - Ctrl-A now goes to the beginning of the line.
Happy now?
Over the last 3 or 4 years I've seen all but one of my friends leave ICQ, because the client took 20 minutes to load. So they released ICQ Lite. It still has skins. Now they all use MSN, because the client doesn't attempt to manage their life for them.
WinAmp 3 is not replacing WinAmp 2 at all - nobody I've met who's tried it thinks it's better than WinAmp 2.
So they were pretty bad examples.
If you're going to go on a flamewar, please do some research. Go to Desktop Preferences, Keyboard Shortcuts and then choose "Emacs" from the text editing shortcuts dropdown box. There you go, emacs keybindings.
Another good example is the "too many clocks" problem.
That was a brilliant example of all that was wrong with GNOME. Who on Gods green earth needs 4 different type of clock?
So I was pissed off when I upgraded my version of gnome and half the applets I used had gone!
What, just like when you upgraded from KDE1 to KDE2, virtually all your programs had to be ported or rewritten? If you really care about those applets, port them!
Don't even get me started on window managers with maximise buttons!
Huh? What is wrong with a maximise button? Every UI I've ever seen has one of these, and no the Mac way of expanding to some apparently random size is not good, it's a pain in the ass.
Developers should remember who they are developing for, and give more precedence to unix traditions than to windows traditions.
Yes, and the GNOME developers are developing for non-geeks. If you have a problem with that, go fork the old gnome1.4 codebase and continue loading it up with more clocks or something.
You're forgetting a few things, like C is far, far easier to use than C++. I must have learnt ten languages in my lifetime, most of them object oriented, but I still don't know C++. It scares me. I can normally read a language I don't know and figure out what's going on with a bit of effort. C++ eludes me. It's a mess. It appears to have several ways to typecast something. Multiple inheritance and generics makes any non-trivial codebase hard to understand. GLib C is a little easier, at least the base rules are more easily understood, but that's not so great either.
The good thing about GLib C is that it's very easily bound to other languages. The Python GNOME bindings for instance are a delight, and the GTKmm/GNOMEmm bindings are more C++ish than the KDE code is. For instance, you don't need a preprocessor to do signals, and you use the STL a lot more.
I know there are Python bindings for KDE. I've yet to find things that use them. On the other hand, for PyGnome there's Straw, the Redhat tools, for Perl::GTK there's the Mandrake setup tools and a photo gallery program I was working on a few months ago.... etc etc.
The language and framework baggage that the GNOME developers are saddled with make them work twice as hard to achieve the same thing.
GNOME has language baggage? You clearly haven't seen the ridiculous flamewars on kde-multimedia. Presented with a feature-rich, powerful multimedia framework that is light-years ahead of anything else (GStreamer) some of the KDE developers were objecting to it because despite the presence of KDE bindings to use it, you might need to write a GObject in C to extend it.
In fact with more complete bindings you wouldn't even need to do that. Nonetheless, it appeared to be their opinion that nobody could possibly want to use something based on the GObject object model, and made some very condescending remarks about it. It was even suggested that a new framework based on C++ was started, an utterly stupid idea, considering GStreamer hardly needs extending at the moment, and that it'd take 2+ years to get to where GStreamer is today. Nonetheless, that's how tied to C++ those guys were.
Unfortunately, KDE is still highly C++ centric. Perhaps this is no longer the case, but when I last looked KParts was simply a wrapper around the c++ classloader, you couldn't create a KPart in anything other than C++.
Thus, less time available to make things polished, bug-free, etc. Maybe this is why the GNOME leadership is pushing the super-stripped newbie desktop direction.
Instead of wildly speculating, why not go review the flamewars of the GNOME2 development era and you can find out why they simplified things right now. It's pretty simple - the majority of the developers felt there was a need for a clean desktop that had usability designed into the core. The fact that it's written in C had nothing to do with it.
They have one already though. Basically if you change the format of the keys (which only happens rarely), you have to migrate data across when you start, leaving the previous keys intact. Upgrades should in theory be painless, if they aren't it's a bug in the app.
Which means deletion of .gconf and .gnome* so defaults can be restored and things can run again.
Or you could just go and delete the individual config files, using the good old "cd" and "rm".
Why is it a ridiculous attitude? Tell you what, you go away and design this killer GUI, and come back when you've got it all figured out. If it's good, I'm sure somebody will be enthusiastic enough to help you.
Besides, if you want a graphic designer, I'm your man. Linux apps are always sorely lacking in the aesthetics department.
Sorry, but while you were whinging on Slashdot tigert, jimmac, everaldo and tackat were busy making GNOME and KDE pretty. By all means go and help them, but I find the latest GTK2/GNOME2 apps much better looking than their equivalents on Windows or the Mac.
If you're seeing that, it's usually because either you have bad drivers, or because the game was originally developed for Direct3D and then an OpenGL renderer was added later. The OpenGL renderer usually isn't as well optimized in such games, so you get lower framerates.
In games that use OpenGL natively (ie have been well optimized) you normally get higher framerates than on Windows.
Agreed. He also seems to confuse "lots of stuff" with "features". Having 3 media players or whatever they're up to now is not feature rich, it's bloat. OK, OK, so I'm being picky, but I've yet to find features that I really miss from KDE. Sure, I liked the idea of having a mini-webserver in my panel, but I never actually used it.
And Gtk+ and Qt both make very inefficient use of the X11 APIs, giving X11 an undeserved reputation for being slow.
Ignoring the fact that both GTK and Qt have pluggable backends, which entails a very slight loss of tight integration, the real reasons that GTK2 is slower than GTK1.2 are
1) It is double buffered. This slows GTK down, but makes it feel faster and smoother to the user. Net win.
2) It uses anti-aliasing. XRender is still not finished, nor well optimized. Profiling the new gnome-terminal widget for instance showed that most of the time was spent inside the X server. That will speed up over time.
3) If you do opaque resizes of apps, you'll find the content area doesn't stick to the borders. That's due to a problem with the internal scheduler, not due to lack of speed on the part of GTK.
I suppose because they're used to postal mail, which being routed by humans has a certain amount of fault-tolerance in addressing.
I mean, if you're in the UK, do you always provide the post-code? Thought not.
OK, that was an extreme example. Demurrage is where you have negative interest - savers are penalised by having some of their savings creamed off each month.
Obviously in any economic system there will be things you can't do, just like there are with capitalism (ie no abusive monopolies etc)
Before that would happen, Windows would have to be:
a) Free software and
b) No longer controlled by Microsoft.
That simply isn't going to happen, ever. Microsoft have no incentive to let go of Windows, and until that happens Linux will be as important as it always was, not because it's more stable or tweakable or whatever, but because it's owned by everybody.
Well yes, obviously most suggestions are bunk. I didn't notice any stealing or prison sentances in my little suggestion before ... obviously you have to separate the wheat from the chaff .
Here's a freedom-enhanced idea for everyone to get the new antibiotic for however cheap you can imagine: buy all the shares of the company at the market price and then give the drug away once you own it.
That might work once, but afterwards? It doesn't actually _change_ anything, it simply gets around it once, and you'll probably destroy the company at the same time.
See how everyone got the antibiotic they needed and no one's freedom was curtailed?
There are plenty of alternative systems that don't involve curtailing freedom - the idea of demurrage for instance.
A lot of peoples knee-jerk reactions seem to be "this guy is a communist!". What a pile of toss. I never even implied communism or socialism in the parent post, all I said was perhaps we should be trying to think of alternative systems that reward innovation but not by restricting its usage.
Come on people, use your imagination! I know none of us are economists, so what. OK, here's a random brainfart:
What if we made saving money illegal? Let's say that at the end of every month, any personal profit you had made you were forced to spend. It doesn't matter how you spend it, as long as you do. Clearly, everybody would have an average personal wealth of zero - if I earn £850/month and spend £600 on living, having fun etc that leaves me with £250. If I'm now compelled to spend that on something, I have the following choices:
a) Buy things with it. Nice food, clothes, games, computer upgrades, whatever. But, at the end of the day even the most material of us can't buy things forever - you'd simply run out of room. Obviously I wouldn't be able to buy anything with a price of over £250 unless I borrowed money off of other people...... which leads me onto
b) Give it to other people. Perhaps my friend rob wants to buy a new stereo system costing £500, he has £250 left over, so do I, so I give my surplus to him. After all, it's either that or I lose it. Maybe I give it to him on the condition that he repays me, but in practice that'd be quite hard to co-ordinate as if one month he dumped £250 on me, I'd now have £500 to get rid of, instead of only £250.
c) Invest it. This is the obvious thing to do. If somebody has a scheme and they want capital for it, they can go around and get surplus money off of other people, with a certain return. You'll have more money to get rid of next month, but hey, your lifestyle costs will probably go up to match it.
d) Buy insurance. Put some away for a rainy day right? Well, I can't save any more, but I can buy some insurance in case I lose my job, or suddenly I have to pay for something expensive. Insurance is clearly going to be a slop bucket of wealth - balancing the books so that in any one month enough wealth comes in to cover the usual payouts would be fun. Maybe governments or independant organisations would have a license to stockpile capital in case of huge unexpected payouts like in natural disasters or something.
e) Give it to society. I dunno what to do now. Another month, another paycheque. I don't feel like giving my money to a friend, nor investing it. Instead, why not make a proportion of my income a contribution to society? It could be spent on research, on traditional charity, on funding the education system.... whatever.
Companies obviously have to fit into this system. If they can't accumulate capital, how do they grow? Well, by borrowing from those with surplus, and repaying people in shares? Any month they have underspent, the money goes on dividends. Giving the director a nice fat $500,000 bonus isn't much use, nobody can spend that much in a month, not even on investments.
So, somebody disovers a new anti-biotic in some slime, and they want to turn it into a cure for a disease. Why? Altruism? Fame? The fear they or a loved one might get it someday? Why do people write free software? Who knows. Anyway, the investment needed to turn it into a working drug is huge. They go around, finding people who routinely make surplus capital, and ask for a proportion of it. After all, use it or lose it. Individuals who care for society (they do exist!) would donate, and their funding grows, and when the research is completed the drug is released to the world.
Well, it's almost certainly an incredibly stupid idea. It's probably riddled with holes, I can think of a few problems with it now. It might have been tried before. It'd probably be ridiculously unstable. The idea is 99% guaranteed to be junk.
BUT that doesn't matter! Don't focus on the idea. Ignore the idea. I don't want to see replies flaming me for what a stupid idea it is, I know that already. Focus on the meta-idea: we should be dreaming up alternatives to the current system.
Come on, give me some credit, I'm an incredibly poor economist but at least I'm trying. The current system IS a big hack, it DOES have big problems. The solution isn't to deride people who point this out by making smart-alec comments about socialism, the solution is to sit down and try and create something better. It doesn't have to be original or massively inventive, it just has to be better than what we currently have. Previous attempts failed - so what? BeOS and OS/2 died, yet we have Linux.
If so many people can sit down and bend the rules of the software economy by creating a complete OS from scratch, I think there is enough intelligence and willpower at least here to do the same for economics. It's worth a shot.
ps - I would never normally say this, but if you have some spare mod points please mod this up, I think it's very very important that people lose this lethargy over thinking about economics and society. I know we're mostly hackers, but all I see here is bitching about how evil the RIAA is, or software patents, or even microsoft when really they are simply products of the system. I want people to see this post and sit back and think about it, and maybe out of the discussion something cool will appear.