I only mentioned motif because it's what the gimp was originally written in. Personally I would go for qt every time, makes programming a joy, it really does. Just try writing the same program in gtk+ and qt.
The IDE is the equivalent of a compiler in this respect though. It's compiling the higher level "IDE-lang" into c++ or whatever you're using - which is what we did when c++ was invented and you wanted to use objects - you'd compile the c++ into plain c, which you could have written yourself, but each compiler would do it slightly differently.
I use qt designer to do my guis a lot, and save them as.ui files. I can have my ide turn these files into.py or.cpp files, or alternatively I can load them directly at runtime. Is loading them at runtime somehow "better"? Because it's pretty much equivalent to using a library to expand the language, but the only difference between it and ide code generation is that it's slower.
I will say one thing for your side though. What we need is a standard way of handling code generation, so you can move your code generator templates between IDEs. Otherwise you have lock-in, which is never a good thing.
My point is that if the linux kernel maintainers aren't going to release a stable kernel, and alan or andrew said they wouldn't, it was down to distro makers to make sure their kernels are stable, then I don't have that choice any more, at least if I want a stable system. I can program reasonably well, but not well enough to maintain a kernel, and my system *must* be stable. So now I'm forced to rely on distros for my kernels, not ones that I've compiled with the source from kernel.org, which is a shame. It's no longer possible to build everything yourself unless you can maintain your own kernel.
Even with current 2.4? What about current 2.2? If there is a bug, report it. They do get fixed, and when nothing else is happening to the code, the number of bugs tends to zero.
One advantage is that there are more open source fans reading the source than litigious companies, so hopefully we will spot any infringement before they do.
Increasing abstraction is necessarily the way of things. When assembly language was invented, some said it was a crutch for people who couldn't write machine code properly. Ditto for fortran and assembly, ditto for garbage collection, ditto for any feature that makes coding easier.
The next wave is the ides that make the repetitive coding unnecessary. Sure, it makes things slower, but it makes *developing* them faster, and machines are getting faster but developers aren't. I don't think there's anything to be afraid of.
The key word there is "some". A kernel which is stable on some systems and unstable on others is not yet a stable kernel. And I can assure you that 2.6 is unstable on my system.
Am I the only one that thinks that GTK+ is just plain terrible outside of X?
Yes, but you don't even need the second part. GTK+ is horrible to program with. It's a kludge that has outlived its purpose, with openmotif and gpled qt available there really is no reason for it to exist. Now I know a lot of people have done a lot of work on it, and there is lots of legacy code around, but it's really time to move on. Gtk is just fundamentally hard to use.
You can call me a troll if you like, but as a developer who has tried several toolkits, none of them was harder to use than GTK+. Like aRts, it did its job well and was good as a stopgap, but now it's time to move on.
This is really like saying that in music, CD quality recording and mastering is good enough, since virtually no-one can hear the difference between CD and mastertape.
Yes, exactly, and that's *true*. The only reason you can ever hear the difference is the crappy mastering done because people want it to sound as loud as possible. So say bye-bye to your dynamic range, it's everything at 0db for you.
Try it some time. Get a cd of two sounds which differ by the smallest amount you can get them to on the cd - just use your favourite audio editor in 16 bit 44100hz mode, and make one sound the tiniest bit louder than the other, or the tiniest bit higher frequency. Then put them together continuously with each a random length - or rather, get a friend to do so, so you don't know where they are. Then listen to this. As hard as you like. And try and spot the transition. You won't be able to. No-one can. That's why those specs were chosen for the cd. It wasn't a question of someone saying "hmm, 44100hz sounds like enough". They tested it, and made the cd high enough quality that no one could tell the difference. That's the best record company "golden ears" they had. If you think you can tell the difference you are deluding yourself, plain and simple.
Why are you all bitching about the interface? It works perfectly fine for me on linux with fluxbox; I just toss all the windows onto an empty virtual desktop.
Good for you. However, it should also be configurable to work fine for the rest of us. Yes, we may want to use bad window managers. The program should be able to handle it gracefully.
I'm not arguing for removing the current interface. I'm not even arguing for making it non-default. But I want an option to change it to the kind of interface I like, not just the one the developers prefer.
And yes, I am willing to code it myself if that's what it takes. But I can't keep up with the development of the rest of the gimp doing it alone. There are enough people who want an alternative UI that the gimp devs should, if nothing else, start a project where people interested in implementing it can sign up and organise things.
No, please don't do that. A name is just a name, and I prefer a more playful title than "image editor". I've never met anyone who's actively offended by it, even my mum. At the moment linux should still be about appealing to people who are interested in computers, and they tend to prefer a program with an interesting name. I don't want to go all corporate and mainstream quite yet.
So, presumably you are not using the crappy window managers that come with KDE/Gnome, what is it really that makes PS stand out?
No, the poster is using the crappy window manager that comes with windows, and has no choice about that, so it would be nice if their image editing program could handle it gracefully. Yes, it's the wm's fault. However, photoshop works around the wm's flaws, which is why its ui is better. And, to be useable on windows, the gimp is going to have to as well.
The reality is more like
When there's a 2.6 that is actually useable on my system (crashes less than once every 30 mins) I'll believe that. 2.7 is not ready to split off until then.
Well, I actually think it's "2.5 should not have become 2.6 until it didn't crash so frequently on my system". But I can understand that 2.6.0 had bugs, it was after all the first stable release, and so had a much wider audience than 2.5 did. The problem is that they've kept adding new features to it, and, like it or not, new code has bugs. What they need to do is put the cool new things in 2.7, and just concentrate on squashing bugs in 2.6. After about 3 releases of pure bugfix, no more features, the kernel tends to become stable. But as long as they're constantly adding new features to 2.6, it will be unstable.
If it's the default, then possibly, although I tried repeatedly with different configurations. I definitely used the version of the nvidia drivers which supposedly had a fix for that, though, so I don't think that's the problem.
Err, what you need is a distro. Essentially every distro is a little fork. And the emphasize different things.
Why should I need a distro? I remember when Linux was used mostly by hobbyists, and you downloaded the sources for the programs you wanted and compiled them yourself. Heck, there weren't even distros around until about '93. And you expected that to work as well or better than getting a prepackaged version, because that was probably what the person developing the program had done.
I have nothing against distros, but it's sad if doing things your own way is no longer possible.
Have you ever tried to maintain code with the amount of ifdefs that linux may need to use? A code-folding ide helps somewhat, but it's still horrible. When more than about 1/3 of the source is ifdefed, it's time to fork.
I have. Not all my applications, but I do have to recompile svgalib and sometimes some applications which use it. And don't get me started on the cdrtools shennanigans.
The development process is going so smoothly that there may not need to be a 2.7 at all; for the first time in linux development history the developers are able to make useful improvements to linux while keeping it somewhat stable.
I call bullshit on this. When there's a 2.6 that is actually useable on my system (crashes less than once every 30 mins) I'll believe that. But I don't think it's going to happen until 2.7 splits off
I don't think that's what it is. I think it's saying that they'll fork off versions for really big patchsets they don't want to merge into the main kernel. So there may be 2.6-hard-realtime, 2.6-reiserfs-v4, 2.6-ip-v9, etc. trees. Which is something to be concerned about, although not necessarily a bad thing.
And they are not covered by the Geneva convention since they are not following the Geneva convention when they fight. To qualify for the protections offered you have to follow the guidelines set.
However, you are entitled to the protections of the convention until this has been determined by a competent tribunal (see article 5). Which so far none of the people at Guantanamo have got.
As has been made clear elsewhere, no it wouldn't. It just means you have to be able to install LSB-compliant rpms - i.e. you need rpm installed and you have to persuade it that you have a few core system libraries installed, which although slightly tricky without making a fake rpm, is not impossible.
As I posted above, those of us who like stability are not yet on 2.6. 2.6 is still a development kernel, at least as much as 2.5 was, and if you weren't willing to run 2.5 you shouldn't run 2.6.
I am still running 2.4, partly because I need the connexant modem drivers and want the free, full speed version, but mostly for one simple reason: 2.6 is not stable.
2.6 changes far too much for a stable kernel. Every version I've tried crashes when switching out of X and then back on my cheap-as-hell TNT2. That I could accept - it's new, after all. What I can't accept is the attitude of the kernel maintainers. They inserted cryptoloop and then took it out in the space of about 5 minor versions. They ripped out the perfectly good cd writing code and replaced it with a buggy, undertested interface, for no other reason as far as I can tell than that Linus doesn't like it. But the last straw for me was Andrew saying that it's up to the distributions to make sure their kernels are stable. Is Pat Volkerding expected to stabilise the zillions of lines of code in the kernel all by himself? Because that's what it seems to be implying. But, more importantly, the linux people are distributing as a stable piece of software, something which they admit is not suitable for end users.
If this goes on, we need a fork. I don't say this lightly, but the maintainers of linux seem to have abandoned the hobbyists it was started for for the sake of the big business redhat/novell distros. And that's not a kernel I want to be relying on. Anyone with me? Or should I just go over to one of the BSDs?
It can be, depending entirely on your setup. Obviously there's no way you can declare gentoo to be compliant, because someone could set up their system to have man pages be installed in/opt/foobar and that would break things. But if you do things normally, and emerge rpm, lots of gentoo systems are LSB-compliant.
Well, ifconfig is the standard:)
Seriously, isn't netconfig available on every distribution? And since servers need network configuration more than desktops do, I can't imagine any gui tool becoming standard. Just learn netconfig, it's actually just as easy as a gui tool, just not quite as pretty looking.
Slashdot Mirror
I only mentioned motif because it's what the gimp was originally written in. Personally I would go for qt every time, makes programming a joy, it really does. Just try writing the same program in gtk+ and qt.
I use qt designer to do my guis a lot, and save them as .ui files. I can have my ide turn these files into .py or .cpp files, or alternatively I can load them directly at runtime. Is loading them at runtime somehow "better"? Because it's pretty much equivalent to using a library to expand the language, but the only difference between it and ide code generation is that it's slower.
I will say one thing for your side though. What we need is a standard way of handling code generation, so you can move your code generator templates between IDEs. Otherwise you have lock-in, which is never a good thing.
My point is that if the linux kernel maintainers aren't going to release a stable kernel, and alan or andrew said they wouldn't, it was down to distro makers to make sure their kernels are stable, then I don't have that choice any more, at least if I want a stable system. I can program reasonably well, but not well enough to maintain a kernel, and my system *must* be stable. So now I'm forced to rely on distros for my kernels, not ones that I've compiled with the source from kernel.org, which is a shame. It's no longer possible to build everything yourself unless you can maintain your own kernel.
Even with current 2.4? What about current 2.2?
If there is a bug, report it. They do get fixed, and when nothing else is happening to the code, the number of bugs tends to zero.
One advantage is that there are more open source fans reading the source than litigious companies, so hopefully we will spot any infringement before they do.
The next wave is the ides that make the repetitive coding unnecessary. Sure, it makes things slower, but it makes *developing* them faster, and machines are getting faster but developers aren't. I don't think there's anything to be afraid of.
The key word there is "some". A kernel which is stable on some systems and unstable on others is not yet a stable kernel. And I can assure you that 2.6 is unstable on my system.
Yes, but you don't even need the second part. GTK+ is horrible to program with. It's a kludge that has outlived its purpose, with openmotif and gpled qt available there really is no reason for it to exist. Now I know a lot of people have done a lot of work on it, and there is lots of legacy code around, but it's really time to move on. Gtk is just fundamentally hard to use.
You can call me a troll if you like, but as a developer who has tried several toolkits, none of them was harder to use than GTK+. Like aRts, it did its job well and was good as a stopgap, but now it's time to move on.
Yes, exactly, and that's *true*. The only reason you can ever hear the difference is the crappy mastering done because people want it to sound as loud as possible. So say bye-bye to your dynamic range, it's everything at 0db for you.
Try it some time. Get a cd of two sounds which differ by the smallest amount you can get them to on the cd - just use your favourite audio editor in 16 bit 44100hz mode, and make one sound the tiniest bit louder than the other, or the tiniest bit higher frequency. Then put them together continuously with each a random length - or rather, get a friend to do so, so you don't know where they are. Then listen to this. As hard as you like. And try and spot the transition. You won't be able to. No-one can. That's why those specs were chosen for the cd. It wasn't a question of someone saying "hmm, 44100hz sounds like enough". They tested it, and made the cd high enough quality that no one could tell the difference. That's the best record company "golden ears" they had. If you think you can tell the difference you are deluding yourself, plain and simple.
Good for you. However, it should also be configurable to work fine for the rest of us. Yes, we may want to use bad window managers. The program should be able to handle it gracefully.
I'm not arguing for removing the current interface. I'm not even arguing for making it non-default. But I want an option to change it to the kind of interface I like, not just the one the developers prefer.
And yes, I am willing to code it myself if that's what it takes. But I can't keep up with the development of the rest of the gimp doing it alone. There are enough people who want an alternative UI that the gimp devs should, if nothing else, start a project where people interested in implementing it can sign up and organise things.
Oh, and give me back my pimp ass newsreader!
No, the poster is using the crappy window manager that comes with windows, and has no choice about that, so it would be nice if their image editing program could handle it gracefully. Yes, it's the wm's fault. However, photoshop works around the wm's flaws, which is why its ui is better. And, to be useable on windows, the gimp is going to have to as well.
Well, I actually think it's "2.5 should not have become 2.6 until it didn't crash so frequently on my system". But I can understand that 2.6.0 had bugs, it was after all the first stable release, and so had a much wider audience than 2.5 did. The problem is that they've kept adding new features to it, and, like it or not, new code has bugs. What they need to do is put the cool new things in 2.7, and just concentrate on squashing bugs in 2.6. After about 3 releases of pure bugfix, no more features, the kernel tends to become stable. But as long as they're constantly adding new features to 2.6, it will be unstable.
If it's the default, then possibly, although I tried repeatedly with different configurations. I definitely used the version of the nvidia drivers which supposedly had a fix for that, though, so I don't think that's the problem.
Why should I need a distro? I remember when Linux was used mostly by hobbyists, and you downloaded the sources for the programs you wanted and compiled them yourself. Heck, there weren't even distros around until about '93. And you expected that to work as well or better than getting a prepackaged version, because that was probably what the person developing the program had done.
I have nothing against distros, but it's sad if doing things your own way is no longer possible.
Have you ever tried to maintain code with the amount of ifdefs that linux may need to use? A code-folding ide helps somewhat, but it's still horrible. When more than about 1/3 of the source is ifdefed, it's time to fork.
I have. Not all my applications, but I do have to recompile svgalib and sometimes some applications which use it. And don't get me started on the cdrtools shennanigans.
I call bullshit on this. When there's a 2.6 that is actually useable on my system (crashes less than once every 30 mins) I'll believe that. But I don't think it's going to happen until 2.7 splits off
I don't think that's what it is. I think it's saying that they'll fork off versions for really big patchsets they don't want to merge into the main kernel. So there may be 2.6-hard-realtime, 2.6-reiserfs-v4, 2.6-ip-v9, etc. trees. Which is something to be concerned about, although not necessarily a bad thing.
However, you are entitled to the protections of the convention until this has been determined by a competent tribunal (see article 5). Which so far none of the people at Guantanamo have got.
As has been made clear elsewhere, no it wouldn't. It just means you have to be able to install LSB-compliant rpms - i.e. you need rpm installed and you have to persuade it that you have a few core system libraries installed, which although slightly tricky without making a fake rpm, is not impossible.
As I posted above, those of us who like stability are not yet on 2.6. 2.6 is still a development kernel, at least as much as 2.5 was, and if you weren't willing to run 2.5 you shouldn't run 2.6.
2.6 changes far too much for a stable kernel. Every version I've tried crashes when switching out of X and then back on my cheap-as-hell TNT2. That I could accept - it's new, after all. What I can't accept is the attitude of the kernel maintainers. They inserted cryptoloop and then took it out in the space of about 5 minor versions. They ripped out the perfectly good cd writing code and replaced it with a buggy, undertested interface, for no other reason as far as I can tell than that Linus doesn't like it. But the last straw for me was Andrew saying that it's up to the distributions to make sure their kernels are stable. Is Pat Volkerding expected to stabilise the zillions of lines of code in the kernel all by himself? Because that's what it seems to be implying. But, more importantly, the linux people are distributing as a stable piece of software, something which they admit is not suitable for end users.
If this goes on, we need a fork. I don't say this lightly, but the maintainers of linux seem to have abandoned the hobbyists it was started for for the sake of the big business redhat/novell distros. And that's not a kernel I want to be relying on. Anyone with me? Or should I just go over to one of the BSDs?
It can be, depending entirely on your setup. Obviously there's no way you can declare gentoo to be compliant, because someone could set up their system to have man pages be installed in /opt/foobar and that would break things. But if you do things normally, and emerge rpm, lots of gentoo systems are LSB-compliant.
Well, ifconfig is the standard :)
Seriously, isn't netconfig available on every distribution? And since servers need network configuration more than desktops do, I can't imagine any gui tool becoming standard. Just learn netconfig, it's actually just as easy as a gui tool, just not quite as pretty looking.