Do you know the meaning of that OR at the start of the first phrase? anyways, you sounded like having to recompile your distro kernel to add support to your printer, you didn't specified that you where using a custom kernel, and soo I can think anything.
Nope, I didn't directly accused of laying, I pointed out posibilities behind you comment, which didn't sounded quite right to me, do not think is something against you, but i'm also tired of tons of people making up stuff just because some OSS didn't worked as they expected, I get comments like "X didn't worked when I tried Y, is shit broken fix it!" i just reply "Y with X is not quite the right thing to do, try Z instead Y" finally it works, but people often jump into attack mode because they don't RTM.
So yes, I'm suspicious once in a while of comments like yours that are not totally clear to me. Given that we are even.
That's what happen when you rely on Close applications which are "free", if Linus is so convinced of BitKeeper then buy a damm license like everyone does with close software.
If he can't buy licenses because his "friend" revoked all kind of licenses, then he's not his friend and Linus is being dumb by defending a person as McVoy.
If the license is too expensive, then Linus shouln't have used it to begin with, should have thinked of that case too, but he didn't.
No point on keeping arging the BitKeeper case, lets just move on.
syousef: "I'm getting goddamn sick of this, and I'm a developer. I'm also damn tired of defending it. I've had comp sci students roll their eyes at me when I had to recompile my kernel to add support for a printer so we could print data off in Linux."
Or you are making that up or you really have problems with that printer.
I think the first because printer support is not something that comes with the kernel, you don't need to recompile the kernel just to add support for a printer, thats done by CUPS and it's drivers which dosen't have anything to do with the kernel.
But maybe you had problems with USB which got away from compiling latest kernel, or parallel port support was not compiled in your distro kernel. And no, firmaware uploading to printer has nothing to do with firmaware kernel driver, CUPS can send the firmaware by USB or parallel port.
C# is some sort of a nice language, but i prefer C++ mind you.
One thing nobody has ever got right is the development tools, MS development tools are the best out there for graphical applications.
You can preach all you want that we have Gtk, Qt, Mono, etc. But of all those only Qt has a sort of IDE that integrates GUI creation with code in a point and click fashion.
Many will flame that if we are able to do apps ala point and click we'll get many sucky apps like windows ones, yeah sure, but is really a life saver when you don't have the time to hard code all GUI or use Glade to load a pregenerated GUI and then tweak it at your needs.
We need a damm IDE that can compete with MS Visual Studio, don't talk about monodevelop it dosen't get near to even SharpDevelop, Eclipse is not quite there, Anjuta 2 with Glade 3 might be a good candidate but is still on alpha stages.
All we need is better GUI tools, not new languages!
I'm never gonna buy another lexmark again!, first, they Linux support is reather bad, but the ink cardbridge price is totally insane.
I bought a cheapo Lexmark Z605 a few months back, which was around 50 that day, it included 2 inks (one b/w other color). The cardbridges didn't lasted, I had to buy another set, but then I saw the prices, heck! 30 for b/w and 35 for color, OMG!, that was more than the printer itself!
I buyed only a b/w cartbridge, after it was empty I went for a recharge from a provider near my house, wich will fill it for around $15, but the cardbridge was in bad state and reather tricky to refill, they suggested me to buy another printer, as it was only round $30 for the same printer model.
WTF are they thinking? do I need to trash my printer and buy a new one just to get those damm cardbridges?
Yes NetBSD 2.0 uses XFree86 4.4, which is a little broken I may say. It lacks Shadow module, this is broken and won't compile, and thus is not on the package. Vesa driver and others (trident) won't work without turning shadowfb off, which is not too much of a hasle, but on my trident card (from a old laptop, soo NO upgrades!) without shadowfb it won't work without disabling XAA Acceleration, which makes it really slow, almost unbearable.
This dosen't happen with XFree86 4.3 or any X.org version, where turning ShadowFB on enables XAA acceleration just fine.
Actually the main reason for buying my current music player (to not say mp3 player) is that it can play ogg files along with mp3s.
And is not some sort of unkown or unsuported player, is an iRiver, which i found quite sleek. As well is supported on Linux too (but needs extra software, open source of course).
Here at home we still use those Rotary-Dial Phones, they are here since the house was wired with phone lines and we haven't replace them, at first we did test a button phone, but the buttons got stuck fairly fast, thus, we never used that kind of phone again, and till today that's the main excuse to no "upgrade", altough with new phones this shouldn't be a problem anymore.
Still, i would like to change at least one of the two phones in my home, just to call cellphones numbers, which are around 12 digits!, quite a pain.
In reality, what you really own is the DVD, the box and all the physical stuff on it; now, the media on it is us just licensed.
You can do whatever you want with your DVD (play it, use it as a frisby, etc.) however there are limits on what you can do with the media (cannot copy that data on it, cannot play it for money, cannot play it on public, etc, etc).
Is not like you own the movie or something, just the physical DVD disk.
While is true the there is "free market" in that you have options and even free options, Microsoft has the market really monopolised, why?
Well is not common to buy a Computer without Windows, it is bundled on almost all new computers, even if you put another operating system after buying it you already have payed the M$ Tax.
Microsoft has all the marked for them, why? because people don't know anything else, M$ monopoly on personal computers is so big that people don't know anything about computers, they think that windows is part of the computer. Their Mindshare (gained trough monopolistic tactics) is so big that is almost imposible to go against (you cannot sucessfully sell anything else).
Thus, there isn't really a free market, that's only an illusion, period.
That is what's goverment for!, to keep an eye on companies trying to monopolise the market, to keep the market free. But obviusly your goverment is dirty and corrupt and don't give a damm about his people (altough i shouldn't blame them, they are reciving tons of bucks from companies, who would reject something like that?)
Anyways, you got what you want, people don't care, fine, the monopoly will be there always.
Linux ACPI Support is actually standards compliant, as written by the specs, even intel writted it!, the problem are manufacturers which sell laptops with broken acpi, and windows that has a broken acpi support that only works with broken implementations (thus everyone has to write broken code to make it run on windows), tough windows code suposly detects broken bios and enables workarounds, this will only promote the broken bios that companies put on your laptops.
Yes, GConf and Windows Registry aren't the same at low level.
But when i want to change an advanced option on say GEdit, i need to go and open GConf, search the right key and change it. It really feels like hacking the windows registry to get things done.
Is not about the underlaying technology, is about the user feeling, isn't gnome philosopy that simple is better? that less is more?, then why do they what me to open another program just to configure the program i'm currently using?, this just causes more confusion.
A better idea should be to embed the gconf editor (using only the keys of the program) into and advanced section.
Really, i don't think that opening a separate programs to change some key on something that seems like windows registry editor is any better than having 10 tabs filled with options.
It depends very much on who install, after all each distro is geared towards a specific user, and some are just as general as posible.
I have installed both Slackware and Mandrake systems, and Mandrake autodetects most of things and makes my life easy, however i use slackware on my machine because i know it very well and everyting is very tight, i know what to install and what not, and I compile the kernel with only the stuff i need, i know my hardware and have enough knowledge to do it.
It's all about needs, some may need lots of customability, other a Just Works (TM) type of system, etc.
This may be true at certain extent, as now is really easy to install software with urpmi or apt-get, however there are still many problems around linux.
The fractmentation of package managers, many distributions use RPM, however there is Debs, and even plain Slackware packages (which are mostly just a gzipped tarball).
There is no standard packaging procedure, for developers is just easier to give the source code instead of compiled binaries, or just give static linked ones.
Some developers say: "Is the job of the distributor to compile and package the software".
Prepackaged binaries should be like: download, double click, maybe follow some on screen installation instructions. bam done, all installed. This is what windows users are familiar with.
Don't known why this remind me so much of the Kernel... o yeah, you are talking about the kernel developers aren't you?
Well now being serious, the Linux kernel is a clear example of this, when i install the NVidia drivers it works ok, but i recompile the kernel, even the same version of the kernel and it will broke all drivers... I have to recompile the NV drivers again... WTF? i just added support for some other Filesystem on the kernel!!!, is this enough to broke binary drivers!?
I Agree, Objective C is easy to learn and very flexible. However the downside of it is that there us no Graphical Toolkit for it besides GNUstep, the GTK Objective C bindings are soo old! (from Gtk 1.0).
I was working on a library for creating database oriented applications, i used Objective-C all the way, added transparent access to various databases (protocols are great), however the problem was the GUI, i tried to make a more current wrapper for Gtk, however Gtk2 is horrible complex!!!!
Why the hell do i need to get the GtkTextBuffer from the GtkTextView and later use 2 GtkTextIter to extract some text from the GtkTextBuffer. All this because that little multiline text widget...
The same thing happends wide spread in all Gtk2, you need to create a ton of objects from objects to do anything. As previously mentioned somewhere, is just the API bloat.
Now back on topic, i think they should enfocate more on providing bindings for all sort of languajes so everyone can use Gtk/Gnome instead of focusing on one languaje that *has* the majority of developers out there.
The problem for us developers is not really the directory layout, for this there is LSB. Tough some people use the/opt dir for things like kde or OO.org (like slackware does with kde).
The problem comes with the libraries, i can compile my program on one system and it works, but in other systems it may fail because of the ABI incompatibility of some libraries you need (an example of this is glibc, just try to recompile it and watch how your system stops working).
Thats why we distribute the source, binaries are just painful to mantain. And no, i don't like statically linked stuff either.
I think you are mistaking OSS Developers for some sort of enterprise company (some are, but not all of us).
Most of us start a project that is useful for us only, later we found that it can be useful for other people, then we make our software Open Source, for everybody to use, share and modify.
We are not about users, we are about sharing. You can modify it if you don't like it, it works for me. OSS people are very kind and care about their users, but their users don't help, they always keep yelling out loud of how the software isn't what they want, that why the interface is ugly, blah, blah, blah.
Stop this crap at once, you don't pay me for coding, i already giving it for free, if you don't like it you can look for other projects or buy some software that does what you want!. Don't bother me with nonsenses, the code is there, help!
They didn't take in count all the popular windows viruses out there, this is not quite objetive.
If you expose a Linux machine to the internet, is unlikly that it will get cracked, and this can be better if you keep up to date all important software and don't run stupid things as ROOT or have running unecesary services.
Now if you put a Windows Machine on the internet, is likly that it will get a virus or will crash. Tought if you keep your machine updated it can be safe for a while, until the new exploit gets out, or M$ locks your machine for ever.
The problem for the masses is not hackers (i don't think there is enough hackers to crack all the servers out there one by one), but viruses and other exploits out there, this is where windows is very vulnerable, anyone remember the RPC problem? it will shutdown your windows box and you don't even need to touch it.
It does make sense from the license point, this will help to avoid using GPL code in close code (statically linking, mixing code, etc).
However, this attitudes makes Linux hard to work with, I have to live with half working drivers (yeah it works, but not all the functionability is supported, this to me is incomplete), hardware manufacturers won't open all. This is not by any means helping the adoption of Linux (on the desktop).
I don't know if I should take Linux seriously any more, by any means. As now that i can't choose between closed and open drivers (yeah I know, I can, but creating close drivers is becoming more dificult, that's my point).
Anyway the idea of creating a driver layer between the driver (closed drivers) and the kernel with a stable API/ABI sound quite good now. We would only need to modify the layer source for new kernels and support the stable API/ABI.
Sure, we can test all the code that comes in with others code base, however may things araise from this:
* Implementations of Standars: what if 2 implementations of a Standard (Open standard) that have been maded separatly comes up with a very similar code or even the same code (counting out variable names and stuff like that), current US laws and IP rights are very extreme as you can patent almost any process, algorithms, and other stuff.
Anyways, we are a comunity that makes and uses that code, we don't want any malicius code running on our machines anyway. We are a community, not a bunch of terrorists, or an organization creating a conspirancy to over take the world by hooking everyone with our software and latter control them. (ala M$)
Making a already propetary software Open Source is not as simple as just opening the source to everyone, a lot of OpenSource projects started small, with people that know each other and they spread the word, latter, after many people get known of the project, there was already a comunity around it.
However this may not be the case of Java, lots of developers are interested on getting their hands on it, there is room for improvement (speed improvements:-)).
The netscape case was different, however look at mozilla today!
Slashdot Mirror
Do you know the meaning of that OR at the start of the first phrase? anyways, you sounded like having to recompile your distro kernel to add support to your printer, you didn't specified that you where using a custom kernel, and soo I can think anything.
Nope, I didn't directly accused of laying, I pointed out posibilities behind you comment, which didn't sounded quite right to me, do not think is something against you, but i'm also tired of tons of people making up stuff just because some OSS didn't worked as they expected, I get comments like "X didn't worked when I tried Y, is shit broken fix it!" i just reply "Y with X is not quite the right thing to do, try Z instead Y" finally it works, but people often jump into attack mode because they don't RTM.
So yes, I'm suspicious once in a while of comments like yours that are not totally clear to me. Given that we are even.
That's what happen when you rely on Close applications which are "free", if Linus is so convinced of BitKeeper then buy a damm license like everyone does with close software.
If he can't buy licenses because his "friend" revoked all kind of licenses, then he's not his friend and Linus is being dumb by defending a person as McVoy.
If the license is too expensive, then Linus shouln't have used it to begin with, should have thinked of that case too, but he didn't.
No point on keeping arging the BitKeeper case, lets just move on.
syousef:
"I'm getting goddamn sick of this, and I'm a developer. I'm also damn tired of defending it. I've had comp sci students roll their eyes at me when I had to recompile my kernel to add support for a printer so we could print data off in Linux."
Or you are making that up or you really have problems with that printer.
I think the first because printer support is not something that comes with the kernel, you don't need to recompile the kernel just to add support for a printer, thats done by CUPS and it's drivers which dosen't have anything to do with the kernel.
But maybe you had problems with USB which got away from compiling latest kernel, or parallel port support was not compiled in your distro kernel. And no, firmaware uploading to printer has nothing to do with firmaware kernel driver, CUPS can send the firmaware by USB or parallel port.
I don't see a problem to include Apache on a Linux distro, even if apache license is not GPL compatible.
The GPL license dosen't cover materials distributed on the same medium as GPL Software.
Altough it would be nice having Apache to be GPL conpatible, soo we can use GPL plugins with it.
C# is some sort of a nice language, but i prefer C++ mind you.
One thing nobody has ever got right is the development tools, MS development tools are the best out there for graphical applications.
You can preach all you want that we have Gtk, Qt, Mono, etc. But of all those only Qt has a sort of IDE that integrates GUI creation with code in a point and click fashion.
Many will flame that if we are able to do apps ala point and click we'll get many sucky apps like windows ones, yeah sure, but is really a life saver when you don't have the time to hard code all GUI or use Glade to load a pregenerated GUI and then tweak it at your needs.
We need a damm IDE that can compete with MS Visual Studio, don't talk about monodevelop it dosen't get near to even SharpDevelop, Eclipse is not quite there, Anjuta 2 with Glade 3 might be a good candidate but is still on alpha stages.
All we need is better GUI tools, not new languages!
I'm never gonna buy another lexmark again!, first, they Linux support is reather bad, but the ink cardbridge price is totally insane.
I bought a cheapo Lexmark Z605 a few months back, which was around 50 that day, it included 2 inks (one b/w other color). The cardbridges didn't lasted, I had to buy another set, but then I saw the prices, heck! 30 for b/w and 35 for color, OMG!, that was more than the printer itself!
I buyed only a b/w cartbridge, after it was empty I went for a recharge from a provider near my house, wich will fill it for around $15, but the cardbridge was in bad state and reather tricky to refill, they suggested me to buy another printer, as it was only round $30 for the same printer model.
WTF are they thinking? do I need to trash my printer and buy a new one just to get those damm cardbridges?
Enough of it, i'm gonna buy an HP or a Cannon.
Yes NetBSD 2.0 uses XFree86 4.4, which is a little broken I may say. It lacks Shadow module, this is broken and won't compile, and thus is not on the package. Vesa driver and others (trident) won't work without turning shadowfb off, which is not too much of a hasle, but on my trident card (from a old laptop, soo NO upgrades!) without shadowfb it won't work without disabling XAA Acceleration, which makes it really slow, almost unbearable.
This dosen't happen with XFree86 4.3 or any X.org version, where turning ShadowFB on enables XAA acceleration just fine.
NetBSD should move to X.org.
Just My 2 Cents.
Actually the main reason for buying my current music player (to not say mp3 player) is that it can play ogg files along with mp3s.
And is not some sort of unkown or unsuported player, is an iRiver, which i found quite sleek. As well is supported on Linux too (but needs extra software, open source of course).
Here at home we still use those Rotary-Dial Phones, they are here since the house was wired with phone lines and we haven't replace them, at first we did test a button phone, but the buttons got stuck fairly fast, thus, we never used that kind of phone again, and till today that's the main excuse to no "upgrade", altough with new phones this shouldn't be a problem anymore.
Still, i would like to change at least one of the two phones in my home, just to call cellphones numbers, which are around 12 digits!, quite a pain.
I'm glad they provide a bittorrent for it, their FTP server has been full for a while!
In reality, what you really own is the DVD, the box and all the physical stuff on it; now, the media on it is us just licensed.
You can do whatever you want with your DVD (play it, use it as a frisby, etc.) however there are limits on what you can do with the media (cannot copy that data on it, cannot play it for money, cannot play it on public, etc, etc).
Is not like you own the movie or something, just the physical DVD disk.
While is true the there is "free market" in that you have options and even free options, Microsoft has the market really monopolised, why?
Well is not common to buy a Computer without Windows, it is bundled on almost all new computers, even if you put another operating system after buying it you already have payed the M$ Tax.
Microsoft has all the marked for them, why? because people don't know anything else, M$ monopoly on personal computers is so big that people don't know anything about computers, they think that windows is part of the computer. Their Mindshare (gained trough monopolistic tactics) is so big that is almost imposible to go against (you cannot sucessfully sell anything else).
Thus, there isn't really a free market, that's only an illusion, period.
That is what's goverment for!, to keep an eye on companies trying to monopolise the market, to keep the market free. But obviusly your goverment is dirty and corrupt and don't give a damm about his people (altough i shouldn't blame them, they are reciving tons of bucks from companies, who would reject something like that?)
Anyways, you got what you want, people don't care, fine, the monopoly will be there always.
Linux ACPI Support is actually standards compliant, as written by the specs, even intel writted it!, the problem are manufacturers which sell laptops with broken acpi, and windows that has a broken acpi support that only works with broken implementations (thus everyone has to write broken code to make it run on windows), tough windows code suposly detects broken bios and enables workarounds, this will only promote the broken bios that companies put on your laptops.
Yeah, i read about this somewere.
Yes, GConf and Windows Registry aren't the same at low level.
But when i want to change an advanced option on say GEdit, i need to go and open GConf, search the right key and change it. It really feels like hacking the windows registry to get things done.
Is not about the underlaying technology, is about the user feeling, isn't gnome philosopy that simple is better? that less is more?, then why do they what me to open another program just to configure the program i'm currently using?, this just causes more confusion.
A better idea should be to embed the gconf editor (using only the keys of the program) into and advanced section.
Really, i don't think that opening a separate programs to change some key on something that seems like windows registry editor is any better than having 10 tabs filled with options.
why this "Standar across alll platforms" sound to me more like:
PC Windows, WinCE and XBox2...
I'm pretty sure they will only support themselves, maybe MacOSX too, but i don't think so.
It depends very much on who install, after all each distro is geared towards a specific user, and some are just as general as posible.
I have installed both Slackware and Mandrake systems, and Mandrake autodetects most of things and makes my life easy, however i use slackware on my machine because i know it very well and everyting is very tight, i know what to install and what not, and I compile the kernel with only the stuff i need, i know my hardware and have enough knowledge to do it.
It's all about needs, some may need lots of customability, other a Just Works (TM) type of system, etc.
This may be true at certain extent, as now is really easy to install software with urpmi or apt-get, however there are still many problems around linux.
The fractmentation of package managers, many distributions use RPM, however there is Debs, and even plain Slackware packages (which are mostly just a gzipped tarball).
There is no standard packaging procedure, for developers is just easier to give the source code instead of compiled binaries, or just give static linked ones.
Some developers say: "Is the job of the distributor to compile and package the software".
Prepackaged binaries should be like: download, double click, maybe follow some on screen installation instructions. bam done, all installed. This is what windows users are familiar with.
Don't known why this remind me so much of the Kernel... o yeah, you are talking about the kernel developers aren't you?
Well now being serious, the Linux kernel is a clear example of this, when i install the NVidia drivers it works ok, but i recompile the kernel, even the same version of the kernel and it will broke all drivers... I have to recompile the NV drivers again... WTF? i just added support for some other Filesystem on the kernel!!!, is this enough to broke binary drivers!?
I Agree, Objective C is easy to learn and very flexible. However the downside of it is that there us no Graphical Toolkit for it besides GNUstep, the GTK Objective C bindings are soo old! (from Gtk 1.0).
I was working on a library for creating database oriented applications, i used Objective-C all the way, added transparent access to various databases (protocols are great), however the problem was the GUI, i tried to make a more current wrapper for Gtk, however Gtk2 is horrible complex!!!!
Why the hell do i need to get the GtkTextBuffer from the GtkTextView and later use 2 GtkTextIter to extract some text from the GtkTextBuffer. All this because that little multiline text widget...
The same thing happends wide spread in all Gtk2, you need to create a ton of objects from objects to do anything. As previously mentioned somewhere, is just the API bloat.
Now back on topic, i think they should enfocate more on providing bindings for all sort of languajes so everyone can use Gtk/Gnome instead of focusing on one languaje that *has* the majority of developers out there.
Come on!, we want Objective-C Bindings!
The problem for us developers is not really the directory layout, for this there is LSB. Tough some people use the /opt dir for things like kde or OO.org (like slackware does with kde).
The problem comes with the libraries, i can compile my program on one system and it works, but in other systems it may fail because of the ABI incompatibility of some libraries you need (an example of this is glibc, just try to recompile it and watch how your system stops working).
Thats why we distribute the source, binaries are just painful to mantain. And no, i don't like statically linked stuff either.
I think you are mistaking OSS Developers for some sort of enterprise company (some are, but not all of us).
Most of us start a project that is useful for us only, later we found that it can be useful for other people, then we make our software Open Source, for everybody to use, share and modify.
We are not about users, we are about sharing. You can modify it if you don't like it, it works for me. OSS people are very kind and care about their users, but their users don't help, they always keep yelling out loud of how the software isn't what they want, that why the interface is ugly, blah, blah, blah.
Stop this crap at once, you don't pay me for coding, i already giving it for free, if you don't like it you can look for other projects or buy some software that does what you want!. Don't bother me with nonsenses, the code is there, help!
They didn't take in count all the popular windows viruses out there, this is not quite objetive.
If you expose a Linux machine to the internet, is unlikly that it will get cracked, and this can be better if you keep up to date all important software and don't run stupid things as ROOT or have running unecesary services.
Now if you put a Windows Machine on the internet, is likly that it will get a virus or will crash. Tought if you keep your machine updated it can be safe for a while, until the new exploit gets out, or M$ locks your machine for ever.
The problem for the masses is not hackers (i don't think there is enough hackers to crack all the servers out there one by one), but viruses and other exploits out there, this is where windows is very vulnerable, anyone remember the RPC problem? it will shutdown your windows box and you don't even need to touch it.
It does make sense from the license point, this will help to avoid using GPL code in close code (statically linking, mixing code, etc).
However, this attitudes makes Linux hard to work with, I have to live with half working drivers (yeah it works, but not all the functionability is supported, this to me is incomplete), hardware manufacturers won't open all. This is not by any means helping the adoption of Linux (on the desktop).
I don't know if I should take Linux seriously any more, by any means. As now that i can't choose between closed and open drivers (yeah I know, I can, but creating close drivers is becoming more dificult, that's my point).
Anyway the idea of creating a driver layer between the driver (closed drivers) and the kernel with a stable API/ABI sound quite good now. We would only need to modify the layer source for new kernels and support the stable API/ABI.
Sure, we can test all the code that comes in with others code base, however may things araise from this:
* Implementations of Standars: what if 2 implementations of a Standard (Open standard) that have been maded separatly comes up with a very similar code or even the same code (counting out variable names and stuff like that), current US laws and IP rights are very extreme as you can patent almost any process, algorithms, and other stuff.
Anyways, we are a comunity that makes and uses that code, we don't want any malicius code running on our machines anyway. We are a community, not a bunch of terrorists, or an organization creating a conspirancy to over take the world by hooking everyone with our software and latter control them. (ala M$)
Making a already propetary software Open Source is not as simple as just opening the source to everyone, a lot of OpenSource projects started small, with people that know each other and they spread the word, latter, after many people get known of the project, there was already a comunity around it.
:-)).
However this may not be the case of Java, lots of developers are interested on getting their hands on it, there is room for improvement (speed improvements
The netscape case was different, however look at mozilla today!