Here is the problem, you see. emacs has a bad learning curve. I have to learn tons of new keybindings, look into the help countless times etc. This is fine if I actually want to learn how to use emacs, but not if I just want to do my stuff. In this case, the environment must not get in the way. Most IDEs do get in the way, unfortunately. Personally, I found VC to be the best in this regard. KDevelop, Anjuta, they all bug me with tons of autogenerated code, autotools in general (they break so easily:/ ), the rigid directory structure etc.
Actually, the best thing would be an X11 editor which allows you to specify a custom build command, can be used via a GUI (not by keybindings), and has the behaviour of kate/gedit/nedit when editing. Oh, and an integrated terminal emulator would be nice. geany Comes close, but is make-only, and unstable.
If you think the Visual Studio is a good environment there is no hope for you.
Stuff your arrogance where the sun doesn't shine.
I for one have been using Visual Studio for more than six years. I used 6, 7 and 8. 6 IS crap, yes. But the rest of your posting radiates ignorance. The typical UNIX way (make/emacs/vi/shell) is not The Way To Enlightenment. I don't use VS for code generation, I use it because it automates stuff I just don't want to handle all the time. Building? One click. Debugging? MUCH easier than with gdb. Quick overview & access to all files? Done.
I do develop for Linux, too, and it constantly bugs me that I have to switch to the shell, type make/scons/whatever, see the error output, switch back to the editor, look for the file in the file requester, open it, switch back to see the exact error, switch back to the editor.... vi and emacs are damn confusing, gvim is ok, but doesn't have a file overview panel like VS has. My favourite editors in Linux are kate, nedit and gedit, but none of them have all helpful tools VS has. Oh, and then there is gdb. Debugging multithreaded stuff with gdb - yeah right. gdb often simply misses breakpoints, does not find the source (even when I specified the exact path in the source command), watching variables is unnecessarily difficult etc. gdb is an absolute nightmare to use. ddd is better, kdbg is best, but debugging is one of the things where an IDE shines: since it has knowledge about the overall project structure and the files it consists of, debugging can be much easier. Then there are additional benefits, like refactoring tools (mostly in Java IDEs though - see IDEA).
Could it be legal to write a simple script/program or include automatization in the program to fetch additional plugins easily? Kinda like the Firefox extension mechanism. If so, the GPL issue would not be in the users' way.
So, how about providing scripts/frontends for easy installation of the nvidia/ati drivers?
The GPL has its problems when a GPLed application has a plugin functionality, because then the plugins have to be GPLed, too, or else they cannot be used. Usually, plugins are loaded as shared objects (that is, they are linked dynamically at run-time), so the GPL applies here too. This can be solved either by providing a different ABI (rather hard to achieve) or by putting the application under another license.
Also, GPL for libraries is almost always a bad idea. Unlike programs, libraries are linked, technically weaving code together, but no lib code gets modified. Thats why the LGPL exists. xvid is GPLed, but this is OK because xvid cannot be used in closed-source binaries anyway (because of mpeg4 patents).
Well, if you change your graphic card, you have to edit xorg.conf. If you want WPA, you need to edit the config file. If the autodetection failed in detecting your monitor capabilites, you have to edit xorg.conf. If you want to set up a simple share, you need samba. (The KDE frontend writes invalid smb.conf files here.)
Perhaps you can explain why if linux dissatisfies you so, why you continue to use it.
Because in sum I like it. My real problem are not the problems in current distros, its the attitude of developers and Linux zealots, their justification of GUI failures.
How exactly do you backup your windows network share configuration?
I don't. Because it is MUCH easier to share directories in Windows. 10 seconds clicking - done. In Samba, I have to dig tons of manpages and howtos to achieve the same. Yes, yes, I know Samba scales better, but I don't want to set up a server for 5 billion users - I just want to share a directory for data exchange between laptop and PC. It took me eons to get the rights stuff to work. Correct charset translation still doesn't work (and I looked for this in Google for a damn long time, tried every configuration I found, nada). German characters like "Ö", "Ä", "ß" are displayed wrongly, Norway chars - the same etc. I gave up and simply use standardized substitutes (ae for ä etc.)
is it really so simple to configure a printer in windows, or is it just that you're more familiar with windows.
It is. Right now I have the problem that for unknown reasons, after a dist-upgrade, printing no longer works. Weird stuff like "@PJL JOB NAME="Ghost" " ist printed instead of the actual contents I wanted to print. I asked in IRC chnanels, message boards etc. Zero answers. Now I reboot into Windows for printing. And, I did not change ANYTHING to soft- or hardware - just apt-get dist-upgrade. How NICE. So I have to become a CUPS guru after all.
started using linux because I sought a challenge. I'm more interested in power, flexibility, and freedom. If you "want to do my stuff and nothing in my way", that's why they make macs.
Just perfect. You justify the GUI failure by artificially declaring Macs for mundanes and Linux for enlightened people.
Remember what I wrote earlier: And saying "then use Windows" is simply admitting the abysmal failure of being user-friendly. Its like saying "yeah, we admit it, Linux just plain sucks for everyday business work". Replace Windows with Mac, and it matches your quote.
I started with an Amiga 500, moved on to a 386, then a 486, then a Pentium 3, and now a Pentium 4. I used AmigaOS, DOS, BeOS, FreeBSD, Windows 3.11/95/98/2000/XP, Gentoo, Debian, and now Ubuntu. I coded for all these systems, and in the past it was "cool" to tweak them. In DOS, I had several custom CONFIG.SYS and AUTOEXEC.BAT files just to get some games to work (those damn 640K...) In AmigaOS, I had custom startup-sequence scripts etc. For this, editing config files indeed is fine.
But you know what? I don't want to become a SMB expert just to share a damn directory. I don't want to become a CUPS guru just to print a page. I don't want to become an X11 crack just to see something other than text. In short: I want to do my stuff and nothing in my way.
IS THIS SO DAMN HARD TO UNDERSTAND?
And then this nonsense:
SMB? That's windows networking stuff. Fine. Let us discuss why "Point, Click, Done!" is inferior to actually reading a man page and understanding what you're doing. What are the security settings on the folder you just shared? Did you share it for everyone on the computer? Everyone on your LAN? Everyone on the Internet? Did you share all the subfolders? Did the subfolders contain any links to other folders? Can people on the internet write to my share? Hey! SexyBoobies.EXE! Awesome.. CLICK!
You know, I can configure all of this with the Share dialog in Windows. No problem. Trying to do it by editing smb.conf took me WEEKS to get it to work. It was one of the most frustrating experiences I had when installing my first Linux distro. I wrote this smb.conf ONCE and reuse it just to get it to work quickly (I won't go the agonizing way through dozens of useless-because-incorrect howtos and cryptic man pages again).
Give this to an experienced UI designer (no, coders usually are NOT good UI designers), and you will be amazed. See the Apple solution.
2) I don't have to tweak the registry to share a directory. I don't have to tweak the registry just to install graphics drivers. I don't have to tweak the registry to configure the WPA supplicant. In short: I don't have to tweak the registry as often as I have to edit a configfile in Linux.
3) Just because SuSE fails its OK not to try at all? What kind of logic is that?
4) "A few options are availably through dialogs of explorer" - THESE ARE THE OPTIONS I AM SPEAKING OF! The Average Joe options! Stop putting Windows as an example, I know Windows hides all advanced stuff, and its not the point!
5) Why do you assume that Windows is everybody's ideal when it comes to GUI design? Holding Apple as an ideal would make much more sense, since they ARE experienced in this area (and succeed in simple Average Joe stuff admirably).
Granted, these features are not important for John Sixpack wanting to share one directory. But once you have a few dozen shares distributed over a few dozen machines all this simple user interfaces simply are a nightmare.
And this is exactly why manual configfiles editing should be OPTIONAL. Instead, it is MANDATORY, which is just as bad as not being able to do it at all. The solution is to provide BOTH. A simple GUI for everyday tasks Joe Sixpack wants, and the configfile for the power user. Simple tasks should be manageable with a GUI. Why should Joe learn to edit the xorg.conf just to permanently set the vertical frequency? Or changing the currently used graphics card because he bought a new one? OF COURSE configfiles are a good thing for power users and admins. But don't forget Joe.
What is this obsession people have with linux being accepted mainstream? What Raymond claims is true, if linux (and FOSS in general) wants to be adopted mainstream it's needs to compromise it's ideals.
But maybe linux doesn't need to be mainstream. Operating systems tend to be developed to accomodate the lowest commond denominator. Microsoft has a lot of grandmas and grandpas with eMachines to satisfy. Apple has a lot of clueless art students to "empower". If linux's lowest common denominator is a bunch of Klingons, that's a good thing.
Translation: Linux must not become useful for non-elite people!
If linux's demographic was magically expanded to include grandma, grandpa and emo art students that aren't willing to learn how to compile a kernel, it would have to contend with the point-and-click, plug-and-play mentality of it's clueless user base, meaning that developers would be forced to spend time making linux idiot proof instead of optimizing, debugging, enhancing and advancing.
Translation: We don't want to spend time learning how to design (not write!) GUIs that don't suck. GUI HIGs - oh no! User friendliness - go away, antichrist!
I, on the other hand, would encourage linux, and all open source projects to minimize their user bases. I'm not suggesting that FOSS should be kept a secret, but you certainly don't have to waste resources buying an idiotic full page ad in the New York Times. A good open source project (like a good democracy) demands an educated user base, not a herd of people who do things because advertisments tell them to.
Translation: Linux is not for Joe Average, and never shall be! Even though it is possible, but we don't want to support those mundanes!
That being said, there's no reason you can't have linux for stupid people (ubuntu (no offense, I use ubuntu)) and linux for smart people (freebsd, ha). That's the beauty of open source software.
This is a contradiction. So there shall be a Joe Average Linux? Ah, you finally see that Linux is about choices?
In summary, if you already use linux and you still bought an ipod instead of the countless non-DRM, usbstorage friendly media players that are better and cheaper, then you are stupid and you should probably buy a mac. If you already own an ipod and you want to get into linux, life sucks, get started reverse engineering iTunes.
Words fail me to describe this one.
People like you are obstacles in the way of finally developing a Linux for the masses. Writing good GUIs is not "wasting time"; instead, it is one of the most difficult and challenging tasks in application development. This is why apple has LOTS of experts in this area, and this is why Linux (and many Windows) GUIs simply fail. If you want your elitist Linux, then have it, but get out of the way when it comes to creating a Linux for everyone.
I really hope Ubuntu finally succeeds in this, so that you can see how wrong you are. All these problems you redefine into "Linux features" are actually just design errors. Having to manually edit xorg.conf is not wise, not l33t, its simply STUPID. (Even in Ubuntu it is required sometimes). There is NO good GUI for xorg.conf editing (no, xorgcfg is not good AT ALL, it is a UI design antithesis). There is NO good GUI for configuring smb.conf just for simple sharing. In Windows, right-click on the folder, press "Share", and then "Share this folder". Done.
And saying "then use Windows" is simply admitting the abysmal failure of being user-friendly. Its like saying "yeah, we admit it, Linux just plain sucks for everyday business work".
The real problem is that Linux devs have their goals carved in stone. They never rethink what they consider to be worthy pursuing, which is a shame. It should be every developer's goal to provide a good user interface, instead this is seen as l4m3, which is just plain idiotic. Creating a good, task-oriented GUI is not l4m3, it is a tremendous sucess.
Where is the reason for using Theora rather than RealVideo or WMV? There is absolutely zero support for Theora, zero tools, zero stream servers, zero tutorials & docs, and no experienced users. MPEG-4 is THE de-facto standard for videos, successfully phasing out MPEG-1.
Vorbis is in a better shape; Icecast/Shoutcast support it. But MP3 is the de-facto standard. It is also easier to implement MP3 in a microcontroller for an mp3 player than Vorbis. Also, support is still much worse than for WMA/RealAudio. Guess why Amazon uses those.
The binary nvidia/ati drivers are a MUST for 3d games in Linux and for XGL. Whine all you want, without those drivers Linux just plain sucks as a gaming platform and falls back to 2000's GUIs. (Forget about SVG for the GUI, forget about XGL as backend...)
I hate to say this (I don't like ESR), but he is right. If Linux shall become a serious alternative to Windows for all the Joe Averages out there, there has to be a compromise. But hey, Linux is all about choices, right? If you don't like it, stay with the purist Linux. The other 99% choose the Desktop Linux one.
"But I wonder why a "Joe Sixpack" user would need to configure X and whether they would even need to know about Samba or xorg."
As an AP mentioned before, xorg.conf HAS to be touched at times because of unsuccessfull autoconfiguration (and changed hardware, like a new video card). As for smb.conf: it IS in Joe's range to share one of his folders so his buddies can watch the photos he made. Unfortunately, this means digging through HOWTOs and smb.conf.
Nice ideas. But irrelevant for the mass market. These projects may yield usable hardware in what, 10 years? By then, they will be hopelessly outdated. So, open hardware for the typical geek, but not for anybody else.
Now, there are two groups: elitists who want to keep Linux hard to grasp (and absolutely HATE Joe Average), and the ones who want Linux to rule the desktop. The second group HAS to accept binary drivers, applications, games, or else mass adoption simply will not happen.
WPA is still a nightmare (even in Ubuntu). It is not trivial to install nvidia/ati drivers (especially ati), which are essential for useful 3D acceleration. The whole access rights stuff is hard to grasp at first (ESPECIALLY with Samba, which is also horrible to configure BTW). GUIs are not as responsive as in OSX or Windows (I suspect the font rendering to play a large role in this). Configuring X is *still* awful (and necessary for setting trivial stuff like the physical screen size for correct DPI). As long as your hardware is recognized, the driver issue is non-existant (especially Ubuntu has a wonderful autoconfiguration). But once something does not work well (even the tiniest bit), you better spend months learning everything about Linux.
Oh, and the terminal should not be necessary for configuring *anything* for a desktop. Not samba (for sharing), not wpa_supplicant, not xorg.conf,.... stuff like Apache is another thing entirely - one does not expect Joe Average to use Apache, but Joe might be interested in encrypted WLAN (and WEP is crap).
The problem is: There is no alternative to Flash Video. ALL other media players are flawed, freezes the browser while loading its components, or crashing, and taking the browser right along with it. This is especially true in Linux. The Flash Video player never crashes, never freezes the browser, it just works. MPlayer plugin? Crashes a LOT. gxine plugin? Starts a new gxine window, which is an insanely STUPID idea, and also crashes a lot. Kaffeine starts a new window, too, but at least doesnt crash. The totem plugin is very good, but still unstable. The VLC plugin crashes.
Also, there is the licensing issue of course.
In Windows, its a LOT better. Windows Media Player, while being bloatware and just plain ugly, works (and does not open a new player window....). For those who want a lean player, Media Player Classic is there, and provides a plugin, too. And it just works, too.
I would love to see a STABLE, non-freezing plugin that loads almost instantly, is lean, very usable, and streams perfectly. On the server side, a useful theora-stream server (since mpeg4 and avc are not free as in MONEY, they are not that hot for noncommercial stuff).
One other thing: are the FLV specs known? If so, it should not be impossible to write a new plugin (or extend current ones) for FLV video playback.
Indeed, GUI development is one of the areas where managed environments shine. However, it IS possible to deal with this in C++. I agree that if I want to put a label in a window, I dont want to have to handle its deallocation, i.e. I want a "fire-and-forget" feature. I accomplished this in my wrapper by letting the parent deal with its children. It is a very simple concept, but works very well. I can write "new Label(mainwindow,"Here is my label",22,11);" and do not ever have to worry about a delete call or about memory leaks.
IMO the best thing is to make the GUI purely data-driven, however. Stuff the entire layout in a XML, and the application only has to connect its event handlers to the named signals (assuming a signals-based event dispatching mechanism). AFAIK libglade does it like this. The huge advantage is that redesigning the interface no longer requires code modification. Don't like the button position? Move it down, you dont even have to recompile.
Sort of. The Windows version has the same license, but has no VisualC support - this is included in the commercial version only. The free one only supports MinGW.
E-UAE is a backport of WinUAE's code back to Linux, so it has a lot of the same features and performance of WinUAE. I'll admit that the rest of the base for emulators is rather poor, but that seems more due to a lack of demand than a lack of talent. But there are a lot of them out there (including VisualBoyAdvance, ZSNES, Snes9x, etc.) that work exactly like their Windows counterparts.
E-UAE was unknown to me, and seems to be in a rather experimental state. Still, I welcome the WinUAE backport.
I won't pretend that the Gimp is better than Photoshop, but, really, who needs Photoshop's features besides graphics pros? (I'm talking about the guys that'd probably spring for ZBrush if they really need it, not Joe Schmoe who wants to make neat pics for his Intarweb site.)
"Besides graphics pros" is one hell of an exclusion. It means that no graphics studios will use Linux. As for Joe, his problem is that nearly EVERYTHING is written for Photoshop. 99,9999% of all tutorials and books available are for PS. It is only logical to choose PS over GIMP then.
3D Studio Max is supplanted by SoftImage|XSI and others depending on the guys using it, be it because 3DS is too expensive or whatever. I do know for a fact that game design guys use everything from Maya to SoftImage and everything in-between, depending on who owns what licenses. Maya and others also have Linux ports, which, I hear, are identical to the Windows version. If you have one of those licenses, tell then you want a Linux version. It's not that hard. But in the mean time, there's always Blender, Aoi, and K3d....
It does not matter that 3DS is being supplanted, it still is widely used.
Moving to another way of doing things is always a compromise. If you or Joe Sixpack isn't willing to compromise on anything, then stay where you're at and don't say a goddamned word.
This does not match. So compromise, but Linux shall take over the Desktop? This excuse is widely used and STILL pointless. With this type of thinking Linux will remain an exotic OS on the desktop. The main Linux Desktop mantra should be: support everything. Not: adapt yourself to our will.
No you didn't. You listed a bunch of stuff for which there are numerous choices, but your unwillingness to try anything "different" is causing you to judge Linux and its users unfairly.
At least for Photoshop there is no choice, for reasons I already mentioned. Max is a different matter, but still, tons of plugins and toolchains are 3DSMax only. And I TRIED different stuff, it was inferior (see: UAE for Linux). I am not judging Linux, I am stating my opinion, and there is NOTHING you can do to stop me.
Go write a Photoshop replacement that isn't GIMP. Oh, too hard? Not for you? Then stay on Windows. You've been given options. Look into them or go away.
Again, this can be translated to: the thousands of pros shall stay away from Linux. Kinda counter-productive to the Linux adoption, isnt it? Also, Joe Schmoe is much better off with PS if he wants to learn it by books, tips, tricks, tutorials etc. and eventually (after years of learning) getting a job as PS pro. Try that with GIMP (assuming anyone would want a GIMP pro).
You've been given options. Look into them or go away.
Pure ignorance. I suggested stuff, but as always, this is ignored. Other devs actually LOOK into the proposed points, and use them for improvement of their software.
Yes, I realize this negative attitude turns some users away.
Not SOME. ALMOST EVERY USER.
Most of the things I've listed show up in Gentoo Portage in a very quick search. I'm certain that.deb or.rpm packages exist for most if not all of them. If only source is available, then start packaging it and provide that for other users! Push for your packages to be added to major repositories!
And what if I were an user with not enough skills? No rights to criticize? "Shut up unless you are a l33t coder?"
Stop whining about a problem that doesn't really exist.
Yeah, Linux is pretty common among PC desktops after all. Oh wait..
n order to develop such an application the developer is highly restricted in the resources available. As compared with writing a GPL program, where any GPL code is freely available for usage. A programmer would even need to check if they can use standard libraries...
Since most libraries are LGPL- or BSD-licensed, this is no problem. (I don't know any standard lib that is GPL-licensed, simply because GPL makes no sense for libraries - can you imagine the HUGE problems that a GPLed glibc would bring?) "highly restricted" is absolute nonsense.
However there's the side effect that using that "binary blob" renders the entire kernel unsupported.
Unsupported by whom? The kernel devs? Irrelevant to the end user, as long as the distro devs support it.
Well, as soon as this fails because of compile errors (most likely some developer versions of needed libraries missing), things become order of magnitudes more difficult. And good luck trying to find help for that. No,.deb/.rpm are there for a reason.
Linux is no market for PC games, but ignoring them is an insanely stupid thing to do; game development has become an ENORMOUS industry and is one good reason to buy a PC. WoW for Linux would be a HUGE boost for the Linux desktop. Also, most graphics artists use Windows or Mac because of Photoshop. PS is THE industry standard, GIMP would have to mimic its surface 1:1 to stand a chance.
Slashdot Mirror
Well, try to move away one inch from the enforced directory structure in Anjuta.
Here is the problem, you see. emacs has a bad learning curve. I have to learn tons of new keybindings, look into the help countless times etc. This is fine if I actually want to learn how to use emacs, but not if I just want to do my stuff. In this case, the environment must not get in the way. Most IDEs do get in the way, unfortunately. Personally, I found VC to be the best in this regard. KDevelop, Anjuta, they all bug me with tons of autogenerated code, autotools in general (they break so easily :/ ), the rigid directory structure etc.
Actually, the best thing would be an X11 editor which allows you to specify a custom build command, can be used via a GUI (not by keybindings), and has the behaviour of kate/gedit/nedit when editing. Oh, and an integrated terminal emulator would be nice. geany Comes close, but is make-only, and unstable.
Rubbish. I can write programs faster in VS than other in their vim environments. Its not the tool, stupid: its the skills.
Oh, and I don't use make. I use scons. Also, I don't use vi.
You must be recently freed of Windows where you are simply told what you want, here in the world of FOSS you have choices.
Hah. In VC, I can choose my directory structure freely. The source can be anywhere, libs can be anywhere, etc.
In Linux IDEs I am told how my directory structure is supposed to look like.
If you think the Visual Studio is a good environment there is no hope for you.
Stuff your arrogance where the sun doesn't shine.
I for one have been using Visual Studio for more than six years. I used 6, 7 and 8. 6 IS crap, yes. But the rest of your posting radiates ignorance. The typical UNIX way (make/emacs/vi/shell) is not The Way To Enlightenment. I don't use VS for code generation, I use it because it automates stuff I just don't want to handle all the time. Building? One click. Debugging? MUCH easier than with gdb. Quick overview & access to all files? Done.
I do develop for Linux, too, and it constantly bugs me that I have to switch to the shell, type make/scons/whatever, see the error output, switch back to the editor, look for the file in the file requester, open it, switch back to see the exact error, switch back to the editor.... vi and emacs are damn confusing, gvim is ok, but doesn't have a file overview panel like VS has. My favourite editors in Linux are kate, nedit and gedit, but none of them have all helpful tools VS has. Oh, and then there is gdb. Debugging multithreaded stuff with gdb - yeah right. gdb often simply misses breakpoints, does not find the source (even when I specified the exact path in the source command), watching variables is unnecessarily difficult etc. gdb is an absolute nightmare to use. ddd is better, kdbg is best, but debugging is one of the things where an IDE shines: since it has knowledge about the overall project structure and the files it consists of, debugging can be much easier. Then there are additional benefits, like refactoring tools (mostly in Java IDEs though - see IDEA).
This brings up an interesting question.
Could it be legal to write a simple script/program or include automatization in the program to fetch additional plugins easily? Kinda like the Firefox extension mechanism. If so, the GPL issue would not be in the users' way.
So, how about providing scripts/frontends for easy installation of the nvidia/ati drivers?
The GPL has its problems when a GPLed application has a plugin functionality, because then the plugins have to be GPLed, too, or else they cannot be used. Usually, plugins are loaded as shared objects (that is, they are linked dynamically at run-time), so the GPL applies here too. This can be solved either by providing a different ABI (rather hard to achieve) or by putting the application under another license.
Also, GPL for libraries is almost always a bad idea. Unlike programs, libraries are linked, technically weaving code together, but no lib code gets modified. Thats why the LGPL exists. xvid is GPLed, but this is OK because xvid cannot be used in closed-source binaries anyway (because of mpeg4 patents).
Well, if you change your graphic card, you have to edit xorg.conf. If you want WPA, you need to edit the config file. If the autodetection failed in detecting your monitor capabilites, you have to edit xorg.conf. If you want to set up a simple share, you need samba. (The KDE frontend writes invalid smb.conf files here.)
Perhaps you can explain why if linux dissatisfies you so, why you continue to use it.
Because in sum I like it. My real problem are not the problems in current distros, its the attitude of developers and Linux zealots, their justification of GUI failures.
How exactly do you backup your windows network share configuration?
I don't. Because it is MUCH easier to share directories in Windows. 10 seconds clicking - done.
In Samba, I have to dig tons of manpages and howtos to achieve the same. Yes, yes, I know Samba scales better, but I don't want to set up a server for 5 billion users - I just want to share a directory for data exchange between laptop and PC. It took me eons to get the rights stuff to work. Correct charset translation still doesn't work (and I looked for this in Google for a damn long time, tried every configuration I found, nada). German characters like "Ö", "Ä", "ß" are displayed wrongly, Norway chars - the same etc. I gave up and simply use standardized substitutes (ae for ä etc.)
is it really so simple to configure a printer in windows, or is it just that you're more familiar with windows.
It is. Right now I have the problem that for unknown reasons, after a dist-upgrade, printing no longer works. Weird stuff like "@PJL JOB NAME="Ghost" " ist printed instead of the actual contents I wanted to print. I asked in IRC chnanels, message boards etc. Zero answers. Now I reboot into Windows for printing. And, I did not change ANYTHING to soft- or hardware - just apt-get dist-upgrade. How NICE. So I have to become a CUPS guru after all.
started using linux because I sought a challenge. I'm more interested in power, flexibility, and freedom. If you "want to do my stuff and nothing in my way", that's why they make macs.
Just perfect. You justify the GUI failure by artificially declaring Macs for mundanes and Linux for enlightened people.
Remember what I wrote earlier: And saying "then use Windows" is simply admitting the abysmal failure of being user-friendly. Its like saying "yeah, we admit it, Linux just plain sucks for everyday business work". Replace Windows with Mac, and it matches your quote.
Boy, you are wrong.
I started with an Amiga 500, moved on to a 386, then a 486, then a Pentium 3, and now a Pentium 4. I used AmigaOS, DOS, BeOS, FreeBSD, Windows 3.11/95/98/2000/XP, Gentoo, Debian, and now Ubuntu. I coded for all these systems, and in the past it was "cool" to tweak them. In DOS, I had several custom CONFIG.SYS and AUTOEXEC.BAT files just to get some games to work (those damn 640K...) In AmigaOS, I had custom startup-sequence scripts etc. For this, editing config files indeed is fine.
But you know what? I don't want to become a SMB expert just to share a damn directory. I don't want to become a CUPS guru just to print a page. I don't want to become an X11 crack just to see something other than text. In short: I want to do my stuff and nothing in my way.
IS THIS SO DAMN HARD TO UNDERSTAND?
And then this nonsense:
SMB? That's windows networking stuff. Fine. Let us discuss why "Point, Click, Done!" is inferior to actually reading a man page and understanding what you're doing. What are the security settings on the folder you just shared? Did you share it for everyone on the computer? Everyone on your LAN? Everyone on the Internet? Did you share all the subfolders? Did the subfolders contain any links to other folders? Can people on the internet write to my share? Hey! SexyBoobies.EXE! Awesome.. CLICK!
You know, I can configure all of this with the Share dialog in Windows. No problem. Trying to do it by editing smb.conf took me WEEKS to get it to work. It was one of the most frustrating experiences I had when installing my first Linux distro. I wrote this smb.conf ONCE and reuse it just to get it to work quickly (I won't go the agonizing way through dozens of useless-because-incorrect howtos and cryptic man pages again).
Give this to an experienced UI designer (no, coders usually are NOT good UI designers), and you will be amazed. See the Apple solution.
And get a clue.
1) no one finds it "cool" to tweak the registry.
2) I don't have to tweak the registry to share a directory.
I don't have to tweak the registry just to install graphics drivers.
I don't have to tweak the registry to configure the WPA supplicant.
In short: I don't have to tweak the registry as often as I have to edit a configfile in Linux.
3) Just because SuSE fails its OK not to try at all? What kind of logic is that?
4) "A few options are availably through dialogs of explorer" - THESE ARE THE OPTIONS I AM SPEAKING OF! The Average Joe options! Stop putting Windows as an example, I know Windows hides all advanced stuff, and its not the point!
5) Why do you assume that Windows is everybody's ideal when it comes to GUI design? Holding Apple as an ideal would make much more sense, since they ARE experienced in this area (and succeed in simple Average Joe stuff admirably).
Granted, these features are not important for John Sixpack wanting to share one directory. But once you have a few dozen shares distributed over a few dozen machines all this simple user interfaces simply are a nightmare.
And this is exactly why manual configfiles editing should be OPTIONAL. Instead, it is MANDATORY, which is just as bad as not being able to do it at all. The solution is to provide BOTH. A simple GUI for everyday tasks Joe Sixpack wants, and the configfile for the power user. Simple tasks should be manageable with a GUI. Why should Joe learn to edit the xorg.conf just to permanently set the vertical frequency? Or changing the currently used graphics card because he bought a new one? OF COURSE configfiles are a good thing for power users and admins. But don't forget Joe.
What is this obsession people have with linux being accepted mainstream? What Raymond claims is true, if linux (and FOSS in general) wants to be adopted mainstream it's needs to compromise it's ideals.
But maybe linux doesn't need to be mainstream. Operating systems tend to be developed to accomodate the lowest commond denominator. Microsoft has a lot of grandmas and grandpas with eMachines to satisfy. Apple has a lot of clueless art students to "empower". If linux's lowest common denominator is a bunch of Klingons, that's a good thing.
Translation: Linux must not become useful for non-elite people!
If linux's demographic was magically expanded to include grandma, grandpa and emo art students that aren't willing to learn how to compile a kernel, it would have to contend with the point-and-click, plug-and-play mentality of it's clueless user base, meaning that developers would be forced to spend time making linux idiot proof instead of optimizing, debugging, enhancing and advancing.
Translation: We don't want to spend time learning how to design (not write!) GUIs that don't suck. GUI HIGs - oh no! User friendliness - go away, antichrist!
I, on the other hand, would encourage linux, and all open source projects to minimize their user bases. I'm not suggesting that FOSS should be kept a secret, but you certainly don't have to waste resources buying an idiotic full page ad in the New York Times. A good open source project (like a good democracy) demands an educated user base, not a herd of people who do things because advertisments tell them to.
Translation: Linux is not for Joe Average, and never shall be! Even though it is possible, but we don't want to support those mundanes!
That being said, there's no reason you can't have linux for stupid people (ubuntu (no offense, I use ubuntu)) and linux for smart people (freebsd, ha). That's the beauty of open source software.
This is a contradiction. So there shall be a Joe Average Linux? Ah, you finally see that Linux is about choices?
In summary, if you already use linux and you still bought an ipod instead of the countless non-DRM, usbstorage friendly media players that are better and cheaper, then you are stupid and you should probably buy a mac. If you already own an ipod and you want to get into linux, life sucks, get started reverse engineering iTunes.
Words fail me to describe this one.
People like you are obstacles in the way of finally developing a Linux for the masses. Writing good GUIs is not "wasting time"; instead, it is one of the most difficult and challenging tasks in application development. This is why apple has LOTS of experts in this area, and this is why Linux (and many Windows) GUIs simply fail. If you want your elitist Linux, then have it, but get out of the way when it comes to creating a Linux for everyone.
I really hope Ubuntu finally succeeds in this, so that you can see how wrong you are. All these problems you redefine into "Linux features" are actually just design errors. Having to manually edit xorg.conf is not wise, not l33t, its simply STUPID. (Even in Ubuntu it is required sometimes). There is NO good GUI for xorg.conf editing (no, xorgcfg is not good AT ALL, it is a UI design antithesis). There is NO good GUI for configuring smb.conf just for simple sharing. In Windows, right-click on the folder, press "Share", and then "Share this folder". Done.
And saying "then use Windows" is simply admitting the abysmal failure of being user-friendly. Its like saying "yeah, we admit it, Linux just plain sucks for everyday business work".
The real problem is that Linux devs have their goals carved in stone. They never rethink what they consider to be worthy pursuing, which is a shame. It should be every developer's goal to provide a good user interface, instead this is seen as l4m3, which is just plain idiotic. Creating a good, task-oriented GUI is not l4m3, it is a tremendous sucess.
And now back to the real world.
Where is the reason for using Theora rather than RealVideo or WMV? There is absolutely zero support for Theora, zero tools, zero stream servers, zero tutorials & docs, and no experienced users. MPEG-4 is THE de-facto standard for videos, successfully phasing out MPEG-1.
Vorbis is in a better shape; Icecast/Shoutcast support it. But MP3 is the de-facto standard. It is also easier to implement MP3 in a microcontroller for an mp3 player than Vorbis. Also, support is still much worse than for WMA/RealAudio. Guess why Amazon uses those.
The binary nvidia/ati drivers are a MUST for 3d games in Linux and for XGL. Whine all you want, without those drivers Linux just plain sucks as a gaming platform and falls back to 2000's GUIs. (Forget about SVG for the GUI, forget about XGL as backend...)
I hate to say this (I don't like ESR), but he is right. If Linux shall become a serious alternative to Windows for all the Joe Averages out there, there has to be a compromise. But hey, Linux is all about choices, right? If you don't like it, stay with the purist Linux. The other 99% choose the Desktop Linux one.
"But I wonder why a "Joe Sixpack" user would need to configure X and whether they would even need to know about Samba or xorg."
As an AP mentioned before, xorg.conf HAS to be touched at times because of unsuccessfull autoconfiguration (and changed hardware, like a new video card). As for smb.conf: it IS in Joe's range to share one of his folders so his buddies can watch the photos he made. Unfortunately, this means digging through HOWTOs and smb.conf.
Nice ideas. But irrelevant for the mass market. These projects may yield usable hardware in what, 10 years? By then, they will be hopelessly outdated. So, open hardware for the typical geek, but not for anybody else.
Now, there are two groups: elitists who want to keep Linux hard to grasp (and absolutely HATE Joe Average), and the ones who want Linux to rule the desktop. The second group HAS to accept binary drivers, applications, games, or else mass adoption simply will not happen.
Define "Just Work".
.... stuff like Apache is another thing entirely - one does not expect Joe Average to use Apache, but Joe might be interested in encrypted WLAN (and WEP is crap).
WPA is still a nightmare (even in Ubuntu). It is not trivial to install nvidia/ati drivers (especially ati), which are essential for useful 3D acceleration. The whole access rights stuff is hard to grasp at first (ESPECIALLY with Samba, which is also horrible to configure BTW). GUIs are not as responsive as in OSX or Windows (I suspect the font rendering to play a large role in this). Configuring X is *still* awful (and necessary for setting trivial stuff like the physical screen size for correct DPI).
As long as your hardware is recognized, the driver issue is non-existant (especially Ubuntu has a wonderful autoconfiguration). But once something does not work well (even the tiniest bit), you better spend months learning everything about Linux.
Oh, and the terminal should not be necessary for configuring *anything* for a desktop. Not samba (for sharing), not wpa_supplicant, not xorg.conf,
Douglas Engelbart DID invent the mouse. http://en.wikipedia.org/wiki/Douglas_Engelbart is correct. Verify it in any computer history book.
The problem is:
There is no alternative to Flash Video.
ALL other media players are flawed, freezes the browser while loading its components, or crashing, and taking the browser right along with it. This is especially true in Linux. The Flash Video player never crashes, never freezes the browser, it just works. MPlayer plugin? Crashes a LOT. gxine plugin? Starts a new gxine window, which is an insanely STUPID idea, and also crashes a lot. Kaffeine starts a new window, too, but at least doesnt crash. The totem plugin is very good, but still unstable. The VLC plugin crashes.
Also, there is the licensing issue of course.
In Windows, its a LOT better. Windows Media Player, while being bloatware and just plain ugly, works (and does not open a new player window....). For those who want a lean player, Media Player Classic is there, and provides a plugin, too. And it just works, too.
I would love to see a STABLE, non-freezing plugin that loads almost instantly, is lean, very usable, and streams perfectly. On the server side, a useful theora-stream server (since mpeg4 and avc are not free as in MONEY, they are not that hot for noncommercial stuff).
One other thing: are the FLV specs known? If so, it should not be impossible to write a new plugin (or extend current ones) for FLV video playback.
Indeed, GUI development is one of the areas where managed environments shine. However, it IS possible to deal with this in C++. I agree that if I want to put a label in a window, I dont want to have to handle its deallocation, i.e. I want a "fire-and-forget" feature. I accomplished this in my wrapper by letting the parent deal with its children. It is a very simple concept, but works very well. I can write "new Label(mainwindow,"Here is my label",22,11);" and do not ever have to worry about a delete call or about memory leaks.
IMO the best thing is to make the GUI purely data-driven, however. Stuff the entire layout in a XML, and the application only has to connect its event handlers to the named signals (assuming a signals-based event dispatching mechanism). AFAIK libglade does it like this. The huge advantage is that redesigning the interface no longer requires code modification. Don't like the button position? Move it down, you dont even have to recompile.
Sort of. The Windows version has the same license, but has no VisualC support - this is included in the commercial version only. The free one only supports MinGW.
E-UAE is a backport of WinUAE's code back to Linux, so it has a lot of the same features and performance of WinUAE. I'll admit that the rest of the base for emulators is rather poor, but that seems more due to a lack of demand than a lack of talent. But there are a lot of them out there (including VisualBoyAdvance, ZSNES, Snes9x, etc.) that work exactly like their Windows counterparts.
.deb or .rpm packages exist for most if not all of them. If only source is available, then start packaging it and provide that for other users! Push for your packages to be added to major repositories!
E-UAE was unknown to me, and seems to be in a rather experimental state. Still, I welcome the WinUAE backport.
I won't pretend that the Gimp is better than Photoshop, but, really, who needs Photoshop's features besides graphics pros? (I'm talking about the guys that'd probably spring for ZBrush if they really need it, not Joe Schmoe who wants to make neat pics for his Intarweb site.)
"Besides graphics pros" is one hell of an exclusion. It means that no graphics studios will use Linux. As for Joe, his problem is that nearly EVERYTHING is written for Photoshop. 99,9999% of all tutorials and books available are for PS. It is only logical to choose PS over GIMP then.
3D Studio Max is supplanted by SoftImage|XSI and others depending on the guys using it, be it because 3DS is too expensive or whatever. I do know for a fact that game design guys use everything from Maya to SoftImage and everything in-between, depending on who owns what licenses. Maya and others also have Linux ports, which, I hear, are identical to the Windows version. If you have one of those licenses, tell then you want a Linux version. It's not that hard. But in the mean time, there's always Blender, Aoi, and K3d....
It does not matter that 3DS is being supplanted, it still is widely used.
Moving to another way of doing things is always a compromise. If you or Joe Sixpack isn't willing to compromise on anything, then stay where you're at and don't say a goddamned word.
This does not match. So compromise, but Linux shall take over the Desktop? This excuse is widely used and STILL pointless. With this type of thinking Linux will remain an exotic OS on the desktop. The main Linux Desktop mantra should be: support everything. Not: adapt yourself to our will.
No you didn't. You listed a bunch of stuff for which there are numerous choices, but your unwillingness to try anything "different" is causing you to judge Linux and its users unfairly.
At least for Photoshop there is no choice, for reasons I already mentioned. Max is a different matter, but still, tons of plugins and toolchains are 3DSMax only. And I TRIED different stuff, it was inferior (see: UAE for Linux). I am not judging Linux, I am stating my opinion, and there is NOTHING you can do to stop me.
Go write a Photoshop replacement that isn't GIMP. Oh, too hard? Not for you? Then stay on Windows. You've been given options. Look into them or go away.
Again, this can be translated to: the thousands of pros shall stay away from Linux. Kinda counter-productive to the Linux adoption, isnt it?
Also, Joe Schmoe is much better off with PS if he wants to learn it by books, tips, tricks, tutorials etc. and eventually (after years of learning) getting a job as PS pro. Try that with GIMP (assuming anyone would want a GIMP pro).
You've been given options. Look into them or go away.
Pure ignorance. I suggested stuff, but as always, this is ignored. Other devs actually LOOK into the proposed points, and use them for improvement of their software.
Yes, I realize this negative attitude turns some users away.
Not SOME. ALMOST EVERY USER.
Most of the things I've listed show up in Gentoo Portage in a very quick search. I'm certain that
And what if I were an user with not enough skills? No rights to criticize? "Shut up unless you are a l33t coder?"
Stop whining about a problem that doesn't really exist.
Yeah, Linux is pretty common among PC desktops after all.
Oh wait..
n order to develop such an application the developer is highly restricted in the resources available. As compared with writing a GPL program, where any GPL code is freely available for usage. A programmer would even need to check if they can use standard libraries...
Since most libraries are LGPL- or BSD-licensed, this is no problem. (I don't know any standard lib that is GPL-licensed, simply because GPL makes no sense for libraries - can you imagine the HUGE problems that a GPLed glibc would bring?) "highly restricted" is absolute nonsense.
However there's the side effect that using that "binary blob" renders the entire kernel unsupported.
Unsupported by whom? The kernel devs? Irrelevant to the end user, as long as the distro devs support it.
Well, as soon as this fails because of compile errors (most likely some developer versions of needed libraries missing), things become order of magnitudes more difficult. And good luck trying to find help for that. No, .deb/.rpm are there for a reason.
Linux is no market for PC games, but ignoring them is an insanely stupid thing to do; game development has become an ENORMOUS industry and is one good reason to buy a PC. WoW for Linux would be a HUGE boost for the Linux desktop. Also, most graphics artists use Windows or Mac because of Photoshop. PS is THE industry standard, GIMP would have to mimic its surface 1:1 to stand a chance.