And yes, I can't think of anything more interesting
.
This is a draft of an open letter I'm going to post on various Linux forums but since there are lots of opinions here, I'll treat you lot to a sneak preview!
Its basically about the current state of the linux desktop, and my personal views on it all.
Incidentally, after using Windows XP and hating it, switching back to Slackware 9.0 and writing this, my own arguments have convinced me to install Windows 2000 on my fast machine. Its fantastic, I never liked XP because it was too childlike, WinXP is probably the most balanced OS I have ever used. Best of all, I dont have to constantly fiddle with it, as is the temptation with Linux (I cant help myself.. give me buttons to press...).
Anyways, enjoy! By the way, Asher will either agree wholeheartedly, or disagree big time. Or both
.
"
An open letter to the GNU/Linux community
We are witnessing today a struggle for dominance in the operating system market. For many years, Microsoft's monopoly was near absolute and now, it is being threatened by the Open Source community. Closed proprietery ownership is being undermined by freedom of source and use. Power is being shifted from the halls of control at the headquarters of a distant and indifferent multinational corporation back to the fingertips of the users.
Most within this community, and indeed many objective observers argue that Linux is a superior kernel to Windows. From reading reviews, tests, and from personal experience, I can say that Linux is faster, more secure, more stable, more adaptable and portable, and more open than Microsoft's offering.
However, these are characteristics that are most essential in areas such as servers, clusters, and other security conscious and networking applications. The new frontier for Linux is the desktop.
This is a far different cup of tea to the servers. Desktop users want ease of use, they want speed, they want the ability to run their applications. They are not interested so much in stability or security, a factor which, as we have seen recently, has backfired upon them. The superiority of one kernel over another is largely irrelevant to them - as they only care about what they can see and do with what they can see.
This cannot be changed. There will be no shift in the desktop market back to using the commandline, or gaining power at the expense of useability. Mice are being worn out at a faster and faster rate, while keyboards gather dust! As such, the GNU/Linux community has to work within that boundary in order to create a viable desktop system.
Despite the inferiority of the kernel, Windows is still the king of desktops, both in market share, and in my opinion, quality.
I own two computers. One was bought in 1999, has an AMD K6/II processor clocked at 500 MHz with 128 MB SD-RAM. The other was bought in January 2003, has a Pentium IV processor clocked at 2.8 GHz with 512 MB DDR-RAM. Both of these systems dual boot. The former runs Slackware 9.0 and Windows 2000 Pro. The latter also runs Slackware 9.0, alongside Windows XP Home. Both Slackware systems run KDE 3.1, though on the faster system, I sometimes run Gnome 2.2.
On both of these vastly different boxes, Windows consistently outperforms Linux on the desktop. I am not referring to stability, graphics capability, or even the actual speed of the OS, I am referring to the interface.
Windows, which in both cases we know to be bloated, feels more responsive, more consistent, somehow, more established and professional. The environment provides basic tools and games (notepad, pinball etc), from which you build up the system using other software. It remains somewhat utilitarian, while the "control panel" is logically laid out. One does not run into the annoying problem of changing permissions or logging into root whenever one needs to install software.
KDE in the other hand, which is perhaps the most viable of Linux desktops in my opinion (thats not to say Gnome isn't a great interface, but KDE takes the cake), reflects some of the principles behind the success of Linux thus far among the hacker community (Please note, I use the word hacker to mean software developer, not the destructive and irritating nuisance of those who seek to harm other systems - people who should be properly termed "crackers"). Those principles are of choice, and of leaving the control over specifics of the system in the hands of the user.
While this is fantastic for those who are capable or willing to use that power, it is beyond the abilities of most desktop users, for whom the computer is little more than an appliance or a means to an end. As a result, despite the obvious effort of KDE to encourage consistency and develop its own applications, it is a confusing, incoherent and bloated environment. It feels, at least to me, as though certain things, for example the control centre, and various pop-up menus, get in the way of doing what I want to do. For someone who is very computer literate, yet an author, it seems inefficient, whereas in Windows, the control panels and windows seem tucked away until you need them at the right time.
When using KDE, I feel like I am constantly wrestling with the interface whereas in Windows, I am better able to concentrate on using the application - the interface is working with me.
Another issue is one of speed. On my slower box, Windows 2000 runs at an acceptably quick pace, while KDE is extraordinarily sluggish. That is also the case on my faster machine, the same version of KDE responds noticeably slower than a newer version of Windows. Windows and applications load faster, implimentation of changes occur faster, and even the pop-up menus on mouse clicks load much faster.
Another issue, one that is perhaps more easily solved (though there should be a way to automate it, if there is, I am unaware of it), is that of permissions. As I touched upon earlier, installing an application involves unzipping a *.tar.gz or *.tar.bz2 file, checking the permissions, chmod'ing it to executable if it isn't, su'ing to root (all in a terminal emulator, though I know I can do it graphically) and running the setup program using yet another shell command. A user should be able to install programs for himself, in his "home" area, that would still function, and not require root authorisation, yet would not be system wide thus not compromising the security of the system. This means that a user's area would have to be an isolated area where security for that one area is determined by root and himself. Its the same balancing act that admins on a Windows network juggle with as a matter of course.
I know that there are various package management systems that simplify the process, but these only work, or work best on certain distributions making the process something of a minefield. Indeed, I am aware, if not entirely understanding, of problems with the major RPM system. As such, major open source projects as well as companies providing products for Linux distribute their programs in a form that will run on all systems, bypassing the package management issue.
Compiling programs from source is a similar problem. This is something that Windows users rarely if ever do, which is a great shame for them as it is a great way of increasing performance. In Linux, it requires a series of commandline inputs, not to mention the plethora of potential dependency problems one can encounter. I would like there to be a way to compile and install such programs automatically, graphically, with one or two clicks.
Dependencies are a problem that is solved quite simply when a program, especially a large one, has many of its dependencies distributed with it, the setup process detecting which of these it needs to install. As such, one only requires basic libraries and a relatively recent kernel. This would go a very long way to simplifying the installation process, and as most users only use a relatively few number of large applications (office suites, music/movie players, games, web browsers etc), then a larger download would be an acceptable price to pay, as is the case with Windows apps.
A problem, partially solved by the "checkinstall" program for Slackware is the inability to manage compiled programs and those installed by other means than the package management system, after they are installed. Removing them means yet a manual process.
I would like a system where packages installed with a package management system, from external builds and from source are installed and removed within the same window, all graphically.
Let me make it clear that I do not hate KDE. For what it is, it is a good interface and a credit to those who develop it, but it is being held back by numerous factors, many of whom I have not covered. These are flaws that it is possible to correct. They do nothing to detract from its great potential.
There is one very promising solution for the GNU/Linux desktop. That takes the form of Athene. This is an environment that is literally built at run-time from editable XML scripts. To the best of my understanding as someone who does not develop software, it is a scripted system not relying on binaries. This has the advantage of actively encouraging open-source techniques, as the software itself is at its maximum potential when the user has the option to edit the software itself! As such, any Microsoft equivalent, if it was not open source, would have to have additional safeguards against the user being able to read and edit the scripts, that would unnecessarily consume resources. Even so, this would most likely be insufficient to "stop the tide" ;-).
Onto the interface itself. At the time of writing, it is skeletal at best, with few applications, and none to my knowledge that have not been developed by the parent company, Rocklyte.
Nonetheless, it has many of the best aspects of Windows outlined above. It is also fast - blisteringly fast. It is highly configurable, yet presents a simple and discrete method of configuration to the user. You are even presented with a choice of interfaces when you start up, each with different charms, including one that aims to emulate Windows 2000. Developers are of course, only the ones who will get their hands dirty in the source, so the best part of the configurability is somewhat hidden from the lay user. More information is on the website: http://www.rocklyte.com/athene/
The free version runs on top of the Linux kernel, and can also run atop Windows, though I haven't tested this as of yet. As such, it is possible to create a small, efficient system consisting of the superior Linux kernel, the numerous tools and libraries required for various software, and Athene in userland.
The Athene concept needs to be worked on further, both in that project itself and in clones. This could well develop into a killer application for GNU/Linux. In the immediate future, some way of integrating it with the Wine project would help greatly in bridging the gap between Windows and GNU/Linux. In my ignorance, I believe it would be good to have it as a part of the kernel, either having the Wine API compiled in or as a module, or maybe integrated into an Athene-like project, but common sense dictates that this would require too much work, and if it was possible it would have been done before. Nonetheless, it is an important project.
I hope this will help in some way to make GNU/Linux a more powerful and popular alternative to Windows, after all, as more users switch, the open source community will benefit from more developers, ideas, and the satisfaction of creating software that more and more people use.
Ben J. Elijah
As an afterthought, another vital approach in bringing Windows users across to Linux will be to appeal to the gamers community. The abilities of the kernel, translated into graphical capabilities can be exploited if more games are developed. To kick off this process, we need a good open source game, preferably a first-person shooter. There are several engines available, many based upon Quake, a popular FPS. A full GPL'd game, with maps, players, bots, weapons, cheats, great gameplay and of course, a multiplayer mode, will help immeasurably in bringing people across. Integrated with Athene, it could well bring a flood of Windows users across to the GNU/Linux camp.
"
EDIT: Got rid of annoying quoted blue text.

This is a draft of an open letter I'm going to post on various Linux forums but since there are lots of opinions here, I'll treat you lot to a sneak preview!

Its basically about the current state of the linux desktop, and my personal views on it all.
Incidentally, after using Windows XP and hating it, switching back to Slackware 9.0 and writing this, my own arguments have convinced me to install Windows 2000 on my fast machine. Its fantastic, I never liked XP because it was too childlike, WinXP is probably the most balanced OS I have ever used. Best of all, I dont have to constantly fiddle with it, as is the temptation with Linux (I cant help myself.. give me buttons to press...).
Anyways, enjoy! By the way, Asher will either agree wholeheartedly, or disagree big time. Or both

"
An open letter to the GNU/Linux community
We are witnessing today a struggle for dominance in the operating system market. For many years, Microsoft's monopoly was near absolute and now, it is being threatened by the Open Source community. Closed proprietery ownership is being undermined by freedom of source and use. Power is being shifted from the halls of control at the headquarters of a distant and indifferent multinational corporation back to the fingertips of the users.
Most within this community, and indeed many objective observers argue that Linux is a superior kernel to Windows. From reading reviews, tests, and from personal experience, I can say that Linux is faster, more secure, more stable, more adaptable and portable, and more open than Microsoft's offering.
However, these are characteristics that are most essential in areas such as servers, clusters, and other security conscious and networking applications. The new frontier for Linux is the desktop.
This is a far different cup of tea to the servers. Desktop users want ease of use, they want speed, they want the ability to run their applications. They are not interested so much in stability or security, a factor which, as we have seen recently, has backfired upon them. The superiority of one kernel over another is largely irrelevant to them - as they only care about what they can see and do with what they can see.
This cannot be changed. There will be no shift in the desktop market back to using the commandline, or gaining power at the expense of useability. Mice are being worn out at a faster and faster rate, while keyboards gather dust! As such, the GNU/Linux community has to work within that boundary in order to create a viable desktop system.
Despite the inferiority of the kernel, Windows is still the king of desktops, both in market share, and in my opinion, quality.
I own two computers. One was bought in 1999, has an AMD K6/II processor clocked at 500 MHz with 128 MB SD-RAM. The other was bought in January 2003, has a Pentium IV processor clocked at 2.8 GHz with 512 MB DDR-RAM. Both of these systems dual boot. The former runs Slackware 9.0 and Windows 2000 Pro. The latter also runs Slackware 9.0, alongside Windows XP Home. Both Slackware systems run KDE 3.1, though on the faster system, I sometimes run Gnome 2.2.
On both of these vastly different boxes, Windows consistently outperforms Linux on the desktop. I am not referring to stability, graphics capability, or even the actual speed of the OS, I am referring to the interface.
Windows, which in both cases we know to be bloated, feels more responsive, more consistent, somehow, more established and professional. The environment provides basic tools and games (notepad, pinball etc), from which you build up the system using other software. It remains somewhat utilitarian, while the "control panel" is logically laid out. One does not run into the annoying problem of changing permissions or logging into root whenever one needs to install software.
KDE in the other hand, which is perhaps the most viable of Linux desktops in my opinion (thats not to say Gnome isn't a great interface, but KDE takes the cake), reflects some of the principles behind the success of Linux thus far among the hacker community (Please note, I use the word hacker to mean software developer, not the destructive and irritating nuisance of those who seek to harm other systems - people who should be properly termed "crackers"). Those principles are of choice, and of leaving the control over specifics of the system in the hands of the user.
While this is fantastic for those who are capable or willing to use that power, it is beyond the abilities of most desktop users, for whom the computer is little more than an appliance or a means to an end. As a result, despite the obvious effort of KDE to encourage consistency and develop its own applications, it is a confusing, incoherent and bloated environment. It feels, at least to me, as though certain things, for example the control centre, and various pop-up menus, get in the way of doing what I want to do. For someone who is very computer literate, yet an author, it seems inefficient, whereas in Windows, the control panels and windows seem tucked away until you need them at the right time.
When using KDE, I feel like I am constantly wrestling with the interface whereas in Windows, I am better able to concentrate on using the application - the interface is working with me.
Another issue is one of speed. On my slower box, Windows 2000 runs at an acceptably quick pace, while KDE is extraordinarily sluggish. That is also the case on my faster machine, the same version of KDE responds noticeably slower than a newer version of Windows. Windows and applications load faster, implimentation of changes occur faster, and even the pop-up menus on mouse clicks load much faster.
Another issue, one that is perhaps more easily solved (though there should be a way to automate it, if there is, I am unaware of it), is that of permissions. As I touched upon earlier, installing an application involves unzipping a *.tar.gz or *.tar.bz2 file, checking the permissions, chmod'ing it to executable if it isn't, su'ing to root (all in a terminal emulator, though I know I can do it graphically) and running the setup program using yet another shell command. A user should be able to install programs for himself, in his "home" area, that would still function, and not require root authorisation, yet would not be system wide thus not compromising the security of the system. This means that a user's area would have to be an isolated area where security for that one area is determined by root and himself. Its the same balancing act that admins on a Windows network juggle with as a matter of course.
I know that there are various package management systems that simplify the process, but these only work, or work best on certain distributions making the process something of a minefield. Indeed, I am aware, if not entirely understanding, of problems with the major RPM system. As such, major open source projects as well as companies providing products for Linux distribute their programs in a form that will run on all systems, bypassing the package management issue.
Compiling programs from source is a similar problem. This is something that Windows users rarely if ever do, which is a great shame for them as it is a great way of increasing performance. In Linux, it requires a series of commandline inputs, not to mention the plethora of potential dependency problems one can encounter. I would like there to be a way to compile and install such programs automatically, graphically, with one or two clicks.
Dependencies are a problem that is solved quite simply when a program, especially a large one, has many of its dependencies distributed with it, the setup process detecting which of these it needs to install. As such, one only requires basic libraries and a relatively recent kernel. This would go a very long way to simplifying the installation process, and as most users only use a relatively few number of large applications (office suites, music/movie players, games, web browsers etc), then a larger download would be an acceptable price to pay, as is the case with Windows apps.
A problem, partially solved by the "checkinstall" program for Slackware is the inability to manage compiled programs and those installed by other means than the package management system, after they are installed. Removing them means yet a manual process.
I would like a system where packages installed with a package management system, from external builds and from source are installed and removed within the same window, all graphically.
Let me make it clear that I do not hate KDE. For what it is, it is a good interface and a credit to those who develop it, but it is being held back by numerous factors, many of whom I have not covered. These are flaws that it is possible to correct. They do nothing to detract from its great potential.
There is one very promising solution for the GNU/Linux desktop. That takes the form of Athene. This is an environment that is literally built at run-time from editable XML scripts. To the best of my understanding as someone who does not develop software, it is a scripted system not relying on binaries. This has the advantage of actively encouraging open-source techniques, as the software itself is at its maximum potential when the user has the option to edit the software itself! As such, any Microsoft equivalent, if it was not open source, would have to have additional safeguards against the user being able to read and edit the scripts, that would unnecessarily consume resources. Even so, this would most likely be insufficient to "stop the tide" ;-).
Onto the interface itself. At the time of writing, it is skeletal at best, with few applications, and none to my knowledge that have not been developed by the parent company, Rocklyte.
Nonetheless, it has many of the best aspects of Windows outlined above. It is also fast - blisteringly fast. It is highly configurable, yet presents a simple and discrete method of configuration to the user. You are even presented with a choice of interfaces when you start up, each with different charms, including one that aims to emulate Windows 2000. Developers are of course, only the ones who will get their hands dirty in the source, so the best part of the configurability is somewhat hidden from the lay user. More information is on the website: http://www.rocklyte.com/athene/
The free version runs on top of the Linux kernel, and can also run atop Windows, though I haven't tested this as of yet. As such, it is possible to create a small, efficient system consisting of the superior Linux kernel, the numerous tools and libraries required for various software, and Athene in userland.
The Athene concept needs to be worked on further, both in that project itself and in clones. This could well develop into a killer application for GNU/Linux. In the immediate future, some way of integrating it with the Wine project would help greatly in bridging the gap between Windows and GNU/Linux. In my ignorance, I believe it would be good to have it as a part of the kernel, either having the Wine API compiled in or as a module, or maybe integrated into an Athene-like project, but common sense dictates that this would require too much work, and if it was possible it would have been done before. Nonetheless, it is an important project.
I hope this will help in some way to make GNU/Linux a more powerful and popular alternative to Windows, after all, as more users switch, the open source community will benefit from more developers, ideas, and the satisfaction of creating software that more and more people use.
Ben J. Elijah
As an afterthought, another vital approach in bringing Windows users across to Linux will be to appeal to the gamers community. The abilities of the kernel, translated into graphical capabilities can be exploited if more games are developed. To kick off this process, we need a good open source game, preferably a first-person shooter. There are several engines available, many based upon Quake, a popular FPS. A full GPL'd game, with maps, players, bots, weapons, cheats, great gameplay and of course, a multiplayer mode, will help immeasurably in bringing people across. Integrated with Athene, it could well bring a flood of Windows users across to the GNU/Linux camp.
"
EDIT: Got rid of annoying quoted blue text.
Comment