GNOME and KDE? XFree 4.0? Kernel 2.4? Compilers? TK, TCL, Perl, etc? I have a hard time believing that. Of course, if you're using the thing bare-bones, this stuff isn't necessary, but when I say BeOS takes ~100MB, I'm talking full install. Compilers, media players, IDEs, etc (perl, and all the GNU tools). All the stuff I mentioned is absolutely necessary for most Linux distros. Take, for example GCC. You can't really get by without a compiler. Say you have XFree 4.0.1. NVIDIA's drivers don't yet have an RPM for that, so you've got to compile it. Upgrading the kernel usually necessitates a recompile (in order to take full advantage of it.)
I'm talking about a usable DESKTOP Linux distribution. Even Slackware, which is pretty sevelte, takes up 150-200MB base install, plus 100-something for X, and another 100 something for GNOME and KDE. This includes compilers and stuff like Perl, TK, etc that a Linux distro is more or less unusable (as a desktop) without. I can tell you exactly where all the bloat is.
A) Libraries. There are tons of libraries on a Linux system.
B) X. X takes up around 70MB on my machine. The BeOS GUI is inside the 3.1MB app server.
C) Gnome and KDE. The BeOS WM is also inside the 3.1MB app server, and the "DE" is spread out through different servers.
You can set up a fully usable BeOS machine in under 50MB. (Graphical, everything.) Try that with Linux.
Exactly why I can play more than a dozen MP3s (14 or 15 on my PII 300) on BeOS (which btw has a two year old version of GCC) while Linux chokes on a measly half-dozen? Also why I can go along not even noticing that some wayard process is chewing up 100% of my CPU, while something like a big compile or untar job makes GNOME noticablly choppier?
I only had two points. I'm not talking about monolithic in the technical sense (I rarely talk about anything in the strict technical sense) but in the end-result sense. One often has to recompile the kernel in order to insert new hardware. This is because there is a close association with the hardware and the drivers. Thus, it is monolithic. The fact that a kernel upgrade requires me to recompile doesn't surprise me, the fact that I have to recompile the modules as well does. This implies way to close of a link between kernel and driver. The vast-majority of the cases may be true for trival hardware, but is decidely not true for major items like video-cards and sound cards.
I never said monolithic is bad. I said that since QNX is not monolithic, ALSA doesn't have to be part of the kernel to work, and thus the kernel need not be Open Source.
Really, Linux is monolithic. Sure it has kernel modules, but they can't be reliably used between kernel versions. So often times, upgrading the kernel or changing the hardware (in case you haven't compiled that driver for your kernel before) usually requires a kernel recompile.
You really do know what the appropriate title for this article should have been, don't you? Sonique to become multi-platform. I'm sure it wouldn't have been accepeted had it read like this...
Sonique to come to BeOS:
.......
....
.....
......
.....
.....
.....
Oh, yea, and it will run on Linux and MacOS too.
1) Several older Linux 3D programs are statically linked. And what do you mean "most" modern 3D hardware. You've got the NVIDIA cards, the Matrox cards, the ATI cards (not the Radeon, the important one), and the 3DFX cards. What about S3, Intel, NeoMagic, etc? And "not to the same level of quality" is severly candy-coating it. Even the best Linux3D drivers (NVIDIAs) are still significantly slower than their Windows counterparts.
2) As I recall, OpenAL isn't complete yet. And it is a middle-level API. The lower-level drivers don't seem to exist yet. Also, Creative is listed as a member company, but doesn't actually have any drivers yet.
3) I'm arguing that USB isn't a big issue because the installed base of USB devices is pretty small. Winmodem support is a big issue because the installed base of winmodems is quite large. Installed base is king, as is market momentum. Right now, PS/2 and Winmodems have both. USB and regular modems do not.
Re:It's not the broken mods that piss people off..
on
New Q3A Patch And Mods
·
· Score: 2
I meant "shouldn't". And for all those brain-dead moderators who don't get the gist of my post (Subtlty 101 should be a require class) I mean that football players work hard, sysadmins work hard. They both bring stuff to humanity. Entertainment is just as important as getting work done, plus football players bring tons of money into the economy. (Ecomonics 101 should also be a required course.) As such, they both deserve what they get.
Re:It's not the broken mods that piss people off..
on
New Q3A Patch And Mods
·
· Score: 2
Sysadmins should be paid as much as they are because all they basically do is cater to an industry that purpously makes tools hard to maintain. Not so pretty when the shoe is on the other foot isn't it?
Thank you. I do agree that Linux does seem to be heading in the right direction, and I'm sorry for being inflamatory myself. However, it just bothers me how many people are blinded by Linux's success and don't see the long road still left to travel.
Is it an operating system that shelters the user from the underlying system? Is it an OS that does
not burden the user with any learning? Is it an OS that has pretty little buttons and widgets
grafted and pasted into the inner workings of the system.
>>>>>>>>
Stupid thinking. Take a good long look at MacOS X or BeOS. You'll see that it presents the whole power of the system to you without REQUIRING you to be hacker to use it. Things like organized, simple directory structure. Sane, simple package management. Well planned configuration files (MacOS X's XML idea kicks ass), a consistant environemnt. All of these go a long way to make an OS "good." (BTW: Eliteist bastard. If an OS FORCES learning upon the user, than its useless as a desktop OS. Not everyone cares about computers, and if you want Linux to succeed in the desktop market, you have to understand users who need to do advanced work, but couldn't care less how the computer works.)
I for one, believe that Unix (in general) can be both a desktop, and a server... it is adaptable. The
modular design is what makes it that way. Things like KDE and Gnome are more like extra
layers on the cake, but what do you expect? That everything should come in a perfect little
package, in one homogenious mixture... a cookie cutter OS?
>>>>>>
Integration. That's all I need to say. There is nothing wrong with being moduler and adaptable as long as the finished product LOOKS smooth and polished. You know, they used to whitewash castles to look like they're made of one big piece of rock, even though they're made up of seperate blocks. Same concept. Take a look at the average car (not American). The things are very moduler, engines can be taken from one, replaced with another. Transmissions can be swapped, etc. The whole system is very adaptable, yet looks like its polished, consistant product.
The purpose of these linux
distributions is to bring everything together. Sure, I could build a linux system from scratch, but it
is much easier to download or get a CD and run a quick install.
>>>>>>>
But they don't. The distros are still terribly disorganized.
This is purely because of user stupidity. Plain and simple. The common user is both fearful and
stupid. (I really mean that.) People will not change because they are afraid of it, because they hate
learning. The number one question is "what if xx will not run under linux?". But this is no excuse
to write everything off and say that everyone should stick with windows. Windows is pure crap...
>>>>>>>>
I've been using UNIX-like systems for years. I still can' stand them. Its not user stupidity, but the fact that half of the population doesn't give a damn how the system works. Linux people don't think in terms of commercial product, they think in terms of CS project. If Linux is to succeed as a desktop OS, it has to cater to the same people as Win98. Half those users just don't want to learn how the blasted thing works. To them it is a tool, like a hammer. I'm sure they don't force you to learn how a chainsaw works when you cut down a tree, do they?
I'm not a newbie Linux user. I've compiled my fair share programs. (X, the kernel, dozens of utilities, KDE, Qt, KDevelop, etc) I think there IS a problem. Nobody ever reads the documentation, they want to click on something to compile it, click to install, and click to uninstall. (Or type a command instead of click.) Also, there is no GUI for making packages. The ability to compile software is potentially on of the greatest advatnages of Linux, and it is wasted on those who understand make. It wouldn't be hard to write a program that would put a GUI to make, but getting people to standardize on one system (make config/ make / make install perferably) would be impossible. I never said anything about a registry. WTF are you talking about? I have no problem with the current make, as long as there is a GUI for it, and all software uses the same system.
Actually, push is pretty usefull. Of course, I have it turned off myself (because of the speed hit) but some people want it, and implemented correctly, it could be pretty nifty.
Alright, you've said your peace, and I think you're a fool, so I will now say mine.
>>>>>
Fair enough. I think you're a fool too;)
Because there's more to life than shrink-wrapped hundred dollar office suites. AbiWord, Gnumeric, KOffice, are all coming. Additionally, StarOffice is
being drastically reworked into something that doesn't suck.:)
>>>>>
Don't ever talk to me about what WILL happen. People are touting Linux as a competitor to Windows NOW, not in the future. Right now, there is nothing comparable to Office or Wordperfect Office on Linux. Believe it or not, some people actually USE the extended features in Office. I for one, could never live without Publisher. If you need to write more than just plain documents, then you need the power of "a hundred dollar shrink wrapped office suite." Right now, WP Office 2K is the only (usable) one on Linux and its better in Windows.
Meet the tip of the iceberg. What, did you expect Linux games to multiply overnight? These things take time.
>>>>
Like I said, people are hyping it NOW!
Once you get into the system itself, you're doing more than 80% of the Windows userbase will ever do. Let's draw some parallels. If/etc is essentially very
similar to the Windows registry and.ini files, how many end users will ever touch their registry, or even realize a file called win.ini exists on their system?
>>>>>>>>>
Lot's of people know how to registry hack and edit.ini files. You fail to remember that the majority of the work-force grew up in DOS. However, look at it this way. How many desktop users install updates to their system via ActiveUpdate? How many install service packs (in essence recompiling your kernel, updating X, updating GNOME and KDE, and your libraries.) How many people install new drivers for their hardware? All this stuff is just too damn hard in Linux. For example, there is no standard way to install a driver. Some, you can just add a line in modules.rc. Other, you've got to edit modules.conf. Still others need to be compiled and installed via installers. Installing the NVIDIA drivers (which, sadly is indicative of how the bulk of Linux installs are like) requires opening up a terminal, compiling a program, and manually replacing files. Only a few applications "get it" in Linux. Helix GNOME is one such application. However, if one app does something well, you can't just point to it and say "oh, we're easy too!"
If you believe/etc can be condensed into a dozen files and retain all their information and ease of access (An all-encompasing GUI is not easier to access
than opening a text file.) then you either aren't familiar with the Unix way of thinking, or you aren't interested in retaining the configurability and
flexibility that Unix offers.
>>>>>>>>>>>>>>>
I'm quite familiar with "the UNIX way." I started with Slackware 3.5, moved to Redhat, then Mandrake, then Slackware, then finally FreeBSD. I use BeOS most of the time, which is quite UNIX-ish in the CLI. I still don't like it. The UNIX way of thinking is not appropriate for a desktop computer. People like organization, and aren't willing to put up with the additional learning curve that the UNIX directory structure presents. For example, what is the difference between/usr/bin and/usr/local/bin on a single-user machine? I'm not suggesting that all UNIXes change. I'm just saying that the "UNIX way" isn't going to cut it for the average desktop.
Of course, if Linux is to be this grand desktop OS that people want to be, I fear it may lose that anyway. And that would indeed be nothing less than a tragic loss. (Which is why I don't think Linux in its current state should be a desktop OS. At least, not the way you seem to think a desktop OS works)
>>>>>>>
Duh, WTF are you argueing with me if you agree with me!
I'm going to be as brief as possible. First, have you ever even read about how Debian works as opposed to RedHat? There are distros besides RedHat, you
know.
>>>>
I know. I've used Slackware, Mandrake, RedHat, and SuSE. (I'll never run Debian since they aren't cutting edge) They all work more or less the same way. The NVIDIA drivers install the same way in all of them, ALSA installs the same way, the kernel compiles the same way.
Second, comparing BeOS to Linux doesn't work. BeOS is the project of a single development group lead by a single program management group. How can you seriously expect Linux to have the same goals?
>>>>>>
Let's be clear here. I'm not talking about Linux in general. That would be stupid. I'm talking about Linux as a competitor to Windows on the desktop. That's a very specific genere, and to succeed in that area, the goals better be pretty damn similar.
Now, if you want to compare a distro to BeOS, that's fine.
>>>>>
Why do people think the distros are so damn different? Do they all only run the software that comes with their distro? All the Linuxes work more or less the same way.
Third, personal telnet server? Are you
completely disconnected from reality? SSH, my friend. Telnet has no place in the hands of end users that don't know anything about security, nor should
they be expected to.
>>>>>>>
Good god. I'm talking about a specific way of configuring something, not the technical merits of the configuration. Fine, if you're so inclined, SSH should be configured with a check-box and a password field. (Which would be useless as the only account on BeOS is root.)
Fourth, your suggestion that Linux and Windows should be held to the same standard is a repeat of your faulty reasoning behind comparing it to BeOS. See above.
>>>>>>
The faulty reasoning is on your part. If you say that Linux is going to be a competitor with Windows (which everybody including Torvalds and the poster I'm replying to is) then it damn well better be held to the same standard.
Interesting, you've been judging and condemning it like one. And throughout this paragraph you expect it to be one. "Consistancy is a good thing for an OS"
is irrelevant if Linux isn't an OS. "managerial problems an OS has" is meaningless, since Linux isn't an OS.
>>>>>>
When I said "Linux isn't an OS" I was pointing out a problem. You seem to misunderstand me. I have nothing against Linux. I think it is a fine system if it suits you. However, I think it is useless as a destkop OS (at the moment, I'm sure it will get better.) Linux ISN'T an OS. In order to compete in the desktop market, it HAS to be.
Now, to point out why none of this matters. Linux isn't a product. It's a kernel. Linux distributions are products, and I'm sick of people comparing this
mythical thing that is Linux to operating systems. If you want to bitch about RedHat, call it RedHat Linux. Don't give the other distros crap for the
product of one company. Further. Don't expect the contents of any distro to be perfect, flowing, and totally consistent until a company shows up that
writes every single application in-house. Then, and only then, does it fit into your world of what an OS is. Then, and only then, may you compare it to
Windows and BeOS.
>>>>>>>>
Again, you mistake me. I'm not talking about Linux in general, but Linux in the context of a Windows-competitor. In that case, the Linux refers to the distros implicitly. It doesn't matter how many different people are chugging along behind the scenes, if the end result is inconsistant and hard to use, then that's a problem for the desktop market. The point is that nobody (statistically) in the desktop market cares about Linux as a movement, but Linux as an OS. All the "managerial stuff" I refer too hinders Linux. If the distro makers can't get such a disparate group of coders to unify, then maybe Linux wasn't meant to be a mass market, desktop OS.
1) UNIX's beauty is in the eye of the beholder. I personally don't see a huge mass of code to be terribly beautiful. I personally dig streamlined, slick, simple systems. SysV (and UNIX in general) is far from that.
Those twelve config files would not necessarily be ugly. It would just take actual THINKING on the part of the distro maker.
However, Fred is an odd entitiy. Very few people actually HAVE a USB zip drive, a USB mouse, and a USB scanner. Take a look at all the commercial computer systems (Compaq, Dell, HP). How many of them ship with USB hardware?
1) Linux 3D acceleration is still weak. Software support is iffy (those stupid statically compiled apps) its still slower than Windows, and the fastest implementation (NVIDIA's) isn't officially supported and isn't quite perfectly stable. The state of 3D on Linux is still in Microsoft-beta quality.
2) 3D sound may be supported by various apps, but then they're writing their own 3D sound engine (unlikely.) I'm talking about the DirectSound3D or Aureal3D hardware everyone has, but Linux can't use.
3) I put the winmodem arguement, because winmodems prevent more people from using Linux than USB devices. Very few people have critical USB devices, but millions have Winmodems. I don't care if its supported to the maximum extent of the patant law, the fact remains that winmodem support is a bigger issue than USB support (which is what the author of the article I'm responding to said.)
4) Tell that to all the people who've bought force-feedback joysticks for $200. They're pretty cool technology, and they are an entire class of devices unusable on Linux. For every USB device, there are dozens of PS/2/parallel/serial equivilents. However, force-feedback devices cannot be used at ALL in Linux.
Slashdot Mirror
GNOME and KDE? XFree 4.0? Kernel 2.4? Compilers? TK, TCL, Perl, etc? I have a hard time believing that. Of course, if you're using the thing bare-bones, this stuff isn't necessary, but when I say BeOS takes ~100MB, I'm talking full install. Compilers, media players, IDEs, etc (perl, and all the GNU tools). All the stuff I mentioned is absolutely necessary for most Linux distros. Take, for example GCC. You can't really get by without a compiler. Say you have XFree 4.0.1. NVIDIA's drivers don't yet have an RPM for that, so you've got to compile it. Upgrading the kernel usually necessitates a recompile (in order to take full advantage of it.)
I'm talking about a usable DESKTOP Linux distribution. Even Slackware, which is pretty sevelte, takes up 150-200MB base install, plus 100-something for X, and another 100 something for GNOME and KDE. This includes compilers and stuff like Perl, TK, etc that a Linux distro is more or less unusable (as a desktop) without. I can tell you exactly where all the bloat is.
A) Libraries. There are tons of libraries on a Linux system.
B) X. X takes up around 70MB on my machine. The BeOS GUI is inside the 3.1MB app server.
C) Gnome and KDE. The BeOS WM is also inside the 3.1MB app server, and the "DE" is spread out through different servers.
You can set up a fully usable BeOS machine in under 50MB. (Graphical, everything.) Try that with Linux.
Exactly why I can play more than a dozen MP3s (14 or 15 on my PII 300) on BeOS (which btw has a two year old version of GCC) while Linux chokes on a measly half-dozen? Also why I can go along not even noticing that some wayard process is chewing up 100% of my CPU, while something like a big compile or untar job makes GNOME noticablly choppier?
I only had two points. I'm not talking about monolithic in the technical sense (I rarely talk about anything in the strict technical sense) but in the end-result sense. One often has to recompile the kernel in order to insert new hardware. This is because there is a close association with the hardware and the drivers. Thus, it is monolithic. The fact that a kernel upgrade requires me to recompile doesn't surprise me, the fact that I have to recompile the modules as well does. This implies way to close of a link between kernel and driver. The vast-majority of the cases may be true for trival hardware, but is decidely not true for major items like video-cards and sound cards.
The beta was released only a few weeks before this one.
I never said monolithic is bad. I said that since QNX is not monolithic, ALSA doesn't have to be part of the kernel to work, and thus the kernel need not be Open Source.
BeZealots!=Linux Zealots. They're just as passionate, but realize when that their OS isn't perfect for everyone.
Really, Linux is monolithic. Sure it has kernel modules, but they can't be reliably used between kernel versions. So often times, upgrading the kernel or changing the hardware (in case you haven't compiled that driver for your kernel before) usually requires a kernel recompile.
Unlike some OSs, QNX is not monolithic. ALSA is just a module somewhere above kernel space, so its not actually part of the OS proper.
God dammit! Why the fuck do we need ANOTHER OpenSource UNIX clone when *BSD is already available!
You really do know what the appropriate title for this article should have been, don't you? Sonique to become multi-platform. I'm sure it wouldn't have been accepeted had it read like this...
Sonique to come to BeOS:
.......
....
.....
......
.....
.....
.....
Oh, yea, and it will run on Linux and MacOS too.
1) Several older Linux 3D programs are statically linked. And what do you mean "most" modern 3D hardware. You've got the NVIDIA cards, the Matrox cards, the ATI cards (not the Radeon, the important one), and the 3DFX cards. What about S3, Intel, NeoMagic, etc? And "not to the same level of quality" is severly candy-coating it. Even the best Linux3D drivers (NVIDIAs) are still significantly slower than their Windows counterparts.
2) As I recall, OpenAL isn't complete yet. And it is a middle-level API. The lower-level drivers don't seem to exist yet. Also, Creative is listed as a member company, but doesn't actually have any drivers yet.
3) I'm arguing that USB isn't a big issue because the installed base of USB devices is pretty small. Winmodem support is a big issue because the installed base of winmodems is quite large. Installed base is king, as is market momentum. Right now, PS/2 and Winmodems have both. USB and regular modems do not.
I meant "shouldn't". And for all those brain-dead moderators who don't get the gist of my post (Subtlty 101 should be a require class) I mean that football players work hard, sysadmins work hard. They both bring stuff to humanity. Entertainment is just as important as getting work done, plus football players bring tons of money into the economy. (Ecomonics 101 should also be a required course.) As such, they both deserve what they get.
Sysadmins should be paid as much as they are because all they basically do is cater to an industry that purpously makes tools hard to maintain. Not so pretty when the shoe is on the other foot isn't it?
Thank you. I do agree that Linux does seem to be heading in the right direction, and I'm sorry for being inflamatory myself. However, it just bothers me how many people are blinded by Linux's success and don't see the long road still left to travel.
Howver, MS did create Outlook and they have the right to do whatever they bloody well want with it.
That's exactly his point. If Outlook does something that needs Exchange, you can't complain about it not working if you're using TradeServer.
Is it an operating system that shelters the user from the underlying system? Is it an OS that does
not burden the user with any learning? Is it an OS that has pretty little buttons and widgets
grafted and pasted into the inner workings of the system.
>>>>>>>>
Stupid thinking. Take a good long look at MacOS X or BeOS. You'll see that it presents the whole power of the system to you without REQUIRING you to be hacker to use it. Things like organized, simple directory structure. Sane, simple package management. Well planned configuration files (MacOS X's XML idea kicks ass), a consistant environemnt. All of these go a long way to make an OS "good." (BTW: Eliteist bastard. If an OS FORCES learning upon the user, than its useless as a desktop OS. Not everyone cares about computers, and if you want Linux to succeed in the desktop market, you have to understand users who need to do advanced work, but couldn't care less how the computer works.)
I for one, believe that Unix (in general) can be both a desktop, and a server... it is adaptable. The
modular design is what makes it that way. Things like KDE and Gnome are more like extra
layers on the cake, but what do you expect? That everything should come in a perfect little
package, in one homogenious mixture... a cookie cutter OS?
>>>>>>
Integration. That's all I need to say. There is nothing wrong with being moduler and adaptable as long as the finished product LOOKS smooth and polished. You know, they used to whitewash castles to look like they're made of one big piece of rock, even though they're made up of seperate blocks. Same concept. Take a look at the average car (not American). The things are very moduler, engines can be taken from one, replaced with another. Transmissions can be swapped, etc. The whole system is very adaptable, yet looks like its polished, consistant product.
The purpose of these linux
distributions is to bring everything together. Sure, I could build a linux system from scratch, but it
is much easier to download or get a CD and run a quick install.
>>>>>>>
But they don't. The distros are still terribly disorganized.
This is purely because of user stupidity. Plain and simple. The common user is both fearful and
stupid. (I really mean that.) People will not change because they are afraid of it, because they hate
learning. The number one question is "what if xx will not run under linux?". But this is no excuse
to write everything off and say that everyone should stick with windows. Windows is pure crap...
>>>>>>>>
I've been using UNIX-like systems for years. I still can' stand them. Its not user stupidity, but the fact that half of the population doesn't give a damn how the system works. Linux people don't think in terms of commercial product, they think in terms of CS project. If Linux is to succeed as a desktop OS, it has to cater to the same people as Win98. Half those users just don't want to learn how the blasted thing works. To them it is a tool, like a hammer. I'm sure they don't force you to learn how a chainsaw works when you cut down a tree, do they?
I'm not a newbie Linux user. I've compiled my fair share programs. (X, the kernel, dozens of utilities, KDE, Qt, KDevelop, etc) I think there IS a problem. Nobody ever reads the documentation, they want to click on something to compile it, click to install, and click to uninstall. (Or type a command instead of click.) Also, there is no GUI for making packages. The ability to compile software is potentially on of the greatest advatnages of Linux, and it is wasted on those who understand make. It wouldn't be hard to write a program that would put a GUI to make, but getting people to standardize on one system (make config/ make / make install perferably) would be impossible. I never said anything about a registry. WTF are you talking about? I have no problem with the current make, as long as there is a GUI for it, and all software uses the same system.
Actually, push is pretty usefull. Of course, I have it turned off myself (because of the speed hit) but some people want it, and implemented correctly, it could be pretty nifty.
Alright, you've said your peace, and I think you're a fool, so I will now say mine. ;)
:)
/etc is essentially very
.ini files, how many end users will ever touch their registry, or even realize a file called win.ini exists on their system?
.ini files. You fail to remember that the majority of the work-force grew up in DOS. However, look at it this way. How many desktop users install updates to their system via ActiveUpdate? How many install service packs (in essence recompiling your kernel, updating X, updating GNOME and KDE, and your libraries.) How many people install new drivers for their hardware? All this stuff is just too damn hard in Linux. For example, there is no standard way to install a driver. Some, you can just add a line in modules.rc. Other, you've got to edit modules.conf. Still others need to be compiled and installed via installers. Installing the NVIDIA drivers (which, sadly is indicative of how the bulk of Linux installs are like) requires opening up a terminal, compiling a program, and manually replacing files. Only a few applications "get it" in Linux. Helix GNOME is one such application. However, if one app does something well, you can't just point to it and say "oh, we're easy too!"
/etc can be condensed into a dozen files and retain all their information and ease of access (An all-encompasing GUI is not easier to access
/usr/bin and /usr/local/bin on a single-user machine? I'm not suggesting that all UNIXes change. I'm just saying that the "UNIX way" isn't going to cut it for the average desktop.
>>>>>
Fair enough. I think you're a fool too
Because there's more to life than shrink-wrapped hundred dollar office suites. AbiWord, Gnumeric, KOffice, are all coming. Additionally, StarOffice is
being drastically reworked into something that doesn't suck.
>>>>>
Don't ever talk to me about what WILL happen. People are touting Linux as a competitor to Windows NOW, not in the future. Right now, there is nothing comparable to Office or Wordperfect Office on Linux. Believe it or not, some people actually USE the extended features in Office. I for one, could never live without Publisher. If you need to write more than just plain documents, then you need the power of "a hundred dollar shrink wrapped office suite." Right now, WP Office 2K is the only (usable) one on Linux and its better in Windows.
Meet the tip of the iceberg. What, did you expect Linux games to multiply overnight? These things take time.
>>>>
Like I said, people are hyping it NOW!
Once you get into the system itself, you're doing more than 80% of the Windows userbase will ever do. Let's draw some parallels. If
similar to the Windows registry and
>>>>>>>>>
Lot's of people know how to registry hack and edit
If you believe
than opening a text file.) then you either aren't familiar with the Unix way of thinking, or you aren't interested in retaining the configurability and
flexibility that Unix offers.
>>>>>>>>>>>>>>>
I'm quite familiar with "the UNIX way." I started with Slackware 3.5, moved to Redhat, then Mandrake, then Slackware, then finally FreeBSD. I use BeOS most of the time, which is quite UNIX-ish in the CLI. I still don't like it. The UNIX way of thinking is not appropriate for a desktop computer. People like organization, and aren't willing to put up with the additional learning curve that the UNIX directory structure presents. For example, what is the difference between
Of course, if Linux is to be this grand desktop OS that people want to be, I fear it may lose that anyway. And that would indeed be nothing less than a tragic loss. (Which is why I don't think Linux in its current state should be a desktop OS. At least, not the way you seem to think a desktop OS works)
>>>>>>>
Duh, WTF are you argueing with me if you agree with me!
I'm going to be as brief as possible. First, have you ever even read about how Debian works as opposed to RedHat? There are distros besides RedHat, you
know.
>>>>
I know. I've used Slackware, Mandrake, RedHat, and SuSE. (I'll never run Debian since they aren't cutting edge) They all work more or less the same way. The NVIDIA drivers install the same way in all of them, ALSA installs the same way, the kernel compiles the same way.
Second, comparing BeOS to Linux doesn't work. BeOS is the project of a single development group lead by a single program management group. How can you seriously expect Linux to have the same goals?
>>>>>>
Let's be clear here. I'm not talking about Linux in general. That would be stupid. I'm talking about Linux as a competitor to Windows on the desktop. That's a very specific genere, and to succeed in that area, the goals better be pretty damn similar.
Now, if you want to compare a distro to BeOS, that's fine.
>>>>>
Why do people think the distros are so damn different? Do they all only run the software that comes with their distro? All the Linuxes work more or less the same way.
Third, personal telnet server? Are you
completely disconnected from reality? SSH, my friend. Telnet has no place in the hands of end users that don't know anything about security, nor should
they be expected to.
>>>>>>>
Good god. I'm talking about a specific way of configuring something, not the technical merits of the configuration. Fine, if you're so inclined, SSH should be configured with a check-box and a password field. (Which would be useless as the only account on BeOS is root.)
Fourth, your suggestion that Linux and Windows should be held to the same standard is a repeat of your faulty reasoning behind comparing it to BeOS. See above.
>>>>>>
The faulty reasoning is on your part. If you say that Linux is going to be a competitor with Windows (which everybody including Torvalds and the poster I'm replying to is) then it damn well better be held to the same standard.
Interesting, you've been judging and condemning it like one. And throughout this paragraph you expect it to be one. "Consistancy is a good thing for an OS"
is irrelevant if Linux isn't an OS. "managerial problems an OS has" is meaningless, since Linux isn't an OS.
>>>>>>
When I said "Linux isn't an OS" I was pointing out a problem. You seem to misunderstand me. I have nothing against Linux. I think it is a fine system if it suits you. However, I think it is useless as a destkop OS (at the moment, I'm sure it will get better.) Linux ISN'T an OS. In order to compete in the desktop market, it HAS to be.
Now, to point out why none of this matters. Linux isn't a product. It's a kernel. Linux distributions are products, and I'm sick of people comparing this
mythical thing that is Linux to operating systems. If you want to bitch about RedHat, call it RedHat Linux. Don't give the other distros crap for the
product of one company. Further. Don't expect the contents of any distro to be perfect, flowing, and totally consistent until a company shows up that
writes every single application in-house. Then, and only then, does it fit into your world of what an OS is. Then, and only then, may you compare it to
Windows and BeOS.
>>>>>>>>
Again, you mistake me. I'm not talking about Linux in general, but Linux in the context of a Windows-competitor. In that case, the Linux refers to the distros implicitly. It doesn't matter how many different people are chugging along behind the scenes, if the end result is inconsistant and hard to use, then that's a problem for the desktop market. The point is that nobody (statistically) in the desktop market cares about Linux as a movement, but Linux as an OS. All the "managerial stuff" I refer too hinders Linux. If the distro makers can't get such a disparate group of coders to unify, then maybe Linux wasn't meant to be a mass market, desktop OS.
1) UNIX's beauty is in the eye of the beholder. I personally don't see a huge mass of code to be terribly beautiful. I personally dig streamlined, slick, simple systems. SysV (and UNIX in general) is far from that.
Those twelve config files would not necessarily be ugly. It would just take actual THINKING on the part of the distro maker.
However, Fred is an odd entitiy. Very few people actually HAVE a USB zip drive, a USB mouse, and a USB scanner. Take a look at all the commercial computer systems (Compaq, Dell, HP). How many of them ship with USB hardware?
1) Linux 3D acceleration is still weak. Software support is iffy (those stupid statically compiled apps) its still slower than Windows, and the fastest implementation (NVIDIA's) isn't officially supported and isn't quite perfectly stable. The state of 3D on Linux is still in Microsoft-beta quality.
2) 3D sound may be supported by various apps, but then they're writing their own 3D sound engine (unlikely.) I'm talking about the DirectSound3D or Aureal3D hardware everyone has, but Linux can't use.
3) I put the winmodem arguement, because winmodems prevent more people from using Linux than USB devices. Very few people have critical USB devices, but millions have Winmodems. I don't care if its supported to the maximum extent of the patant law, the fact remains that winmodem support is a bigger issue than USB support (which is what the author of the article I'm responding to said.)
4) Tell that to all the people who've bought force-feedback joysticks for $200. They're pretty cool technology, and they are an entire class of devices unusable on Linux. For every USB device, there are dozens of PS/2/parallel/serial equivilents. However, force-feedback devices cannot be used at ALL in Linux.
Reality: More people have winmodems than USB devices.