First of all: I'm not a Linux user, but a FreeBSD user, who is very angry at nVidea at the moment. Let me tell you a little story behind the development of our cute little nVidea Riva128 chipset drivers:
1998 computer bought, only drivers available for Windows NT
1999 Drivers released for Windows 95/98, shortly followed by _very_ buggy Linux drivers.
Right now, nVidea stopped their development for the riva128 chipset on Linux (which means it "probably doesn't work"). No support for any videocard under FreeBSD or any other OS besides Linux/NT/98/95. No specs opened, many developers who are *willing* to introduce this chipset only if they had the specs.
Results: when XFree86 developers or Linux kernel developers are willing to change their implementation the nVidea drivers are likely to be incompatible.
And no, I'm not a gamer, I'm just asking for OpenGL hardware acceleration on my system. Tell me why I should stay with nVidea?
Oh, I've seen them enough, and I can remember enough quotes (mostly on E/Gnome), which were something like this:
"Look at these semi-transparant windows!"
"Look at these cool skins!"
"Hey, look at that, it's not like Windows at all!"
etc. etc.
Most of them turned to Linux, found out that they were supposed to edit files by hand, learn to use a shell and couldn't do any game on it, and returned to Windows.
These are the users we like to call lusers (a general IRC terminology). They run it for some time, show of to their friends and stuff, and before they really got a glitch of the beauty of the UNIX design, they are back to their own comfortable Windows.
When you run any UNIX alike OS (whether this is Solaris, Linux, BSD or whatever), you choose mostly because of it's incredible good internal design. It could have been that you didn't really recognised this, but IMHO this is true.
Another key feature of UNIX is it's really long time history. Standard programs have been looked over, and improved over and over (and over) again. This contributes to stability, speed, ease of use and flexibility (note that ease of use is something different than User Friendly, ease of use is the fact that: "once you know it, it's easy").
Point is, neither the command line nor most GUIs are terribly intuitive. But GUIs, for the end user, make a hell of a lot more sense. Unix's underpinnings are great. Its current interface is absolute garbage.
Well, I understand your points, UNIX interface design was initially a bit poor. But the idea of pipes (and pipelines), shell subsitution, input and output redirectors, etc. etc. has been introduced with a thought behind it.
This thought is called flexibility. And I can't underline this term even more. One of the key things why I use UNIX to it's full extend, and learned to love it, is flexibility. Small applications like sed, awk, find, grep, ls, cp and the others only contribute to this. Good editors like vi or emacs even extends this idea.
But there is a drawback in this idea and it is called "User Friendly". This term has been introduced mainly for new users. The need for this term is obvious in two ways.
First of all, not everybody is as techy as the average Slashdot reader. It is completely out of mind to think that a new computer user will pick up the idea behind UNIX and shells easily.
The second drawback in this idea of flexibility is that it keeps open too many ways for a user to interact with the OS. Again: most techies will like this idea of open-mindedness, and are always willing to learn (myself included). But it also introduces doubt in how to act on certain problems. In 10 seconds I can think of 10 different ways of finding a file on a certain operating system. This might be ideal for flexibility, but it leaves the user with a problem on how to choose his/her best bet.
The idea of using GUI's comes in mind. The use of a mouse comes in mind. But as we can see now, it doesn't really solve the problems involved in making things less complex. Instead of reading manual pages, people are now browsing through all the menus, different windows and still help pages. As it's biggest drawback it seems to loose a lot of flexibility. GUI programs tend to be bigger, capable of doing more and more things, but less than the sum of all the small command line utilities.
Of course, the need for graphical applications is very high. We just *need* them, no doubt about it, but as noted above, it also limits a lot of things. My answer: introduce a shell which is understandable for normal users. A shell which understands lines like:
AI shell> get all files ending in tmp in my home place
OK, I've found 10 files for your request
AI shell> go to the place where my temporary files are stored
OK
AI shell> drop the files there
10 files dropped
AI shell> no, I made an error there, put them back
OK, 10 files put back to your home place
AI shell> edit the document I was working on yesterday
2 files found:
foo.doc
bar.doc
AI shell> edit the last document
OK, editor started
user gets a word editor, opening the file bar.doc
This might seem a bit strange, and really difficult to implement, but if something like this would only nearly be possible, it would be a huge leap for new users to overcome the UNIX-anxiety.
Instead of Linux on these machine, the directors have chosen to install the new MovieOS from Miranda. Not only delivers it much better visualisation on camera (20 columns width), it can even be put on one standard 1.44MB diskette (High Density of course).
I hope MovieOS will soon be opensource, so we can enjoy it too.
What you call about the automatic method execution in "Properties" has been designed in C++ for a very long time.
You want your "properties" to be objects of themselves. Then you can overload all of the operators on these objects, thereby allowing "magically" executed code at the same way.
This structure is much more flexible than Delphi's one, and allows for a far better object model.
Another thing what is bothering me about this new "RAD tool" is that it makes programmers lazy. IMHO your tools should be as flexible as possible, allowing you to do everything that can be done on computers. I just can't see how this can be done by clicking and dragging and typing small sets of code thereby linking objects.
This new tool might be of interest to many people who have difficulties understanding the underlying principles of the graphical toolkit libraries, or just don't want to be bothered by it too much. I think of banking applications, educational tools and other low-level stuff.
As great as this tool might look, how much flexibility will this take away? It might be a pain to write your applications in VI, letting it compile by Kylix (typing:make goes much faster than saving your file, importing it and compiling it). Does it allow for all those lowlevel debugging gdb has? Can I create GTK as well as QT applications with it? Does the user need to download another set of large libraries to use my applications? Most questions are gone unanswered for now, but I guess most of it will not be possible.
A yet even more intruiging question: to compile your GPL'ed software developed under Kylix, do I have to get this compiler too? Does this compiler run under other operating systems than Linux too? One of the key signs of UNIX is it's portability, allowing you to port almost all applications to most hardware/and/ software platforms more or less easily. I guess this will be gone with this new Kylix development tool.
I remain with a lot of questions, any of these can be answered?
Well, it was just on the x86 architecture (times are 286). The tricks used were:
-- mov'ing and setting registers as many cycles apart as possible
-- making a table of MUL's of 320 (getting the stuff from mem goes faster than MUL'ing it real time)
-- completely ASM (of course)
-- some other horror stuff (can't really remember, it was years ago).
So I posted this to a certain newsgroup about gaming in response to somebody asking for it. Blatently noting that this should be the fastest way around. It was not:] I'm sorry, but I can't really remember how he made it even faster, it was something about POP'ing the function arguments faster.
Sidenote: it had to run on a 286 (so no fancy wide EAX registers available) and it had to be a function.
We all seem to find new CPU's "kickass", since they can do decoding/encoding faster, will enable faster generated images, etc. etc.
But as once was stated at the first lecture I saw about Moore's Law: If we don't have the technology (or software) to "use" this new hardware, what good is it? The gap between software and hardware is getting larger every day.
Just a small sidenote: apart from me running seti@home and some rendering stuff, my pII-celeron 266 is mostly having a load of 0.02.
I hope you still read this comment. What I wanted to say was: If I want to make an application which everybody can use, I have to make a decision.
Please don't get me wrong, I don't want to take any choice away, I want a little standard.
Many, many years ago somebody thought about ASCII. Somewhat later, somebody else thought about ANSI. Now all your textmode programs run on it. It might be that you have to download ncurses for an extra level of abstraction, but that's it.
What if somebody pops up and says: I want IICSA which has all the 128 characters turned around and you need to turn your monitor 180 degrees to be compatible with it again. Would you choose it? No. Would you code for it? Of course not. Big operating systems need some standards.
I do not object against all the different windowmanagers around. It's just a top level app, it's not bothering any other application in any way, and it keeps the choices of a developer and a user open.
The thing about Framebuffer->Embedded systems, I did not knew, but I fear it's misuse though.
Neh, wait until your start writing some code:]
A good system must rely on it's standards. You can't say "I'm using Linux" nowadays, it's more "I'm using Slackware or I'm using RedHat".
Instead of adding necessery functionality to things which already excist, new stuff seems to pop out of the ground.
On one side this is good, the standard argument is: let us keep the choice
On the other side this goes wrong: if I want to make a graphical application, what toolkit should I use. I can choose between gtk, qt, gnome enhanced gtk, motif/lesstif or plain Xt or Xp. Tell me?
I think lesstif was a pretty good idea, and it looked very good, but I havn't seen much effort on making it better.
Now we even have the choice between using X and using console (framebuffers). Just to be faster? Are you giving up network support for being faster? I hope not.
Post is completely redundant of course, but I must fully agree with you.
But about a standard he was more or less right: Linux is losing his standards.
In what package format should I set my software?.tar.gz?.rpm? suse's format? Should I include binaries for those that can't wait for a good compile? For what systems should I include binaries?
How is my filestructure organised? Do I have a/dev filesystem, or just a/dev directory/partition? Am I using a ReiserFS, or the ext2 filesystem? Should the startup appear with bright "green" and "red" marks to indicate that something is wrong? Or should I deny it all together and have a nice "flashscreen" at startup?
What site should I get software from which I havn't downloaded yet? Freshmeat? cdrom.com?
Maybe this is all a bit redundant, but the Linux community is losing standards. Most of them are to blame by the distros however. We know that, but we don't have the power to stop that. Any clue?
Oh yes, if Linux goes on like this (with even flames betweeen users about which distro is better), the Linux community will stagger to an end.
But no, that's not what is going to happen. Linux will adapt, probably distros will become unusable because they are "too far off".
Another probability is a shift in/which/ opensource software is going to be used. It's not only Linux under the horizon you know? We still have the BSD's and an C++ alike implementation of the Linux kernel (forgot the name, never used it).
You say users need easier installations, but/what/ needs to be different? I recall clueless people installing MSDOS (which involved just as much trouble as earlier Linux distros) pretty easily.
Yep, people are lazy, people don't want to read the documentation anymore and people don't listen to advices.
I can recall a Redhat 7.0 installation for a demonstration computer at a bookstore I worked for last year. It was so enourmously simple! Nobody can confince me that a kid can't do the installation.
But it's not only the installation. An even bigger part of a system is the maintenance. Every system needs it and you can't walk away not doing it. And that's were things are going wrong, because Linux is still based on the Unix prinicple: maintenance by commandline. And yes, there are tons of handy configer tools which allow you to do the stuff the easy way, but they allow you to bugger the system for around 50%. The power still lies in the shell for the other 50%, and that's were new users are getting stuck. Given the fact that most "users" in contrary with geeks, don't read the documentation.
At first I noticed a little inconcistency in the above article. UNIX is not at all the same as Linux. And at the moment it doesn't seem to be UNIX versus Microsofts OS'es, but Linux versus Windows.
Why do "massmedia-reporters" always say things which are already known? Is it because we (as geeks) know more, or is it because they are too uncertain about what is going on?
I think it is the latter, and prepare for it: it's not going to get any better real soon. All you'll read in magazines like that is stuff which says: "Look at me, I know that the world is round, not flat", years after it is proven that the only flat world is Discworld (a series of books by Terry Pratchett).
And yes: Windows is at the moment beaten up by Linux, and no: We're not talking 98 or ME here, we're talking about their previous topprofit Windows NT (or atm 2000) here. So what? Most experienced users have made their bet by now, or will do so in a not so wide timespan. It/is/ known to most administrators or users that Linux (or BSD) is a viable alternative.
Well, most of the hardware on Linux or other free operating systems is supported by the vendors simple by giving the developers the interface, not the actual drivers.
I've asked for months for support from NVidea when I once bought my nvdidia riva128 chipset supported videocard. If at all a reply there was a corporate salemanager who said their standards were closed.
Sorry, but I can't really participate in your cheerfullness with NVidea.
Yay for Open Source indeed!
No, Linux is not the best, neither are the BSD's, neither is VMS, Solaris or even Win2k. But they all might be the best solution for a certain problem.
Alas, I don't think Linux is bad software, no. I don't think Linux is a bad choice for a server system.
And now I wanted to put my reasons for not choosing Linux, but I don't think it would be of any value for you, or anybody. That's because everybody should take his/her of decision.
I was wrong trying to make my reasons clear, but I can say that I'm quite happy now having chosen FreeBSD. I don't feel any reason for going back to Linux. Maybe if I had newer hardware: yes. At the moment: no. But that's my own decision, you've made yours.
Well, first of all: I shouldn't have posted that comment at all (I could have expected a flamewar on a board full of Linux enthousiasts).
It's not that I want to degrade the GPL or don't like Linus'es face;]. It's about a kernel which is in my opinion quickly losing structure and design (I'm not totally sure about that since the last version I compiled was 2.2.17 iirc). And it is about some idealogical reasons of course. Most people fail to understand that Open Source stands on these idealogical reasons.
It's about me looking through Freshmeat and seeing thousends of codeworks being started but never being finished (of course this does not only affect Linux but every system which relies on open source).
It's about distros making it easy for the newcomer to use Linux. Until he fires a shell.
It's about the very beginning of UNIX: a multiuser multiprocessing monolythical system which might even have degraded to a desktop system? Is that what you think Linux should be? Is that what you think Linus wanted when he started Linux is '91? Of course I will keep Linux as a good alternative and am always keeping an eye open for the ongoing development.
So I need to contribute to make my point? Isn't that a bit odd?
I would take the following approuch into consideration: try to seek out what the design problems are, implement a new design besides the old one (so you create backward compatibility while keeping the system stable). And start porting the pieces based on the old system to the new one.
This has happened before with linux's sound support. It would be a good thing to do the same with other parts of the kernel.
What creates loose ends? Bad design or loose ends.
What solves loose ends? Goto start
You can't solve deep design problems in such a huge project, no matter how many effort you put in it. I've had the problem many times, and the only real solution I could find was rewriting the whole bloody thing from the beginning.
-- Explosion in kernel growth, putting security and stability at a lower level than feature richness.
Explosion in growth, yes. Might also have something to do with the increased number of platforms supported. Am I the only one who thinks that the kernel should be split into several parts: one common part, and one part for each
platform? Lack of concern for stability and security is not the reason.
Of course I admire the efforts of many to make Linux an operating system which can be used on many platforms, but I don't agree with you on splitting the kernel into seperate parts. A better idea might be to make "light" kernels which doesn't have the support for other platforms available, and "heavy" kernels which are for development.
-- Defraction and scattering of features. Too many distros who all have their own way of thinking about how the best Linux system should look like.
So what? Pick the one you like and ignore the rest.
I don't agree with you again. The kernel and userland should be working together, not putting each other on a safe distance so that "everybody" runs Linux, but only as far as the kernel. Don't get me wrong, of course the kernel should be a seperate part with userland, but this doesn't mean that they shouldn't be cooperate in their final goal.
-- Loose application of standards. An oldy is the BSD versus System5 style init in Slack vs. Redhat.
Ditto. Some people like sysv, some like bsd. Pick the one you like. Open software is supposed to be about creating choices, not removing them.
That's a standard quote you've got there, but please note that this doesn't only apply to the distro you use, but also to the operating system you use. It's not about what one likes more that the other, it's about having a standard. And writing software depends on relying on standards.
-- Commercialisation of the attempts of many to produce a "free" operating system. Moreover: doing that at such a high rate that the quality of the whole system degrades (I point to the massive distro RedHat, while also looking at some of the smaller ones like SuSE or Debian).
Red Hat, yes. How exactly has Debian become commercialized?
Excuse, completely my fault:] Substitue Debian with any commercial Linux distribution.
To make a simple conclusion: the spirit of creating a free operating system which does better than the commercial ones, and therefor creating good code at every line of C one writes, is and has been disappearing slightly.
Wrong. Linux (the kernel) is still about making the best OS possible. What Red Hat does is their own business and does not control what the rest of the community does.
No, what Red Hat does/is/ that what the community does. It's about making everything easier at the expense of configurability. It's about supporting more stuff at the expense of stability or code correctness.
smaller than Redhat so to say. Debian isn't small in my opinion, and if any distro has the potential to take over the place of leading distro it must be Debian.
I'm advicing everybody who takes a start with Linux to try Debian, but that's a completely other story of course:]
uhm, I ain't really looking forward to fixing 25 megs of stuff from which a lot is based on code which was badly designed in the first place.
Therefor (and for other reasons), I've chosen another OS.
The problem with Linux is a bit like this: we know that some large chunks of the kernel aren't exactly as they should be, but fixing it costs so much time and breaks so many dependencies that it's kept that way.
My hypothesis (and I'm not entirely sure about that), is that the reason why linux-2.4 took so much time was that by altering a small part of the kernel, it created tons of new problems.
See it this way: adding features to a system which isn't designed for such features is the same thing as breaking the system.
Maybe I'm totally wrong, but from the parts of code I've read in the Linux kernel, and the 8 months I was part of the Linux mailinglist, I think I'm not entirely lieing if give these comments:]
Well, I must say that such a comment is of course meant to make a discussion. But I don't want a flame of pro-linux users vs. pro-unix users about what is more important: features or stability.
I want to leave Linux users in their respectfull position, and I'm not argueing that the whole world should use a FreeBSD, this is just my point of view.
And I must say that I'm somebody who expects a lot of a good operating system, but you can't expect everything. Every operating system has it's pros and it's contras, and maybe my comment also was a desperate attempt for new "posix compatible OS" users to look further than only the look and feel, but also take a look at the inside.
As I'm from.nl, my English might not be everything:]
-- Explosion in kernel growth, putting security and stability at a lower level than feature richness.
-- Defraction and scattering of features. Too many distros who all have their own way of thinking about how the best Linux system should look like.
-- Loose application of standards. An oldy is the BSD versus System5 style init in Slack vs. Redhat.
-- Commercialisation of the attempts of many to produce a "free" operating system. Moreover: doing that at such a high rate that the quality of the whole system degrades (I point to the massive distro RedHat, while also looking at some of the smaller ones like SuSE or Debian).
To make a simple conclusion: the spirit of creating a free operating system which does better than the commercial ones, and therefor creating good code at every line of C one writes, is and has been disappearing slightly.
This is not an attempt to start a flame, it's more an attempt to show my reasons why I stepped over to Linux.
Slashdot Mirror
First of all: I'm not a Linux user, but a FreeBSD user, who is very angry at nVidea at the moment. Let me tell you a little story behind the development of our cute little nVidea Riva128 chipset drivers:
1998 computer bought, only drivers available for Windows NT
1999 Drivers released for Windows 95/98, shortly followed by _very_ buggy Linux drivers.
Right now, nVidea stopped their development for the riva128 chipset on Linux (which means it "probably doesn't work"). No support for any videocard under FreeBSD or any other OS besides Linux/NT/98/95. No specs opened, many developers who are *willing* to introduce this chipset only if they had the specs.
Results: when XFree86 developers or Linux kernel developers are willing to change their implementation the nVidea drivers are likely to be incompatible.
And no, I'm not a gamer, I'm just asking for OpenGL hardware acceleration on my system. Tell me why I should stay with nVidea?
Oh, I've seen them enough, and I can remember enough quotes (mostly on E/Gnome), which were something like this:
"Look at these semi-transparant windows!"
"Look at these cool skins!"
"Hey, look at that, it's not like Windows at all!"
etc. etc.
Most of them turned to Linux, found out that they were supposed to edit files by hand, learn to use a shell and couldn't do any game on it, and returned to Windows.
These are the users we like to call lusers (a general IRC terminology). They run it for some time, show of to their friends and stuff, and before they really got a glitch of the beauty of the UNIX design, they are back to their own comfortable Windows.
When you run any UNIX alike OS (whether this is Solaris, Linux, BSD or whatever), you choose mostly because of it's incredible good internal design. It could have been that you didn't really recognised this, but IMHO this is true.
Another key feature of UNIX is it's really long time history. Standard programs have been looked over, and improved over and over (and over) again. This contributes to stability, speed, ease of use and flexibility (note that ease of use is something different than User Friendly, ease of use is the fact that: "once you know it, it's easy").
Point is, neither the command line nor most GUIs are terribly intuitive. But GUIs, for the end user, make a hell of a lot more sense. Unix's underpinnings are great. Its current interface is absolute garbage.
Well, I understand your points, UNIX interface design was initially a bit poor. But the idea of pipes (and pipelines), shell subsitution, input and output redirectors, etc. etc. has been introduced with a thought behind it.
This thought is called flexibility. And I can't underline this term even more. One of the key things why I use UNIX to it's full extend, and learned to love it, is flexibility. Small applications like sed, awk, find, grep, ls, cp and the others only contribute to this. Good editors like vi or emacs even extends this idea.
But there is a drawback in this idea and it is called "User Friendly". This term has been introduced mainly for new users. The need for this term is obvious in two ways.
First of all, not everybody is as techy as the average Slashdot reader. It is completely out of mind to think that a new computer user will pick up the idea behind UNIX and shells easily.
The second drawback in this idea of flexibility is that it keeps open too many ways for a user to interact with the OS. Again: most techies will like this idea of open-mindedness, and are always willing to learn (myself included). But it also introduces doubt in how to act on certain problems. In 10 seconds I can think of 10 different ways of finding a file on a certain operating system. This might be ideal for flexibility, but it leaves the user with a problem on how to choose his/her best bet.
The idea of using GUI's comes in mind. The use of a mouse comes in mind. But as we can see now, it doesn't really solve the problems involved in making things less complex. Instead of reading manual pages, people are now browsing through all the menus, different windows and still help pages. As it's biggest drawback it seems to loose a lot of flexibility. GUI programs tend to be bigger, capable of doing more and more things, but less than the sum of all the small command line utilities.
Of course, the need for graphical applications is very high. We just *need* them, no doubt about it, but as noted above, it also limits a lot of things. My answer: introduce a shell which is understandable for normal users. A shell which understands lines like:
AI shell> get all files ending in tmp in my home place
OK, I've found 10 files for your request
AI shell> go to the place where my temporary files are stored
OK
AI shell> drop the files there
10 files dropped
AI shell> no, I made an error there, put them back
OK, 10 files put back to your home place
AI shell> edit the document I was working on yesterday
2 files found:
foo.doc
bar.doc
AI shell> edit the last document
OK, editor started
user gets a word editor, opening the file bar.doc
This might seem a bit strange, and really difficult to implement, but if something like this would only nearly be possible, it would be a huge leap for new users to overcome the UNIX-anxiety.
Instead of Linux on these machine, the directors have chosen to install the new MovieOS from Miranda. Not only delivers it much better visualisation on camera (20 columns width), it can even be put on one standard 1.44MB diskette (High Density of course).
I hope MovieOS will soon be opensource, so we can enjoy it too.
What you call about the automatic method execution in "Properties" has been designed in C++ for a very long time.
:make goes much faster than saving your file, importing it and compiling it). Does it allow for all those lowlevel debugging gdb has? Can I create GTK as well as QT applications with it? Does the user need to download another set of large libraries to use my applications? Most questions are gone unanswered for now, but I guess most of it will not be possible.
/and/ software platforms more or less easily. I guess this will be gone with this new Kylix development tool.
You want your "properties" to be objects of themselves. Then you can overload all of the operators on these objects, thereby allowing "magically" executed code at the same way.
This structure is much more flexible than Delphi's one, and allows for a far better object model.
Another thing what is bothering me about this new "RAD tool" is that it makes programmers lazy. IMHO your tools should be as flexible as possible, allowing you to do everything that can be done on computers. I just can't see how this can be done by clicking and dragging and typing small sets of code thereby linking objects.
This new tool might be of interest to many people who have difficulties understanding the underlying principles of the graphical toolkit libraries, or just don't want to be bothered by it too much. I think of banking applications, educational tools and other low-level stuff.
As great as this tool might look, how much flexibility will this take away? It might be a pain to write your applications in VI, letting it compile by Kylix (typing
A yet even more intruiging question: to compile your GPL'ed software developed under Kylix, do I have to get this compiler too? Does this compiler run under other operating systems than Linux too? One of the key signs of UNIX is it's portability, allowing you to port almost all applications to most hardware
I remain with a lot of questions, any of these can be answered?
Well, it was just on the x86 architecture (times are 286). The tricks used were:
:] I'm sorry, but I can't really remember how he made it even faster, it was something about POP'ing the function arguments faster.
-- mov'ing and setting registers as many cycles apart as possible
-- making a table of MUL's of 320 (getting the stuff from mem goes faster than MUL'ing it real time)
-- completely ASM (of course)
-- some other horror stuff (can't really remember, it was years ago).
So I posted this to a certain newsgroup about gaming in response to somebody asking for it. Blatently noting that this should be the fastest way around. It was not
Sidenote: it had to run on a 286 (so no fancy wide EAX registers available) and it had to be a function.
We all seem to find new CPU's "kickass", since they can do decoding/encoding faster, will enable faster generated images, etc. etc.
But as once was stated at the first lecture I saw about Moore's Law: If we don't have the technology (or software) to "use" this new hardware, what good is it? The gap between software and hardware is getting larger every day.
Just a small sidenote: apart from me running seti@home and some rendering stuff, my pII-celeron 266 is mostly having a load of 0.02.
Never say something is impossible. I learned this many years ago, when I said a certain putpixel routines was impossible to get any faster.
I was proven wrong.
Please read "Impossible for Dummies", this must brighten you up.
I hope you still read this comment. What I wanted to say was: If I want to make an application which everybody can use, I have to make a decision.
Please don't get me wrong, I don't want to take any choice away, I want a little standard.
Many, many years ago somebody thought about ASCII. Somewhat later, somebody else thought about ANSI. Now all your textmode programs run on it. It might be that you have to download ncurses for an extra level of abstraction, but that's it.
What if somebody pops up and says: I want IICSA which has all the 128 characters turned around and you need to turn your monitor 180 degrees to be compatible with it again. Would you choose it? No. Would you code for it? Of course not. Big operating systems need some standards.
I do not object against all the different windowmanagers around. It's just a top level app, it's not bothering any other application in any way, and it keeps the choices of a developer and a user open.
The thing about Framebuffer->Embedded systems, I did not knew, but I fear it's misuse though.
And don't forget IISCA is coming your way soon.
Neh, wait until your start writing some code :]
A good system must rely on it's standards. You can't say "I'm using Linux" nowadays, it's more "I'm using Slackware or I'm using RedHat".
Instead of adding necessery functionality to things which already excist, new stuff seems to pop out of the ground.
On one side this is good, the standard argument is: let us keep the choice
On the other side this goes wrong: if I want to make a graphical application, what toolkit should I use. I can choose between gtk, qt, gnome enhanced gtk, motif/lesstif or plain Xt or Xp. Tell me?
I think lesstif was a pretty good idea, and it looked very good, but I havn't seen much effort on making it better.
Now we even have the choice between using X and using console (framebuffers). Just to be faster? Are you giving up network support for being faster? I hope not.
I didn't knew this, do you have any link or documentation about this contract?
Post is completely redundant of course, but I must fully agree with you.
.tar.gz? .rpm? suse's format? Should I include binaries for those that can't wait for a good compile? For what systems should I include binaries?
/dev filesystem, or just a /dev directory/partition? Am I using a ReiserFS, or the ext2 filesystem? Should the startup appear with bright "green" and "red" marks to indicate that something is wrong? Or should I deny it all together and have a nice "flashscreen" at startup?
But about a standard he was more or less right: Linux is losing his standards.
In what package format should I set my software?
How is my filestructure organised? Do I have a
What site should I get software from which I havn't downloaded yet? Freshmeat? cdrom.com?
Maybe this is all a bit redundant, but the Linux community is losing standards. Most of them are to blame by the distros however. We know that, but we don't have the power to stop that. Any clue?
Oh yes, if Linux goes on like this (with even flames betweeen users about which distro is better), the Linux community will stagger to an end.
/which/ opensource software is going to be used. It's not only Linux under the horizon you know? We still have the BSD's and an C++ alike implementation of the Linux kernel (forgot the name, never used it).
/what/ needs to be different? I recall clueless people installing MSDOS (which involved just as much trouble as earlier Linux distros) pretty easily.
But no, that's not what is going to happen. Linux will adapt, probably distros will become unusable because they are "too far off".
Another probability is a shift in
You say users need easier installations, but
Yep, people are lazy, people don't want to read the documentation anymore and people don't listen to advices.
I can recall a Redhat 7.0 installation for a demonstration computer at a bookstore I worked for last year. It was so enourmously simple! Nobody can confince me that a kid can't do the installation.
But it's not only the installation. An even bigger part of a system is the maintenance. Every system needs it and you can't walk away not doing it. And that's were things are going wrong, because Linux is still based on the Unix prinicple: maintenance by commandline. And yes, there are tons of handy configer tools which allow you to do the stuff the easy way, but they allow you to bugger the system for around 50%. The power still lies in the shell for the other 50%, and that's were new users are getting stuck. Given the fact that most "users" in contrary with geeks, don't read the documentation.
Just my 5c
At first I noticed a little inconcistency in the above article. UNIX is not at all the same as Linux. And at the moment it doesn't seem to be UNIX versus Microsofts OS'es, but Linux versus Windows.
/is/ known to most administrators or users that Linux (or BSD) is a viable alternative.
Why do "massmedia-reporters" always say things which are already known? Is it because we (as geeks) know more, or is it because they are too uncertain about what is going on?
I think it is the latter, and prepare for it: it's not going to get any better real soon. All you'll read in magazines like that is stuff which says: "Look at me, I know that the world is round, not flat", years after it is proven that the only flat world is Discworld (a series of books by Terry Pratchett).
And yes: Windows is at the moment beaten up by Linux, and no: We're not talking 98 or ME here, we're talking about their previous topprofit Windows NT (or atm 2000) here. So what? Most experienced users have made their bet by now, or will do so in a not so wide timespan. It
Well, most of the hardware on Linux or other free operating systems is supported by the vendors simple by giving the developers the interface, not the actual drivers.
I've asked for months for support from NVidea when I once bought my nvdidia riva128 chipset supported videocard. If at all a reply there was a corporate salemanager who said their standards were closed.
Sorry, but I can't really participate in your cheerfullness with NVidea.
Yay for Open Source indeed!
No, Linux is not the best, neither are the BSD's, neither is VMS, Solaris or even Win2k. But they all might be the best solution for a certain problem.
Alas, I don't think Linux is bad software, no. I don't think Linux is a bad choice for a server system.
And now I wanted to put my reasons for not choosing Linux, but I don't think it would be of any value for you, or anybody. That's because everybody should take his/her of decision.
I was wrong trying to make my reasons clear, but I can say that I'm quite happy now having chosen FreeBSD. I don't feel any reason for going back to Linux. Maybe if I had newer hardware: yes. At the moment: no. But that's my own decision, you've made yours.
Well, first of all: I shouldn't have posted that comment at all (I could have expected a flamewar on a board full of Linux enthousiasts).
;]. It's about a kernel which is in my opinion quickly losing structure and design (I'm not totally sure about that since the last version I compiled was 2.2.17 iirc). And it is about some idealogical reasons of course. Most people fail to understand that Open Source stands on these idealogical reasons.
It's not that I want to degrade the GPL or don't like Linus'es face
It's about me looking through Freshmeat and seeing thousends of codeworks being started but never being finished (of course this does not only affect Linux but every system which relies on open source).
It's about distros making it easy for the newcomer to use Linux. Until he fires a shell.
It's about the very beginning of UNIX: a multiuser multiprocessing monolythical system which might even have degraded to a desktop system? Is that what you think Linux should be? Is that what you think Linus wanted when he started Linux is '91? Of course I will keep Linux as a good alternative and am always keeping an eye open for the ongoing development.
So I need to contribute to make my point? Isn't that a bit odd?
I would take the following approuch into consideration: try to seek out what the design problems are, implement a new design besides the old one (so you create backward compatibility while keeping the system stable). And start porting the pieces based on the old system to the new one.
This has happened before with linux's sound support. It would be a good thing to do the same with other parts of the kernel.
What creates loose ends? Bad design or loose ends.
What solves loose ends? Goto start
You can't solve deep design problems in such a huge project, no matter how many effort you put in it. I've had the problem many times, and the only real solution I could find was rewriting the whole bloody thing from the beginning.
-- Explosion in kernel growth, putting security and stability at a lower level than feature richness.
:] Substitue Debian with any commercial Linux distribution.
/is/ that what the community does. It's about making everything easier at the expense of configurability. It's about supporting more stuff at the expense of stability or code correctness.
Explosion in growth, yes. Might also have something to do with the increased number of platforms supported. Am I the only one who thinks that the kernel should be split into several parts: one common part, and one part for each
platform? Lack of concern for stability and security is not the reason.
Of course I admire the efforts of many to make Linux an operating system which can be used on many platforms, but I don't agree with you on splitting the kernel into seperate parts. A better idea might be to make "light" kernels which doesn't have the support for other platforms available, and "heavy" kernels which are for development.
-- Defraction and scattering of features. Too many distros who all have their own way of thinking about how the best Linux system should look like.
So what? Pick the one you like and ignore the rest.
I don't agree with you again. The kernel and userland should be working together, not putting each other on a safe distance so that "everybody" runs Linux, but only as far as the kernel. Don't get me wrong, of course the kernel should be a seperate part with userland, but this doesn't mean that they shouldn't be cooperate in their final goal.
-- Loose application of standards. An oldy is the BSD versus System5 style init in Slack vs. Redhat.
Ditto. Some people like sysv, some like bsd. Pick the one you like. Open software is supposed to be about creating choices, not removing them.
That's a standard quote you've got there, but please note that this doesn't only apply to the distro you use, but also to the operating system you use. It's not about what one likes more that the other, it's about having a standard. And writing software depends on relying on standards.
-- Commercialisation of the attempts of many to produce a "free" operating system. Moreover: doing that at such a high rate that the quality of the whole system degrades (I point to the massive distro RedHat, while also looking at some of the smaller ones like SuSE or Debian).
Red Hat, yes. How exactly has Debian become commercialized?
Excuse, completely my fault
To make a simple conclusion: the spirit of creating a free operating system which does better than the commercial ones, and therefor creating good code at every line of C one writes, is and has been disappearing slightly.
Wrong. Linux (the kernel) is still about making the best OS possible. What Red Hat does is their own business and does not control what the rest of the community does.
No, what Red Hat does
smaller than Redhat so to say. Debian isn't small in my opinion, and if any distro has the potential to take over the place of leading distro it must be Debian. :]
I'm advicing everybody who takes a start with Linux to try Debian, but that's a completely other story of course
uhm, I ain't really looking forward to fixing 25 megs of stuff from which a lot is based on code which was badly designed in the first place.
:]
Therefor (and for other reasons), I've chosen another OS.
The problem with Linux is a bit like this: we know that some large chunks of the kernel aren't exactly as they should be, but fixing it costs so much time and breaks so many dependencies that it's kept that way.
My hypothesis (and I'm not entirely sure about that), is that the reason why linux-2.4 took so much time was that by altering a small part of the kernel, it created tons of new problems.
See it this way: adding features to a system which isn't designed for such features is the same thing as breaking the system.
Maybe I'm totally wrong, but from the parts of code I've read in the Linux kernel, and the 8 months I was part of the Linux mailinglist, I think I'm not entirely lieing if give these comments
Well, I must say that such a comment is of course meant to make a discussion. But I don't want a flame of pro-linux users vs. pro-unix users about what is more important: features or stability.
.nl, my English might not be everything :]
I want to leave Linux users in their respectfull position, and I'm not argueing that the whole world should use a FreeBSD, this is just my point of view.
And I must say that I'm somebody who expects a lot of a good operating system, but you can't expect everything. Every operating system has it's pros and it's contras, and maybe my comment also was a desperate attempt for new "posix compatible OS" users to look further than only the look and feel, but also take a look at the inside.
As I'm from
-- Explosion in kernel growth, putting security and stability at a lower level than feature richness.
-- Defraction and scattering of features. Too many distros who all have their own way of thinking about how the best Linux system should look like.
-- Loose application of standards. An oldy is the BSD versus System5 style init in Slack vs. Redhat.
-- Commercialisation of the attempts of many to produce a "free" operating system. Moreover: doing that at such a high rate that the quality of the whole system degrades (I point to the massive distro RedHat, while also looking at some of the smaller ones like SuSE or Debian).
To make a simple conclusion: the spirit of creating a free operating system which does better than the commercial ones, and therefor creating good code at every line of C one writes, is and has been disappearing slightly.
This is not an attempt to start a flame, it's more an attempt to show my reasons why I stepped over to Linux.
Thanks alot, I almost volunteered as betatester for the game. And I don't really want to know what he would do with my mailadres.
I kind of wondered how he wanted to achieve so much things in such a short time while his screenshots were still so crappy.
Thanks again, greets