An article at OSNews which I want to bring to your attention. I could use some help in the ensueing discussion against the people dumbassedly asking why we don't just patch the Linux kernel... (yeah, right, hunting a supersonic moving target with a blowgun...)
http://www.osnews.com/story.php?news_id=8162
Why you shouldn't write your own kernel anymore
Why you shouldn't write your own kernel anymore
Every good solution is obvious once you've found it.
-
Curufir
Re:Why you shouldn't write your own kernel anymore
Already read your comments (I have a different alter-ego over there), I'll just stand on the sideline and cheerlead for now (Where are those damn pompoms ;D).
The guy who wrote the article is right in that for a commercial venture it makes much more sense to use an existing OS and modify as/when required.
Having said that it's amusing to see the number of dumbasses that have never even touched system code and still feel qualified to say that Linux/*nix is the "One true way" and soooo much more advanced than anything a hobbyist could create. There are some folks on that board that do know what they're talking about, but picking out their comments from the background noise is an artform in itself.
What most folks don't seem to comprehend is that a large amount of seriously crap code exists in *nix-land, but because they use it every day, and it mostly gets the job done, it's assumed to be the dog's bollocks. Well some of us aren't happy with "adequate" software, and OS dev is just one outlet for that emotion.
"Worse is better", now that guy really knew what he was talking about.
The guy who wrote the article is right in that for a commercial venture it makes much more sense to use an existing OS and modify as/when required.
Having said that it's amusing to see the number of dumbasses that have never even touched system code and still feel qualified to say that Linux/*nix is the "One true way" and soooo much more advanced than anything a hobbyist could create. There are some folks on that board that do know what they're talking about, but picking out their comments from the background noise is an artform in itself.
What most folks don't seem to comprehend is that a large amount of seriously crap code exists in *nix-land, but because they use it every day, and it mostly gets the job done, it's assumed to be the dog's bollocks. Well some of us aren't happy with "adequate" software, and OS dev is just one outlet for that emotion.
"Worse is better", now that guy really knew what he was talking about.
-
Schol-R-LEA
Re:Why you shouldn't write your own kernel anymore
Reasons why programmers write their own kernels, even if it is probably a pointless thing to do in the Bigger Picture:
- I'm doing this for fun, or to learn, or to hone my programming skills. The final result is less important that the process.
- I'm doing research, trying to find new and better ways to do things. By definition, this means starting over from scratch for at least part of the system.
- I have a design idea that is radically different, and want to test the concept.
- The high-level user interface designs I want to use won't work on existing kernels or with existing standards.
- "Screw the standards! Major change can only come from breaking with the past." (regardless of how true this is, it is a common attitude and one reason why many of us do this.)
- I want to try writing an OS in an unusual language or in an unusual manner, one which precludes using a conventional kernel.
- I'm not satisfied with the existing systems, and do not think I ever will be.
and the number one reason for writing your own kernel: - Because I want the gratification of say I did it - let's face it, that's the real reason most of us are here, right?
-
Dreamsmith
Re:Why you shouldn't write your own kernel anymore
Why write your own kernel, when there are ready-made kernels to use? Writing your own code is a pain.
Why build your own birdhouse, when you can just buy one at the store? Woodworking tools are expensive and a pain to use.
Why cook your family a meal, when you can just stop by the fast food joint on the way home? It's such a hassle cleaning up the kitchen.
Why sing, when you can play a CD? Heck, even Rob Zombie has a better singing voice than you.
Why write yet another book, when there are already more books than anyone could read in a lifetime?
Why paint yet another picture, when there are already more paintings in the world than anyone could possibly fit on all their walls?
As soon as the last artist in the world puts away is paintbrush for good, the last novelist retires, people stop singing in the shower, and only designated professionals cook or build things, I promise then I will stop working on my kernel. But not a minute sooner than that...
Why build your own birdhouse, when you can just buy one at the store? Woodworking tools are expensive and a pain to use.
Why cook your family a meal, when you can just stop by the fast food joint on the way home? It's such a hassle cleaning up the kitchen.
Why sing, when you can play a CD? Heck, even Rob Zombie has a better singing voice than you.
Why write yet another book, when there are already more books than anyone could read in a lifetime?
Why paint yet another picture, when there are already more paintings in the world than anyone could possibly fit on all their walls?
As soon as the last artist in the world puts away is paintbrush for good, the last novelist retires, people stop singing in the shower, and only designated professionals cook or build things, I promise then I will stop working on my kernel. But not a minute sooner than that...
-
Legend
Re:Why you shouldn't write your own kernel anymore
Added my 2 cents ...
And directly assaulted POSIX and UNIX in that way
Like I always do ...
And directly assaulted POSIX and UNIX in that way
Like I always do ...
-
mystran
Re:Why you shouldn't write your own kernel anymore
Agree in attitude with Dreamsmith, even if I'm not currently working with an OS kernel. I'm busy with a GUI toolkit instead. (why don't you just use some existing toolkit =?)
-
mystran
Re:Why you shouldn't write your own kernel anymore
Another thing that occured to me, is that many people don't think (whatever they claim) of "Windows" and "Linux" as general purpose operating systems for a generic device called "a computer". Instead, they think of these OS as (depending on their skill level) "part of the device" or "way to make desktop computer from that heap of hardware".
However, many of us think of developing a "general purpose OS" while in fact, there is "Windows" which is one kind of "device" and "Unix" which is another type of device, which can do many similar things (and run the same programs) but are still distinct kind of things with their own philosophies.
Many people even use both of these "device": either they have a separate "Unix-computer" and a "Windows-computer" or they use the same hardware using either dual-boot or some form of emulation.
So, I started wondering, instead of developing a "general purpose OS", should one instead try to "realize a vision of a device" of certain kind. I mean, once you've fullfilled that, it's probably complete enough that IF it is similar enough with the existing "general purpose OS" one can port some application on top of it. Or maybe it can support them using some form of emulation?
My point is, instead of thinking like "I want it to run games, and I want it to be able to surf the web" one should start from more fundamental concepts like Unix's "everything is file" or Mac's "provide a graphical environment where programs look and feel the same, so they are easier to learn".
Why has the target "device" of OS development have to be "a general purpose desktop OS" or "stable and secure server system". Why not start from some more specific goal?
Not like anyone has to do this. If I ever get some higher priority projects "done" and still have time to work on such things, I might do something like that myself. I yeah, I know many of you have ideologies like that... just, wanted to share this thought.
However, many of us think of developing a "general purpose OS" while in fact, there is "Windows" which is one kind of "device" and "Unix" which is another type of device, which can do many similar things (and run the same programs) but are still distinct kind of things with their own philosophies.
Many people even use both of these "device": either they have a separate "Unix-computer" and a "Windows-computer" or they use the same hardware using either dual-boot or some form of emulation.
So, I started wondering, instead of developing a "general purpose OS", should one instead try to "realize a vision of a device" of certain kind. I mean, once you've fullfilled that, it's probably complete enough that IF it is similar enough with the existing "general purpose OS" one can port some application on top of it. Or maybe it can support them using some form of emulation?
My point is, instead of thinking like "I want it to run games, and I want it to be able to surf the web" one should start from more fundamental concepts like Unix's "everything is file" or Mac's "provide a graphical environment where programs look and feel the same, so they are easier to learn".
Why has the target "device" of OS development have to be "a general purpose desktop OS" or "stable and secure server system". Why not start from some more specific goal?
Not like anyone has to do this. If I ever get some higher priority projects "done" and still have time to work on such things, I might do something like that myself. I yeah, I know many of you have ideologies like that... just, wanted to share this thought.
Re:Why you shouldn't write your own kernel anymore
I strongly object... but that's a bit OT.Dreamsmith wrote:
Why sing, when you can play a CD? Heck, even Rob Zombie has a better singing voice than you.
Every good solution is obvious once you've found it.
-
df in singapore
Re:Why you shouldn't write your own kernel anymore
if you havnt worked it out, ceaser = df = man under the bridge = troll
hahaha. i used to like osnews but it is so laughable its just not funny anymore....
hahaha. i used to like osnews but it is so laughable its just not funny anymore....
-
Curufir
Re:Why you shouldn't write your own kernel anymore
Seems some folks on that board just cannot handle any criticism of *nix, GPL, or OSS in general.
I think it's kind of hilarious
.
Although I would have asked the question of why the Nvidia driver needs recompiling if the kernel is recompiled. If that isn't evidence of a shifting driver API then I don't know what is. Having to recompile smells suspiciously like the Nvidia driver is being forced to use absolute addresses for their kernel module. Either way it's not a good sign.
I think it's kind of hilarious
.Although I would have asked the question of why the Nvidia driver needs recompiling if the kernel is recompiled. If that isn't evidence of a shifting driver API then I don't know what is. Having to recompile smells suspiciously like the Nvidia driver is being forced to use absolute addresses for their kernel module. Either way it's not a good sign.