You will see the advantage when you compile a kernel that only contains what you need. I didn't say that the compiled kernel should contain all options available in the "make menuconfig".
Of course being able to compile the kernel yourself is an advantage but my point was all drivers (except the ones need to boot) should be external to the kernel. For example if the kernel should only have ext3 driver (or wharever file system you use) built in so all the other drivers can be loaded. This would save resources because when I'm not reading my ntfs or FAT32 drive the driver doesn't need to be in memory.
And why should NTFS be enabled as default in the kernel? The filesystem is a Windows one and should be considered a bonus.
Yes, it is a bonus but that wasn't my point.
Also, many desktop distros already have support for ntfs. Servers which is where Linux is used most don't have any use for ntfs. at least not most of them.
Redhat doesn't, but it would be better if you could just download and compile another module that can be loaded on demand to read the ntfs drive instead of bloating up the kernel.
They probably use ext3, jfs, xfs, reiserfs or reiser4.
This Linux ntfs driver is pretty shitty as it only really supports read only access. I'm not makeing a point here because ntfs is Windows not Linux and read only access is good enough anyway.
Oh sorry I do have a point, the Linux ntfs driver isn't as good as the Windows driver because the people who wrote it did so by a process of reverse engineering - similar to how most Linux drivers are written.
My main argument is that you get all your drivers you need with one cd, not 10 or 20. They are also up to date and should work better with the current software.
Sorry, while they may be anough to get your system booted no Linux distribution has contained drivers for my printer/scanner, winmodem, and the graphics card driver is always slow.
I personally hate to install all drivers after a reformat
I thought Linux was supposed to be so stable a reformat is rarely needed.
as I have a lot of controllers, other peripherals and so on. It just isn't effective. That's true. I have a high end HP printer and it works fine with the CD driver in Windows, but you still need the disc
The disc that came with the printer?
You should just keep this in a safe place with the printer manual warranty ect.
or an internet connection to get a new one with bugfixes. I use DSL but I think it would be rather crappy to have for ex a 56k connection and download a 150mb driver package (audigy for ex)
I only have a 56k connection and the printer driver was well under 1MB.
because of the CD being lost.
That's your fault for loosing it, you should keep it in a safe place.
I've read and heard a lot about this problem before, where the customers had to pay 10$ to get a new driver cd.
If they kept in in a safe place they wouldn't have this problem and all they need to do is download a small file from the internet anyway.
Think off needing to get a school work done on a Windows machine... You have no drivers installed for the printer and the disc is somehow bad and your internet connection is down. All you have is the Windows disc.
Or you could be in an even same position on a Linux machine, but even worse the printer driver you require simply doesn't exist. This wouln't bother me either way as I would just take the work to school on a disc.
Yes, but that's because Canon is "alive".
Canon could've been dead and it wouln't have made any difference the driver I use on Windows XP (NT5.1) was designed for NT4 which is very old. The only way Canon being alive has benifited is I could ring the support, but I could've found this information on the Intenet anyway.
Many manufacturers end support for certain units and devices which get the customers left in dark like with my bluetooth dongle and probably like 5 gaming units.
I can see how this could be a problem, the only possible advantage of hacker-written drivers is that manufacturer support is not needed to obtain the drivers.
Yes It's fair because I'm referring to a XP SP2 cd, and It's nearly still the same regarding driver support as it was with original XP and SP1.
Ok fair enough but lets not blame the operating system for driver problems as it's the manufacturer's fault.
Hardware manufacturers are the ones who have the best specifications of their stuff but I think they should collaborate more with open source developers
Yes I agree, the problem is manufacturers ofen want to keep their trade secretes - something not compatable with open source software.
so that it will be possible to integrate the drivers into the kernel.
And make it bloated?
No it'd be better to keep them separate and load them when they're needed.
Well, I was mainly thinking about the driver model in the two OS,s .
I'd say the Windows driver model is better for the same reason the Mac model is better than Linux because the manufacturers support them something that Linux just can't help.
Yes but you don't need to dualboot and you're able to enter the system pretty fast. It's not so bad really because you'll only use it when you'll need those "special" apps. I'm currently running XP SP2 in the virtual machine
Now that is a very good idea, I think I'll need to upgrade though, while 256MB of RAM might be good enough for what I currently use my PC for I'll need more to fit two operating systems in memory but it might just be worth it.
but If I want I could just put a Longhorn beta/alpha on it as I don't have to get scared about the system going down because the alpha/beta state. Soon I'll put OSX 10.4 under Pearpc which is available under linux. Will be pretty cool to experiment with. Yes I've heard about that but I don't understand why really. The 2.6 kernel is extremely stable at least for heavy gaming usage and heavy gcc compiling + some video rendering without any hitches. The 2.6 kernel is also much faster.
All I was saying was that it seems strange a Windows driver binary compiled for a very old kernel works on the most recent kernel and a Linux driver designed for a very slightly older kernel won't work even though it's compiled from source.
Maybe I'll go back to the slightly older 2.4.x kernel the next time I can be bothered to install Linux.
Whoosh. OLD!! I haven't seen many kernel upgrades on Windows update or in the patches for a long time. It's mostly small fixes which probably are for modules anyway.
Yes it is a bit old, Windows 2000 was NT 5.0 and Windows XP sp2 is NT 5.1.26 - only a minor update which really takes this piss if you ask me. You pay though the nose for just a minor update with Microsoft software.
The kernel isn't the bad part of Windows anyway I think Cutler did a very good job. The bad part of Windows is all the other shit that's wrapped around the kernel.
