One-letter class names? Is he nuts? That guy never had to maintain code I guess...
Pixel works for MandrakeSoft and responsible for many important modules, for example installation. IMHO it's one of the best Linux installs.
Anyway I agree with you and I'm sure Pixel too. He's pretty busy with many things and it seems this comparison was written "quickly" and mostly for himself to help designing his own language, merd in his free time.
Nevertheless the page is quite interesting, it would be nice to see more on this topic, especially from those people who are actually working on this area, not just as a free time hobby.
Repartitioning on 2.6 kernels can result incorrect partition table for Windows boot and they stop booting. Mandrake 10 and SUSE 9.1 have the same problem. There is more information and potential solutions on this site.
Examples are ndis and ntfs.sys. This trend will continue.
Does Captive NTFS (ntfs.sys) work for you? Users report very slow speed (50-100 KB/sec) and lost data. Even the author cannot promise any" reliability.
But do you think they'd tollerate full R/W support from Linux without trying to break it?
Breaking NTFS is part of the technology evolution. ext3 "broke" ext2, reiser4 "broke" reiserfs, etc. So why you can't mess your filesystem? Because everytime these changes happen the on-disk filesystem version, signatures are also incremented. Not doing so Microsoft would damage itself, for example by W2K, XP destroying Longhorn NTFS. If this version change is detected then no data harm is possible. The old, unmaintained NTFS driver didn't check the NTFS version so when W2K came out the NT4 driver destroyed the W2K NTFS. Kernel developers fixed this only much later on thus the damage was already done.
About WinFS: it's not a filesystem. It's a storage layer on top of NTFS.
They are. But NTFS is a huge project. It takes a lot of time. Today it's developed as a hobby, couple of hours a week or not even that much.
> I would really like to think that in a few months
> someone will figure out how to (safely) lift these restrictions
The developers said in the forums: somebody having lots of time must start active coding (no need for reverse engineering any more) or hire coders and finance the development.
Varies on country. Read the EULA. If it's valid in your country it's apparently a violation.
However the EULA also states that any use of the software not expressly granted to the end user is reserved by Microsoft. This way Microsoft can say OK for friends and NO for competitors. Did you already forget when Microsoft threatened MS Visual FoxPro users some months ago who used the same trick?
The difference isn't reliability but functionality. The rewritten NTFS code is reliable but it's incomplete, it gives "permission denied" or "not supported" for most of the write operations.
The commands, you wrote, should work with either driver without any kernel modification but Captive NTFS is much slower.
Will this driver speed up the creation of a native one? Since now NTFS.SYS is working under Linux although through emulation it should be quite easy to spy to what it's doing, and try to improve the native driver based on that.
IMHO it won't help much. The native driver developers say, they have the knowledge but not the time to finish the driver. NTFS is quite complex.
Partimage doesn't use the NTFS kernel driver. It's a user space NTFS implementation based on the publicly available NTFS information.
it's NTFS support is presumably as unstable as that in the Linux kernel (which claims to be quite immature). I didn't test it in deployment, but I didn't figure it was worth it. Some previous poster told me it worked for him, though.
There are at least 4 different NTFS implementations with different functionality, quality and license for Linux. Which one are you talking about?
And he told it to me without being an ass about it. Learn some respect. Read the post before commenting, please.
It's better to check the facts... ooops apparently there is even a NTFS "ghost" for Linux there, listed as STABLE and called ntfsclone.
Part image does not promise NTFS stuff will work, however. The general problems of M$ making
things difficult are not ever going to go away and those who make tools to interact with NTFS risk bad things from bad laws.
Unfortunately Partimage doesn't use the Linux-NTFS project's stable codebase, at least when I checked some months ago. Partimage developers are indeed guessing NTFS.
I searched the forums on Sourceforge but wasn't able to find any mention of Microsoft either intentionally changing or not changing
the format.
The NTFS format improves, just like ext2 to ext3, etc. The Linux-NTFS project supports all of them up to the latest Windows Server 2003 NTFS format, what actually is the same what XP uses.
XP home OEM installs normally use NTFS, and as far as I am aware(correct me if I am wrong please!) only Mandrake 9.1 can
resize NTFS at the moment. Of course, Mandrake GPL'ing all of their own code means that feature will be in all of the distros rather
soon, but at the moment, it does require third party windows apps to do it, or restricts you to a single distro
You're wrong. Several distros support non-destructive NTFS resizing in some level (GUI, command line, wizzard, etc). One of them is doing it over 2 years, native. Mandrake uses the GPL'ed ntfsresize. It's also a native Linux tool, available for everyone for about one year from the Linux-NTFS project pages who took over the old, broken NTFS code and fixed, improved it.
I've sucessfully resized FAT partitions on several computers using free / demo tools, but what about NTFS? IIRC, XP is part of the NT branch, and thus uses an NTFS variant.
If you read this thread, not only write, then you would easily find the answer;-)
If anyone can find me an ext2 splitter
resize2fs, ext2resize, parted,... usually at least one of these are shipped with the distro one installed.
Really? I thought Microsoft had patents that kept free software from writing to NTFS. Well, certian version of NTFS at least. here is a cluefull letter about NTFS and installs.
Instead of quoting a one year old email, you could check out the Linux-NTFS project carefully who took over the broken NTFS driver, reverse engineered all versions, wrote a new NTFS driver and the NTFS resizer that has been working for a year?
I don't know if it can write to NTFS. NTFS is like the.DOC file format of Word: as soon as someone gets close to figuring it out,
Microsoft changes it slightly, so nobody else can be truly compatible.
These kind of claims were repeatedly rebuted by the Linux-NTFS team. See their mailing lists, forums, web pages for the NTFS documentation. Their problem is the lack of time to implement all the features, NTFS is very complex.
I've looked at Phat but haven't actually used it. Has anyone tried it with NTFS? In a recent kernel, there was support for NTFS
write support, but it was marked "experimental" and "dangerous". Maybe Phat does it anyway....
PhatLinux uses the new NTFS driver, developed by the Linux-NTFS team. Not the broken, old one. Also TopologiLinux is much more popular than PhatLinux, almost half million downloads, and constantly improved. It also runs off an NTFS partition read-write.
I tried using qtparted on my ancient Toshiba Tecra 520 laptop's 4-gig drive *today* and it failed. That was with Knoppix
2003-0606 BTW, the latest non-dvd version. Couldn't even resize hda1 (Fat32) from 1-gig to fill up the rest of the free space on
the drive, which should have been simple.
What was the error message? Did you submit bug report? There are two known problems in libparted popping up rarely but those were fixed. Unfortunately they aren't yet integrated to QTparted, AFAIK. I didn't bother to ask the QTParted author for update because I've never had these problems.
Yeah, and I've found the hard way that ntfsresize is by no means idiot-proof, very far from it. I'm not so much an idiot but sometimes mistyping or forgetful, and that was enough for me to screw up the company laptop...
I think NTFS resizing on Linux is quite idiot-proof. I always use and hand over my Knoppix CD, having QTParted, to my co-workers and friends who either lack or have problem getting Partition Magic to run or work.
I'm sure you messed up the only thing you could, the partition table manipulation. This is filesystem independent and tools like parted, QTParted and DiskDrake handle the entire process a user friendly way. You used the wrong, too low level tools to do the job.
Resizing NTFS partition is not supported by SuSE, and I would gather SuSE is modern enough for you? If resizing NTFS was so very safe, I gather SuSE would have included it by now.
The ntfsresize FAQ says "ntfsresize is in the SuSE 8.2 rescue system".
About its safeness:
"Since July of 2002, when ntfsresize became publicly available, there were many success reports for both enlarging and shrinking Windows
XP/2000/NT4 and Windows Server 2003 NTFS filesystems on both workstation and server versions (Home, Professional, Server, Advanced Server). No destroyed filesystem was reported who followed the instructions correctly".
Do you think they and Mandrake would release it as stable otherwise? Also note, the Linux-NTFS team is not the one who wrote the well known broken NTFS driver, that can destroy data. This is also explained on the above pages. There are two NTFS drivers and most distros use their driver. They also list who use which driver.
There are other myths rebuted and explained on their pages, IMHO it's worth a look.
If you enabled NTFS read/write support in the Linux kernel, you are to blame. NTFS write support is APTLY labeled 'DANGEROUS'. So most likely you hosed the partition.
Quite unlikely. If you read the linux-ntfs project we pages you can find that write is disabled for XP in the *old*, unmaintained driver and the limited write support is said to be safe using the *new* NTFS driver developed by the project. There is also a short list there what NTFS driver vendors use.
Slashdot Mirror
One-letter class names? Is he nuts? That guy never had to maintain code I guess...
Pixel works for MandrakeSoft and responsible for many important modules, for example installation. IMHO it's one of the best Linux installs.
Anyway I agree with you and I'm sure Pixel too. He's pretty busy with many things and it seems this comparison was written "quickly" and mostly for himself to help designing his own language, merd in his free time.
Nevertheless the page is quite interesting, it would be nice to see more on this topic, especially from those people who are actually working on this area, not just as a free time hobby.
Repartitioning on 2.6 kernels can result incorrect partition table for Windows boot and they stop booting. Mandrake 10 and SUSE 9.1 have the same problem. There is more information and potential solutions on this site.
Who are actually in the "western nations" today (aka accepting software patents laws)? USA. Anybody else? Perhaps Canada, Japan?
Google says Bob will use XP with the preinstalled NTFS and it's quite probable he doesn't want to dump it immediately and because Debian still doesn't support non-destructive NTFS resizing thus the install will fail for him.
Does Captive NTFS (ntfs.sys) work for you? Users report very slow speed (50-100 KB/sec) and lost data. Even the author cannot promise any" reliability.
- But do you think they'd tollerate full R/W support from Linux without trying to break it?
Breaking NTFS is part of the technology evolution. ext3 "broke" ext2, reiser4 "broke" reiserfs, etc. So why you can't mess your filesystem? Because everytime these changes happen the on-disk filesystem version, signatures are also incremented. Not doing so Microsoft would damage itself, for example by W2K, XP destroying Longhorn NTFS. If this version change is detected then no data harm is possible. The old, unmaintained NTFS driver didn't check the NTFS version so when W2K came out the NT4 driver destroyed the W2K NTFS. Kernel developers fixed this only much later on thus the damage was already done.About WinFS: it's not a filesystem. It's a storage layer on top of NTFS.
They are. But NTFS is a huge project. It takes a lot of time. Today it's developed as a hobby, couple of hours a week or not even that much.
> I would really like to think that in a few months
> someone will figure out how to (safely) lift these restrictions
The developers said in the forums: somebody having lots of time must start active coding (no need for reverse engineering any more) or hire coders and finance the development.
However the EULA also states that any use of the software not expressly granted to the end user is reserved by Microsoft. This way Microsoft can say OK for friends and NO for competitors. Did you already forget when Microsoft threatened MS Visual FoxPro users some months ago who used the same trick?
Ditch Microsoft then no such troubles.
The commands, you wrote, should work with either driver without any kernel modification but Captive NTFS is much slower.
- Will this driver speed up the creation of a native one? Since now NTFS.SYS is working under Linux although through emulation it should be quite easy to spy to what it's doing, and try to improve the native driver based on that.
IMHO it won't help much. The native driver developers say, they have the knowledge but not the time to finish the driver. NTFS is quite complex.- Since it uses the Linux kernel NTFS support
Partimage doesn't use the NTFS kernel driver. It's a user space NTFS implementation based on the publicly available NTFS information.- it's NTFS support is presumably as unstable as that in the Linux kernel (which claims to be quite immature). I didn't test it in deployment, but I didn't figure it was worth it. Some previous poster told me it worked for him, though.
There are at least 4 different NTFS implementations with different functionality, quality and license for Linux. Which one are you talking about?- And he told it to me without being an ass about it. Learn some respect. Read the post before commenting, please.
It's better to check the facts- Do you realize how that development is done? Hex editor on the partition, make an operation, hex edit, log and analyze the result...
No. That's over, every important thing is known and documented on the linux-ntfs.sf.net site. The lack of developers' free time is the problem.It's free, open source and the Linux-NTFS developers say it's reliable: http://linux-ntfs.sourceforge.net/man/ntfsclone.ht ml
Could somebody check fsutil fsinfo ntfsinfo DRIVE: on Longhorn? It should tell the NTFS version. Linux knows 1.2, 3.0 and 3.1.
I'm sure you messed up the only thing you could, the partition table manipulation. This is filesystem independent and tools like parted, QTParted and DiskDrake handle the entire process a user friendly way. You used the wrong, too low level tools to do the job.
About its safeness: "Since July of 2002, when ntfsresize became publicly available, there were many success reports for both enlarging and shrinking Windows XP/2000/NT4 and Windows Server 2003 NTFS filesystems on both workstation and server versions (Home, Professional, Server, Advanced Server). No destroyed filesystem was reported who followed the instructions correctly".
Do you think they and Mandrake would release it as stable otherwise? Also note, the Linux-NTFS team is not the one who wrote the well known broken NTFS driver, that can destroy data. This is also explained on the above pages. There are two NTFS drivers and most distros use their driver. They also list who use which driver.
There are other myths rebuted and explained on their pages, IMHO it's worth a look.
They don't need anymore. The problem is lack of time and ...
Also, NTFS is a very complex filing system, with many different versions. You don't want to get that wrong.
Very true.
Resizing was a more important goal, and that has been working for many months now.
Agree and based on the ntfsresize FAQ, actually it's just for one year.
Quite unlikely. If you read the linux-ntfs project we pages you can find that write is disabled for XP in the *old*, unmaintained driver and the limited write support is said to be safe using the *new* NTFS driver developed by the project. There is also a short list there what NTFS driver vendors use.