OSDev.org

Just read this article and found it a very interesting read, as here is discussions about file systems quite frequently I wonder what you people have to say about it.

http://blogs.msdn.com/b/b8/archive/2012 ... -refs.aspx

I just got to the second sentence... then...
LIE, LIE and LIE.

Also one of the goals are:
It is doomed from the start...

I don't even have to read it. You can bet on Microsoft "protecting" (or rather, "poisoning") any core technology of theirs with patents / non-disclosure, making it next-to-worthless for our needs. (Worse: in order to properly interact with Windows systems, you have to somehow support it. See NTFS...)

ok, ok, Microsoft is evil
But what about the design and theory of it?

Jezze wrote:

Code: Select all

Today, NTFS is the most widely used, advanced, and feature rich file system in broad use

LIE, LIE and LIE.

Not so much when you read "in broad use." It may be obvious MS propaganda but it does have some truth to it.

Anyway, one of the biggest technical changes from NTFS seems to be their new allocate-on-write model. Instead of updating metadata in place and keeping a journal, it just writes new metadata to a newly allocated chunk of the disk, which apparently helps avoid problems when writes are interrupted.

They also look like they're trying to do things like chkdisk without taking the file system offline, which is pretty interesting.

daemoned wrote:But what about the design and theory of it?

What about it? Even if it's great, you aren't allowed to copy it.

Rusky wrote:

Jezze wrote:

Code: Select all

Today, NTFS is the most widely used, advanced, and feature rich file system in broad use

LIE, LIE and LIE.

Not so much when you read "in broad use." It may be obvious MS propaganda but it does have some truth to it.

Actually, it has to be FAT because every device you carry that does not care about whether you plug it into a mac or any other desktop runs that.

FAT isn't exactly "advanced" or "feature rich" though.

Sounds like it's going to be another ZFS - a great concept, but stifled with patents so there can't ever be a legal F/OSS implementation.

Synon wrote:Sounds like it's going to be another ZFS - a great concept, but stifled with patents so there can't ever be a legal F/OSS implementation.

Yes, it's too bad that ZFS is so stifled with patents that there can't be a legal implementation, even the official implementation must be illegal because they use their own patents - wait, what?

fronty wrote:

Synon wrote:Sounds like it's going to be another ZFS - a great concept, but stifled with patents so there can't ever be a legal F/OSS implementation.

Yes, it's too bad that ZFS is so stifled with patents that there can't be a legal implementation, even the official implementation must be illegal because they use their own patents - wait, what?

Maybe I'm mistaking F/OSS for GPL-compatible but ZFS can't be implemented in (and distributed with) Linux because of the patents:

kerneltrap.org wrote:A recent discussion on the lkml examined the possibility of a Linux implementation of Sun's ZFS. It was pointed out that the file system is released under the GPL-incompatible CDDL, and that Sun has filed numerous patents to prevent ZFS from being reverse engineered. Max Yudin pointed out, "according to Jeff Bonwick's blog Sun issued 56 patents on ZFS, but I have no idea what they patented. Sorry, binary compatible ZFS reimplementation with GPL license might not be legal."

You can port it to Linux and distribute it legally, but it can't be distributed with the kernel because of the patents and the CDDL (according to Theodore T'so they probably chose the CDDL specifically because it is incompatible with the GPL).


[edit] At any rate, I was saying that ReFS would be so stifled with patents that a F/OSS version would be illegal, not ZFS. ZFS is heavily patented but obviously F/OSS versions aren't illegal.

Synon wrote:Maybe I'm mistaking F/OSS for GPL-compatible but ZFS can't be implemented in (and distributed with) Linux because of the patents:

kerneltrap.org wrote:A recent discussion on the lkml examined the possibility of a Linux implementation of Sun's ZFS. It was pointed out that the file system is released under the GPL-incompatible CDDL, and that Sun has filed numerous patents to prevent ZFS from being reverse engineered. Max Yudin pointed out, "according to Jeff Bonwick's blog Sun issued 56 patents on ZFS, but I have no idea what they patented. Sorry, binary compatible ZFS reimplementation with GPL license might not be legal."

You can port it to Linux and distribute it legally, but it can't be distributed with the kernel because of the patents and the CDDL (according to Theodore T'so they probably chose the CDDL specifically because it is incompatible with the GPL).

CDDL is still even approved by FSF, so CDDL-lincensed ZFS implementation isn't just open source but free software. Yes, it isn't compatible with GPL, but still it's open source.

The license is the only issue Linux developers are concerned about, it's not "GPL compatible".. so they whine.

Quite frankly, they're hypocrites. They often have no problem using permissively licensed code, but only after they wrap the GPL around their changes.

Software patents are some made up American problem, Linux undoubtedly violates quite a few of them. FreeBSD seems to be having no problems distributing ZFS.

God damnit! I was going to impliment a allocate on demand file system into my w.i.p. OS, but I guess I can't now. Wow microsoft, WOW.

Synon wrote:

fronty wrote:

Synon wrote:Sounds like it's going to be another ZFS - a great concept, but stifled with patents so there can't ever be a legal F/OSS implementation.

Yes, it's too bad that ZFS is so stifled with patents that there can't be a legal implementation, even the official implementation must be illegal because they use their own patents - wait, what?

Maybe I'm mistaking F/OSS for GPL-compatible but ZFS can't be implemented in (and distributed with) Linux because of the patents:

kerneltrap.org wrote:A recent discussion on the lkml examined the possibility of a Linux implementation of Sun's ZFS. It was pointed out that the file system is released under the GPL-incompatible CDDL, and that Sun has filed numerous patents to prevent ZFS from being reverse engineered. Max Yudin pointed out, "according to Jeff Bonwick's blog Sun issued 56 patents on ZFS, but I have no idea what they patented. Sorry, binary compatible ZFS reimplementation with GPL license might not be legal."

You can port it to Linux and distribute it legally, but it can't be distributed with the kernel because of the patents and the CDDL (according to Theodore T'so they probably chose the CDDL specifically because it is incompatible with the GPL).


[edit] At any rate, I was saying that ReFS would be so stifled with patents that a F/OSS version would be illegal, not ZFS. ZFS is heavily patented but obviously F/OSS versions aren't illegal.

Patents don't protect against reverse engineering software for the purpose of program to program interoperability in the United States. This is the reason projects like ReactOS can exist legally without licenses from Microsoft. Even if Microsoft patents every last bit of ReFS, then from my understanding you can still reverse engineer it and implement it as is. I'm no patent lawyer though, so I could be totally off base though.