Public domain filesystems

[Perica]

..

[AR]

FAT is considered "public domain" (Microsoft recently moved to patent it but I don't think it worked). I'm not sure about EXT*FS, but since it's in Linux you can probably assume it's freely avaliable.

As for Reiser, I'm pretty sure the concepts are public but the only official working implementation is for the Linux Kernel (Under GPL). There are some unofficial docs here though (The docs claim to be under GPL, but GPL only applies to code, so you may not want to copy and paste if that's a problem):
http://www.p-nand-q.com/download/rfstoo ... ?the_id=57

[mystran]

Can filesystem design be copyrighted? They are pretty much the same thing as fileformat. Can fileformats be copyrighted? The code to handle them can, but can the structure itself?

Now, obviously you might be able (at least in some parts of the world) patent some method used in the filesystem, but then it's not about copyrights anymore. Does public domain help if it's patented?

And finally, at least in Europe there are additional laws for protecting interoperability. How do those affect the whole mess?

[Brendan]

Hi,

I'm not a lawyer of any kind, so I'm probably mostly wrong here, but..

Can filesystem design be copyrighted? They are pretty much the same thing as fileformat. Can fileformats be copyrighted? The code to handle them can, but can the structure itself?

I'd say file system designs and file formats can't be copyrighted. They can use trademarks (e.g. "NTFS (tm)") to protect the name, and patents to protect the fundamental design (or parts of it). The reverse would apply for implementation - code can be copyrighted but not patented.

Now, obviously you might be able (at least in some parts of the world) patent some method used in the filesystem, but then it's not about copyrights anymore. Does public domain help if it's patented?

It depends. AFAIK if there's adequate proof that the method existed in the public domain before it was patented then the patent itself (or at least part of it) is invalid - you're only supposed to be able to patent new methods that aren't obvious. The legal definition of "adequate proof" and "obvious" probably aren't the same as my definitions though. If the method was patented before it was used in something that's in the public domain, then people can be sued for any loss this may have caused the patent holder (which includes whoever tried to put it in the public domain and anyone that borrowed the idea from the public domain).

And finally, at least in Europe there are additional laws for protecting interoperability. How do those affect the whole mess?

I want to design and manufacture a device for monitoring the flight path of nuclear ICBM's, do you think these laws would allow me to have access to the design details for the ICBM flight controllers in question (for interoperability purposes)?

I think the interoperability laws might be directly related to anti-monopoly (or fair trade practices) laws - I dont know.


Cheers,

Brendan

[Pype.Clicker]

Well, i'll not be taking patented FS into account here, but

- Filesystem code *can* be copyrighted, much like any code. This means there are rules on how you may use the sourcecode (and the finished product), such like what GPL does.

- Filesystem documents can be copyrighted, much like any book/document can be. In other words, the explanations (as a sequence of carefully chosen words and diagrams) are protected.

But *ideas* cannot be copyrighted per se. So you're free to re-implement a GPL'd FS provided your code and documentation doesn't plagiates existing GPL'd code & documentation.

[Solar]

ReiserFS is "free" if your OS is "free" (using "free" in the GNU sense, i.e. GPL or compatible), but has to be licensed if you want your OS to sell. Means, your ReiserFS code can't be put into PD.

[Candy]

ReiserFS is "free" if your OS is "free" (using "free" in the GNU sense, i.e. GPL or compatible), but has to be licensed if you want your OS to sell. Means, your ReiserFS code can't be put into PD.

If and only if you abided by the rules of the license... If you decide not to accept the license and to reverse-engineer the FS, then implement it, you can spread that as PD. But conversely, you can't be sure it works and you can't call it ReiserFS.

Note, check on the StarFS project Pype & I started, it's intended to be public domain (afaik).

[df]

The UFS's are bsd'd so you can essentialy do as you will with them.
Im not sure what copyright/license minix FS is under...

[Pype.Clicker]

[Note, check on the StarFS project Pype & I started, it's intended to be public domain (afaik).

reinforcing that (PD docs, at least, are guaranteed. any implementor will thus be free to choose the licensing that better suits his convictions)

[Perica]

..

[Pype.Clicker]

http://clicker.sourceforge.net/wiclicke ... php/StarFS

we're still not very far in the design, though