zfs discuss - May 2008 - The ZFS inventor and Linus sitting in a tree?

Here''s a link to a recent blog entry of Jeff Bonwick, lead engineer of
ZFS, showing him with Linus Torvalds, making mysterious comments in a
blog post that''s tagged ZFS.

I hate to be a scaremongerer, but are we about to lose one major
advantage over the "competition"?

I mean, if the Linux folks to want it, fine. But if Sun''s actually
helping with such a possible effort, then it''s just shooting itself in
the foot here, in my opinion.

Regards,
-mg

> Here''s a link to a recent blog entry of Jeff Bonwick, lead
engineer of
> ZFS, showing him with Linus Torvalds, making mysterious comments in a
> blog post that''s tagged ZFS.
Well, here''s the link, anyhow. :S

http://blogs.sun.com/bonwick/entry/casablanca

-mg

On Sunday 18 May 2008 2:30:30 pm Mario Goebbels wrote:
> > Here''s a link to a recent blog entry of Jeff Bonwick, lead
engineer of
> > ZFS, showing him with Linus Torvalds, making mysterious comments in a
> > blog post that''s tagged ZFS.
>
> Well, here''s the link, anyhow. :S
>
> http://blogs.sun.com/bonwick/entry/casablanca
No matter what pitch of the listservs happens to be, it''s good to know
that
everyone is still friends.

-Luke

On May 18, 2008, at 14:01, Mario Goebbels wrote:
> I mean, if the Linux folks to want it, fine. But if Sun''s actually
> helping with such a possible effort, then it''s just shooting
itself in
> the foot here, in my opinion.

As it is, patents and licenses prevent Linux from picking up ZFS, but  
if they were to re-license ZFS under GPL3 or grant a linux project a  
patent license, it could be possible.

I''ve been impressed by the development process ethic under  
OpenSolaris since joining the ZFS list.  There appears to be a real  
commitment to solving root-cause problems, which linux can sometimes  
ignore.  Timeliness could always be better, but such is life.  So,  
[Open]*Solaris has a good argument for it; e.g. real support contracts.

I''ve approached some of my linux-based clients about ZFS and
they''re
reluctant to add another OS to their mix.  I''m happy (and frankly  
prefer) hopping around OS''s, but not everybody is.  ZFS on Linux on  
Thumper would actually be very interesting to many of them.  I think  
that''s good for Sun.  Of course, ZFS on Linux on commodity hardware  
is also interesting to some, but that can be done on FreeBSD today.

Either way, if ZFS doesn''t find its way to Linux, somebody is going  
to re-implement something like it, and then Sun will have to sue  
Linux developers, which really sucks for everybody involved (and  
they''re quick to do it - they threatened to sue me when they
couldn''t
figure out how to take back a try-out server).  Having ZFS as a de- 
facto standard lifts all boats, IMHO.

-Bill
-----
Bill McGonigle, Owner           Work: 603.448.4440
BFC Computing, LLC              Home: 603.448.1668
bill at bfccomputing.com           Cell: 603.252.2606
http://www.bfccomputing.com/    Page: 603.442.1833
Blog: http://blog.bfccomputing.com/
VCard: http://bfccomputing.com/vcard/bill.vcf

On Mon, May 19, 2008 at 10:06 PM, Bill McGonigle <bill at
bfccomputing.com> wrote:
> On May 18, 2008, at 14:01, Mario Goebbels wrote:
>
>> I mean, if the Linux folks to want it, fine. But if Sun''s
actually
>> helping with such a possible effort, then it''s just shooting
itself in
>> the foot here, in my opinion.
>
>
[....]
> they''re quick to do it - they threatened to sue me when they
couldn''t
> figure out how to take back a try-out server).
There''s a story contained within that for sure! :) You brought a smile
to this subscriber when I read it.

> Having ZFS as a de- facto standard lifts all boats, IMHO.
It''s still hard to believe (in one sense) that the entire world
isn''t
beating a path to Sun''s door and PLEADING for ZFS. This is (if
y''all
will forgive the colloquialism) a kick-ass amazing piece of software.
It appears to defy all the rules, a bit like levitation in a way, or
perhaps it just rewrites those rules. There are days I still can''t get
my head around what ZFS really is.

In general, licensing issues just make my brain bleed, but one hopes
that the licensing gurus can get their heads together and find a way
to get this done. I don''t personally believe that Open Solaris *OR*
Solaris will lose if ZFS makes its way over the fence to Linux, I
think that this is a big enough tent for everyone. Sure hope so
anyway, it would be immensely sad to see technology like this not
being adopted/ported/migrated/whatever more widely because of "damn
lawyers" and the morass called licensing.

Perhaps (gazing into a cloudy crystal ball that hasn''t been cleaned in
a while) Solaris/Open Solaris can manage to hold onto ZFS-on-boot
which is perhaps *the* most mind bending accomplishment within the zfs
concept, and let the rest procreate elsewhere. That could contribute
to the "must-have/must-install" cachet of Solaris/OpenSolaris.

Umm, my uninspiring and non-expert contribution to this (unusually)
non-expert thread. Thanks to all involved on this list, sometimes it
seems like every post is a kind of mini tutorial all on its own.

Warm Regards and thanks for all the fish/knowledge nuggets
-Colin
--
Colin J. Raven
"A wide-eyed neophyte staring through the window at the wizards toiling
within"

On Mon, 19 May 2008, Bill McGonigle wrote:
> As it is, patents and licenses prevent Linux from picking up ZFS, but
> if they were to re-license ZFS under GPL3 or grant a linux project a
> patent license, it could be possible.
I don''t believe that either of these is sufficient.  The Linux kernel 
is GPLv2 but with the allowance to use "any later version" removed so 
it does not appear to be compatible with GPLv3.  GPLv2 is compatible 
with GPLv3 by being upgraded to GPLv3.  GPL does not care if patent 
rights are granted for some specific usage situation since it must 
apply for all possible usage situations.

While Linus may be the de-facto Linux spokesperson, he does not hold 
the many thousands of copyrights to Linux so he does not "own" the 
work.  It is a hopeless case.  The FSF was a bit wiser since they made 
sure that everyone who contributed to their projects signed over the 
copyrights to them (a tedious process to be sure).

Bob
=====================================Bob Friesenhahn
bfriesen at simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/

On May 19, 2008, at 16:06, Bill McGonigle wrote:
> As it is, patents and licenses prevent Linux from picking up ZFS, but
> if they were to re-license ZFS under GPL3 or grant a linux project a
> patent license, it could be possible.
What does the CDDL say about patents?
The CDDL provides an explicit patent license for code released under  
the license. This means that you can use, modify, and redistribute  
code released under CDDL without worrying about any patents that the  
contributors of the code (including Sun) might have on the  
contributed technology. The license also includes a provision to  
discourage patent litigation against developers by revoking the  
rights to the code for anyone initiating a patent claim against a  
developer regarding code they have contributed.

http://opensolaris.org/os/about/faq/licensing_faq/#patents

Bill McGonigle <bill at bfccomputing.com> wrote:
> On May 18, 2008, at 14:01, Mario Goebbels wrote:
>
> > I mean, if the Linux folks to want it, fine. But if Sun''s
actually
> > helping with such a possible effort, then it''s just shooting
itself in
> > the foot here, in my opinion.
>
>
> As it is, patents and licenses prevent Linux from picking up ZFS, but  
> if they were to re-license ZFS under GPL3 or grant a linux project a  
> patent license, it could be possible.
This is a claim from the Linux camp.

If it was correct, then neither FreeBSD nor Mac OS X could use ZFS.

The GPL does not prevent non-GPL code to appear in a GPLd project. The GPL only 
prevents GPLd code to appear inside a non-GPL project.

J?rg

-- 
 EMail:joerg at schily.isdn.cs.tu-berlin.de (home) J?rg Schilling D-13353 Berlin
       js at cs.tu-berlin.de                (uni)  
       schilling at fokus.fraunhofer.de     (work) Blog:
http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily

Bob Friesenhahn <bfriesen at simple.dallas.tx.us> wrote:
> While Linus may be the de-facto Linux spokesperson, he does not hold 
> the many thousands of copyrights to Linux so he does not "own"
the
> work.  It is a hopeless case.  The FSF was a bit wiser since they made 
> sure that everyone who contributed to their projects signed over the 
> copyrights to them (a tedious process to be sure).
Linus together with a "few" other people could change the license for
Linux.

He only needs to make sure that these people cover more than 50% of the code 
and that no single contributor who owns more than 5% of the whole source 
remains unasked.

It seems that less than 50 people need to agree on the license change.

J?rg

-- 
 EMail:joerg at schily.isdn.cs.tu-berlin.de (home) J?rg Schilling D-13353 Berlin
       js at cs.tu-berlin.de                (uni)  
       schilling at fokus.fraunhofer.de     (work) Blog:
http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily

On May 19, 2008, at 17:31, David Magda wrote:
> you can use, modify, and redistribute code released under CDDL  
> without worrying about any patents
On May 19, 2008, at 18:12, Joerg Schilling wrote:
> If it was correct, then neither FreeBSD nor Mac OS X could use ZFS.
Somebody correct me if I''m wrong, but AFAIK, both FreeBSD and OSX use  
a non-trival portion of the CDDL''ed code directly (FreeBSD  
implemented a Solaris compatibility layer so they could do this,  
according to the paper on ZFS on FreeBSD) so the patent rights are  
conferred because the CDDL''ed code is used.  Since they''re
both BSD,
the licenses are compatible.

A cleanroom implementation would not use the CDDL''ed code, and so the  
patent rights wouldn''t flow.  It would be neat if the CDDL said you  
could learn from the CDDL''ed code and have patent rights confer, but  
I don''t believe such a clause exists.  Clean-rooming ZFS might be  
easier than tackling GPLv3.

On May 19, 2008, at 18:16, Joerg Schilling wrote:
> He only needs to make sure that these people cover more than 50% of  
> the code
> and that no single contributor who owns more than 5% of the whole  
> source
> remains unasked.
That would be convenient, but do you have cite (it would help my  
understanding of this morass)?  I don''t see where the GPLv2 trumps  
Berne Convention copyright.  The *GNU Project* code does require  
copyright assignment, for this reason (turns out they were on to  
something, IMHO).

I have heard people say that by this time we know where X% (where X  
is very large) of the Linux code came from, and re-writing the  
unattributed parts would not be an impossible task, should GPLv3 be  
desired.  My guess is if 47 of the 50 proverbial core people were on  
board the other 3''s code would get re-done.  Unless it were e.g.,  
Ingo, Alan, Linus.

In my dream world Jonathan dispatched Jeff to talk to Linus and  
broach, "we''ll do it if you do it". :)  If so we can mark the
date on
the calendar when digital cameras come with ZFS flash cards instead  
of FAT32! (I''m only half kidding)

-Bill

-----
Bill McGonigle, Owner           Work: 603.448.4440
BFC Computing, LLC              Home: 603.448.1668
bill at bfccomputing.com           Cell: 603.252.2606
http://www.bfccomputing.com/    Page: 603.442.1833
Blog: http://blog.bfccomputing.com/
VCard: http://bfccomputing.com/vcard/bill.vcf

Frankly, while I love Linux for a bunch of things (it runs my desktop, 
amongst other things), I really think the impetus of "Geeze, can''t
we
get Linux to have THAT" every time some interesting thing comes along is 
a BAD IDEA.

Having a large ecosystem of OSes which can cross-germinate ideas, but 
not necessarily code, is a Good Thing. However, trying to make one OS do 
everything well is a recipe for disaster, let alone code bloat.  It''s 
like trying to design a car that: (a) does 0-60mph under 3 seconds (b) 
can haul 2 tons of cargo (c) get 50mpg and (d) cost under $15k.

Not going to happen, and the compromises that you have to make to 
achieve certain goals make other attributes either difficult or 
impossible to accomplish.   OSes should have a "optimum" targeted user
population - one that  gain the most benefit from the OSes'' strengths, 
and are hurt least by it''s weaknesses.   Trying for a One True OS is a 
silly (and dangerously shortsighted) goal.

In ZFS''s case, if the ideas around it are so attractive, then 
reimplementing the ideas in a Linux-compatible FS module are entirely 
possible, even if certain patents must be worked around. I''m not 100% 
sure, and I''m certainly not one of the lawyers, but my reading of the 
ZFS internals and related IP leaves me with the impression that a 
complete clean-room reimplementation is entirely possible, though it may 
not have the same performance profile.  That is, the ZFS on-disk format 
isn''t IP protected, and the general concepts of how ZFS works (pools, 
CoW, snapshots, etc) are open, it''s just _how_ the guts do these things
which are.

So, I''m actually pretty glad that ZFS will remain solely in the 
CDDL-compatible land, and that if a ZFS-compatible system shows up in 
GPL-land, it will have a different underlying engineering, which may 
provide many more interesting insights into improving what is truly the 
FS-for-the-new-millenium.


(Please, if I''m wrong about our [Sun''s] patent protection of
ZFS''s
internals, I want to know _now_. Speak up and correct me please, folks).

-- 
Erik Trimble
Java System Support
Mailstop:  usca22-123
Phone:  x17195
Santa Clara, CA
Timezone: US/Pacific (GMT-0800)

On May 20, 2008, at 03:09, Erik Trimble wrote:
> That is, the ZFS on-disk format isn''t IP protected, and the
general
> concepts of how ZFS works (pools, CoW, snapshots, etc) are open,  
> it''s just _how_ the guts do these things which are.

I''m also in the decidedly-not-a-lawyer camp too, but when I went and  
looked one time (at less than half of the patents - does anybody have  
a public list?) I think I saw things like on-disk file-tree  
representation, zero-fill and sparse file storage formats, multiple  
copies, extended attributes storage, the way you make snapshots, the  
way you express filesystems and pools hierarchically, the specific  
way write intents have to be done in ZFS - stuff like that.  Again,  
being neither a lawyer nor a ZFS dev, I came away thinking it might  
be possible to make a read-only version that would be OK, or if maybe  
a handful of the patents didn''t issue perhaps a low-performance  
version without some of the features.  But then nobody would want to  
do a thing like that. :)

Everything I''ve read from Sun folk indicates they''d like to
see ZFS
become ubiquitous, so I''m sure they''re going to figure
something out
sooner or later.  It''s worth remembering that some people are still  
waiting for what they consider essential features, so we''re really  
early in the game here, and if we''re still in the Cathedral stage,  
the status quo may be the best bet for now.

-Bill

-----
Bill McGonigle, Owner           Work: 603.448.4440
BFC Computing, LLC              Home: 603.448.1668
bill at bfccomputing.com           Cell: 603.252.2606
http://www.bfccomputing.com/    Page: 603.442.1833
Blog: http://blog.bfccomputing.com/
VCard: http://bfccomputing.com/vcard/bill.vcf

> On May 18, 2008, at 14:01, Mario Goebbels wrote:
>  ZFS on Linux on  
> humper would actually be very interesting to many of
> them.  I think  
> that''s good for Sun.  Of course, ZFS on Linux on
Umm, how many Linux shops buy support and/or HW from Sun ?

It it''s a Linux shop money is (in order) going to these people - IBM,
Redhat, Novell, Dell.

Those all are - technically - Sun competitors in some sphere.

If you consider software stacks, there are only 3 companies in the world with
complete SW stack - Sun, IBM and Microsoft

If you throw HW in the mix, there are only two - Sun and IBM.

Figuring out who''s printing money and who''s contributing most
code is left as an exercise to the reader.

- mritun
 
 
This message posted from opensolaris.org

Bill McGonigle <bill at bfccomputing.com> wrote:
> On May 19, 2008, at 17:31, David Magda wrote:
> > you can use, modify, and redistribute code released under CDDL  
> > without worrying about any patents
>
> On May 19, 2008, at 18:12, Joerg Schilling wrote:
> > If it was correct, then neither FreeBSD nor Mac OS X could use ZFS.
>
> Somebody correct me if I''m wrong, but AFAIK, both FreeBSD and OSX
use
> a non-trival portion of the CDDL''ed code directly (FreeBSD  
> implemented a Solaris compatibility layer so they could do this,  
> according to the paper on ZFS on FreeBSD) so the patent rights are  
> conferred because the CDDL''ed code is used.  Since
they''re both BSD,
> the licenses are compatible.
>
> A cleanroom implementation would not use the CDDL''ed code, and so
the
....

It seems that you are following the wrong idea that the CDDLd original code 
cannot be used for Linux directly. This is obviously wrong.

The GPL does not forbid GPLd code to use non-GPLd code from a GPLd project.
If this was not true, then the GPL would be completely unusable. It is bad to 
see that RMS in his talks always tells you what he _likes_ to do but never what 
the GPL really does (to make it compatible with reality).

Note that the biggest problem for using ZFS in Linux is not the license but the 
incompatible Linux filesystem interface.

> On May 19, 2008, at 18:16, Joerg Schilling wrote:
> > He only needs to make sure that these people cover more than 50% of  
> > the code
> > and that no single contributor who owns more than 5% of the whole  
> > source
> > remains unasked.
>
> That would be convenient, but do you have cite (it would help my  
> understanding of this morass)?  I don''t see where the GPLv2 trumps
> Berne Convention copyright.  The *GNU Project* code does require  
> copyright assignment, for this reason (turns out they were on to  
> something, IMHO).
Looks like a missunderstanding of Copyright...

The European Copyright law explicitely contains a provision that disallows 
minority contributors to control the way of marketing. In an OSS environment, 
this means that you still pave to pay the part of the income (a fraction of 0)
to minority contributors but they cannot control the license.

BTW: Simon Phipps told me last year that the wife of Eric Raymond (beeing a 
lawywer) recently wrote an article that claims that the same applies to US 
Copyright and that for this reason, Linus _could_ change the license together  
with a few additional people.

J?rg

-- 
 EMail:joerg at schily.isdn.cs.tu-berlin.de (home) J?rg Schilling D-13353 Berlin
       js at cs.tu-berlin.de                (uni)  
       schilling at fokus.fraunhofer.de     (work) Blog:
http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily

> On Mon, May 19, 2008 at 10:06 PM, Bill McGonigle
> <bill at bfccomputing.com> wrote:
> > On May 18, 2008, at 14:01, Mario Goebbels wrote:
> >
> >> I mean, if the Linux folks to want it, fine. But
> if Sun''s actually
> >> helping with such a possible effort, then it''s
> just shooting itself in
> >> the foot here, in my opinion.
> >
> >
> 
> [....]
> > they''re quick to do it - they threatened to sue me
> when they couldn''t
> > figure out how to take back a try-out server).
> 
> There''s a story contained within that for sure! :)
> You brought a smile
> to this subscriber when I read it.
> 
> 
> > Having ZFS as a de- facto standard lifts all boats,
> IMHO.
> It''s still hard to believe (in one sense) that the
> entire world isn''t
> beating a path to Sun''s door and PLEADING for ZFS.
> This is (if y''all
> will forgive the colloquialism) a kick-ass amazing
> piece of software.
> It appears to defy all the rules, a bit like
> levitation in a way, or
> perhaps it just rewrites those rules. There are days
> I still can''t get
> my head around what ZFS really is.
> 
> In general, licensing issues just make my brain
> bleed, but one hopes
> that the licensing gurus can get their heads together
> and find a way
> to get this done. I don''t personally believe that
> Open Solaris *OR*
> Solaris will lose if ZFS makes its way over the fence
> to Linux, I
> think that this is a big enough tent for everyone.
> Sure hope so
> anyway, it would be immensely sad to see technology
> like this not
> being adopted/ported/migrated/whatever more widely
> because of "damn
> lawyers" and the morass called licensing.
> 
> Perhaps (gazing into a cloudy crystal ball that
> hasn''t been cleaned in
> a while) Solaris/Open Solaris can manage to hold onto
> ZFS-on-boot
> which is perhaps *the* most mind bending
> accomplishment within the zfs
> concept, and let the rest procreate elsewhere. That
> could contribute
> to the "must-have/must-install" cachet of
> Solaris/OpenSolaris.
Umm, I think it''s too late for that; as I recall, the bits needed for
read-only access had to be made dual CDDL/GPL to be linked with GRUB.
 
 
This message posted from opensolaris.org

"Colin Raven" <colin at clearcutnetworks.com> wrote:
> In general, licensing issues just make my brain bleed, but one hopes
> that the licensing gurus can get their heads together and find a way
> to get this done. I don''t personally believe that Open Solaris
*OR*
I see no license issue here. The problem is a cultural problem and the fact 
that the Linux kernel FS interface is very different from the one used in 
Solaris.

Soneone at the linux side would need to put effort in a port and I
don''t see
this to happen because of missing interest. From my understanding I wwould 
first like to see some kind of interest from the Linux camp.

J?rg

-- 
 EMail:joerg at schily.isdn.cs.tu-berlin.de (home) J?rg Schilling D-13353 Berlin
       js at cs.tu-berlin.de                (uni)  
       schilling at fokus.fraunhofer.de     (work) Blog:
http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily

> > On May 18, 2008, at 14:01, Mario Goebbels wrote:
> >  ZFS on Linux on  
> > humper would actually be very interesting to many
> of
> > them.  I think  
> > that''s good for Sun.  Of course, ZFS on Linux on
> 
> Umm, how many Linux shops buy support and/or HW from
> Sun ?
> 
> It it''s a Linux shop money is (in order) going to
> these people - IBM, Redhat, Novell, Dell.
Linux is supported on Sun equipment.  So is Windows.  A recent report show Sun
is the #3 vendor of x86 servers.

My shop just bought a Sun workstation to dualboot Windows and Solaris.

Regardless, ZFS is compelling enought that I''d like it everywhere.

Just like I like having vi everywhere.  Or perl or python or emacs....
 
 
This message posted from opensolaris.org

On Tue, 20 May 2008, Tom Buskey wrote:
> Regardless, ZFS is compelling enought that I''d like it everywhere.
Agreed--but not at the expense of changing its (ZFS'') license.

-- 
Rich Teer, SCSA, SCNA, SCSECA

CEO,
My Online Home Inventory

URLs: http://www.rite-group.com/rich
      http://www.linkedin.com/in/richteer
      http://www.myonlinehomeinventory.com

On Tue, 20 May 2008, Joerg Schilling wrote:
> The GPL does not forbid GPLd code to use non-GPLd code from a GPLd project.
> If this was not true, then the GPL would be completely unusable. It is bad
to
> see that RMS in his talks always tells you what he _likes_ to do but never
what
> the GPL really does (to make it compatible with reality).
You are correct.  As long as the source code for the complete work is 
distributed in terms which comply with GPL, it is ok.  Note that Linus 
(and others) have more reqirements for Linux than GPL actually 
requires. Debian Linux would never support this sort of blending.
> BTW: Simon Phipps told me last year that the wife of Eric Raymond (beeing a
> lawywer) recently wrote an article that claims that the same applies to US
> Copyright and that for this reason, Linus _could_ change the license
together
> with a few additional people.
It should be clear to any rational person that lawyers expressed 
opinions are not necessarily correct or legally significant.  If the 
opinions of lawyers were correct, then they would not be continually 
contesting each other in court.  One of the lawyers loses in every 
court case which comes to conclusion.  This means that lawyers are 
correct no more than 50% of the time so their opinions are no more 
valid than simply flipping a coin.

The only statement of any consequence is from a court decision.

Bob
=====================================Bob Friesenhahn
bfriesen at simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/

Bob Friesenhahn <bfriesen at simple.dallas.tx.us> wrote:
> On Tue, 20 May 2008, Joerg Schilling wrote:
>
> > The GPL does not forbid GPLd code to use non-GPLd code from a GPLd
project.
> > If this was not true, then the GPL would be completely unusable. It is
bad to
> > see that RMS in his talks always tells you what he _likes_ to do but
never what
> > the GPL really does (to make it compatible with reality).
>
> You are correct.  As long as the source code for the complete work is 
> distributed in terms which comply with GPL, it is ok.  Note that Linus 
> (and others) have more reqirements for Linux than GPL actually 
> requires. Debian Linux would never support this sort of blending.
It may be that you confuse the term "work" in trying to extend it in a
wrong way.

The reason why GPLd code may happily call non-GPLd code is because otherwise
GPLd code would be illegal on AIX, HP-UX or Cygwin (or other closed source 
platforms). This is what the FSF likes to avoid. Given the fact that the GPLv2 
does not mention the term "linking", it is obvious that there is no
difference
between static and dynamic linking. The GPLv2 is completely based on the term 
"work", so we have to be careful when interpreting whether/how the
"work" may
be expanded in a specific case. In order to allow GPLd software on CSS 
platforms (as mentioned above), it is obvious that the term "work"
cannot be
expanded in a way that it includes unrelated code. The only possible way to 
define the "work" seems to be in including only the code that was
_derived_
from the original GPLd code.

On CSS platforms, libc or similar are obviously not derived from GPLd code and 
thus cannot be part of "the work" altough libc code may appear in a
binary
created from a GPLd work.

Conclusion 1) is that the original CDDLd ZFS code is not part of "the
work"
"Linux".

Conclusion 2) is that if the ZFS port is done in a way that does not introduce 
code "derived" from the Linux kernel into ZFS, there will be no
problem with a
CDDLd ZFS used by Linux.



 J?rg

-- 
 EMail:joerg at schily.isdn.cs.tu-berlin.de (home) J?rg Schilling D-13353 Berlin
       js at cs.tu-berlin.de                (uni)  
       schilling at fokus.fraunhofer.de     (work) Blog:
http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily

On May 20, 2008, at 10:42 AM, Joerg Schilling wrote:
> Bob Friesenhahn <bfriesen at simple.dallas.tx.us> wrote:
>
>> ,,,
> It may be that you confuse the term "work" in trying to extend it
> in a wrong way.
...many wise words elided...

Not being a lawyer, and this not being a Legal forum ... can we leave  
license analysis alone?

-- 
Keith H. Bierman   khbkhb at gmail.com      | AIM kbiermank
5430 Nassau Circle East                  |
Cherry Hills Village, CO 80113           | 303-997-2749
<speaking for myself*> Copyright 2008

Keith Bierman wrote:
>
> Not being a lawyer, and this not being a Legal forum ... can we leave  
> license analysis alone?
>
>   
The GNU _itself_ states that it is not allowable in plain English. Why 
people continue to argue about it is beyond me :-)

Common Development and Distribution License (CDDL) 
<http://www.opensolaris.org/os/licensing/cddllicense.txt>

    This is a free software license. It has a copyleft with a scope
    that''s similar to the one in the Mozilla Public License, which
makes
    it incompatible with the GNU GPL
    <http://www.gnu.org/licenses/gpl.html>. This means a module covered
    by the GPL and a module covered by the CDDL cannot legally be linked
    together. We urge you not to use the CDDL for this reason.

    Also unfortunate in the CDDL is its use of the term "intellectual
    property <http://www.gnu.org/philosophy/not-ipr.html>".


(from http://www.gnu.org/licenses/license-list.html#SoftwareLicenses)

-Brian


-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20080520/e23b7b3f/attachment.html>

"Brian H. Nelson" <bnelson at cis.ysu.edu> wrote:
> Keith Bierman wrote:
> >
> > Not being a lawyer, and this not being a Legal forum ... can we leave
> > license analysis alone?
> >
> >   
>
> The GNU _itself_ states that it is not allowable in plain English. Why 
> people continue to argue about it is beyond me :-)
This claim is in conflict with other claims from the FSF and in conflice
with claims made by Eben Moglen (professor of law and laywer from the FSF).


So finally, please stop this license talk.

J?rg

-- 
 EMail:joerg at schily.isdn.cs.tu-berlin.de (home) J?rg Schilling D-13353 Berlin
       js at cs.tu-berlin.de                (uni)  
       schilling at fokus.fraunhofer.de     (work) Blog:
http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily

On Tue, 20 May 2008, Joerg Schilling wrote:
>
> The reason why GPLd code may happily call non-GPLd code is because
otherwise
> GPLd code would be illegal on AIX, HP-UX or Cygwin (or other closed source
> platforms). This is what the FSF likes to avoid. Given the fact that the
GPLv2
It seems that it is time for you to refresh your memory of GPLv2:

   "However, as a special exception, the source code distributed need 
not include anything that is normally distributed (in either source or 
binary form) with the major components (compiler, kernel, and so on) 
of the operating system on which the executable runs, unless that 
component itself accompanies the executable."

This is why GPL programs can use a closed-source libc and other 
libraries which are standard in an operating system.

While GPL does use the term "work", it used to describe a product 
based on the "Program".  The "Program" is the part protected
for
redistribution under GPL.
> Conclusion 2) is that if the ZFS port is done in a way that does not 
> introduce code "derived" from the Linux kernel into ZFS, there
will
> be no problem with a CDDLd ZFS used by Linux.
I have not read CDDL, but if it includes terms which revoke the 
recipient''s rights to use the source code or Program (as was 
previously mentioned here), then it is not compatible with GPL 
distribution requirements.  Failure to adhere to GPL distribution 
requirements simply revokes the right to distribute a work based on 
the Program, and never to revoke the right to use the Program.  The 
user of a GPLed Program is under no obligation to the author of that 
Program other than to similarly provide the source code if he cares to 
share the Program with someone else.

Hint: If the Linux kernel (a "work") was to do something which 
violates GPL, then it can no longer be distributed.  That is why the 
Linux folks are so worried about this ZFS "poison pill".

Bob
=====================================Bob Friesenhahn
bfriesen at simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/

Bob Friesenhahn <bfriesen at simple.dallas.tx.us> wrote:
> On Tue, 20 May 2008, Joerg Schilling wrote:
> >
> > The reason why GPLd code may happily call non-GPLd code is because
otherwise
> > GPLd code would be illegal on AIX, HP-UX or Cygwin (or other closed
source
> > platforms). This is what the FSF likes to avoid. Given the fact that
the GPLv2
>
> It seems that it is time for you to refresh your memory of GPLv2:
>
>    "However, as a special exception, the source code distributed need 
I am sorry top tell you that you are wrong: Ask Eben Moglen (law professor) for 
help, he explained 2-3 years ago why this "exception" is not needed...

But please finally stop this thread. It does not belong here.