back to article It's a no to ZFS in the Linux kernel from me, says Torvalds, points finger of blame at Oracle licensing

Linux kernel jockey, Linus Torvalds, has taken time out to remind open source loyalists that he is no fan of the ZFS file system due, in part, to the sometimes tortuous nature of open source licensing. Torvalds was responding to a question late last week regarding a recent update to the Linux kernel breaking the third party …

  1. jms222

    Hypocritical

    Linux, quite happy to thieve anything from anybody when it's convenient then try to place a viral license on it.

    I need absolutely robust non-rotting fileservers for my customers so it's simply FreeBSD (or FreeNAS) and ZFS. I dare say Solaris would cut it too.

    1. Chris 15
      FAIL

      Viral eh?

      So making sure that the main tenets of the gpl are maintained to the benefit of all, whilst stopping people and organisations getting a free lunch from others' work is a problem for you is it?

      1. Korev Silver badge
        Linux

        Re: Viral eh?

        So making sure that the main tenets of the gpl are maintained to the benefit of all, whilst stopping people and organisations getting a free lunch from others' work is a problem for you is it?

        Yes and no; see the open source software vendors getting rather annoyed by the hyperscalers making money from their software and then paying nothing in return. See MongoDB and Redis for examples.

    2. Captain Scarlet Silver badge

      Re: Hypocritical

      I would like to know where you have information on Linux stealing other peoples code (Obviously from the downvotes not many people agree).

      1. Graham Dawson Silver badge

        Re: Hypocritical

        His arse.

    3. Avatar of They
      Trollface

      Re: Hypocritical

      Oooohhh... Linux = theft. That is a new angle I haven't heard before. Do you work for Oracle or M$?

      Please provide evidence though, I love a good popcorn moment - All those fat lawyers could jump up and down on Linux and sue .... Somebody over it. Heck Oracle could sue Oracle as they are a supporter of Linux, that would be ace..

      We need a 'feeding the troll' icon to go with the 'troll'' icon. Come on El-Reg I just filled in your survey and forgot to request more icons.

      1. Anonymous Coward
        Anonymous Coward

        Re: free = theft?

        Commercialism has misseducated so many people, I'd not be surprised if they really do believe they are in backwards land.

      2. R3sistance

        Re: Hypocritical

        Wait... don't MS and Oracle both have Linux Distributions now, sure for MS it's for their internal routing in Azure called SONiC, Oracle has Oracle Enterprise Linux (OEL) which was based on RHEL. I don't think either would want to call Linux as theft now given that they are both using and redistributing linux distros.

      3. red floyd

        Re: Hypocritical

        Neither. He works for Unxis (aka SCO)

        1. jgarbo
          Pirate

          Re: Hypocritical

          SCO? The Supreme Sewer (pun intended) of Computerland. I'm shocked...

      4. DuncanLarge Silver badge

        Re: Hypocritical

        Maybe he works for SCO?

        Tumbleweed....

        SCO?

        Another tumbleweed...

        Surely someone remembers SCO??

        Crickets...

        1. Doctor Syntax Silver badge

          Re: Hypocritical

          "Surely someone remembers SCO?"

          Yes. It made a really good but maybe overpriced SMB server OS before it turned into a litigation company.

          1. Alan Brown Silver badge

            Re: Hypocritical

            That was a different company. The SCO name changed hands a few times along the way and the SCO that started the litigation was not the Santa Cruz Operations that created SCO - which remained part of Novell at the time.

            In a further twist of irony, it turned out that SCO really dodnt have standing to litigate over alleged SCO violations as they didnt have full copyright licensed to them by Novell - and Novell weren't at all happy about SCO's antics or claims of ownership of SCO

            1. Michael Wojcik Silver badge

              Re: Hypocritical

              it turned out that SCO really dodnt have standing to litigate over alleged SCO violations as they didnt have full copyright licensed to them by Novell

              The question was actually whether Novell had sold the UNIX copyrights to Caldera, later renamed SCO Group.

              In 2007, the court issued a summary judgement saying Novell retained the copyrights. In 2010, the appeals court reversed, stating that "Agreement 2" between Novell and Caldera did not clearly reserve the copyrights to Novell. Later that year, a jury ruled Novell did in fact retain the copyrights.

              Subsequently The Attachmate Group bought Novell. They sold a bunch of Novell IP to a holding company controlled by Microsoft, but retained the UNIX copyrights.

              Then in 2016 Micro Focus bought Attachmate (well, it was one of those complicated reverse-trust merger whatchacallit things, but to a first approximation...). Currently, according to the latest information I can find, the copyrights are still assigned to Novell, as a division of Micro Focus. Woo!

              Of course the UNIX trademark was long ago transferred to The Open Group.

              1. Tom Paine

                Re: Hypocritical

                Why'm I having a Fags, Mags and Bags flashback?

        2. Anonymous Coward
          Anonymous Coward

          Re: Hypocritical

          Are they still going?

          1. sbt
            Devil

            Don't mention the (Unix) war.

            Shh. A few more mentions and the corpse will re-animate to haunt us all!

            1. eldakka

              Re: Don't mention the (Unix) war.

              Shh. A few more mentions and the corpse will re-animate to haunt us all!
              AFAIK, the case is still ongoing.

              According to this Feb 2018 Joint Status Report, all of SCO's copyright claims have been resolved (dismissed), however there are still some outstanding claims:

              • SCO's remaining claim:

                • (1) Unfair Competition
              • IBM’s remaining counterclaims:

                • (1) Contract Claim (Count I)
                • (2) Claims Relating to SCO’s Copying of IBM Code in Linux (Counts VI-VIII)
                • (3) Claims Concerning SCO’s Campaign to Create Fear, Uncertainty and Doubt About IBM’s Products and Services (Counts II-V)

              I haven't been able to find (after 2 minutes searching...) any update since then.

        3. Michael Wojcik Silver badge

          Re: Hypocritical

          This is the Register. Everyone remembers Final SCO, and many people remember Real SCO as well.

      5. Kiwi
        Black Helicopters

        Re: Hypocritical

        Please provide evidence though, I love a good popcorn moment

        If you ate a kiddy bag a week and limited your supply to every kernel of corn that has or will exist during the history of manking, I doubt there'd be enough popcorn to cover your waiting for actual evidence - even twisted paranoid conspiracy-nutter[1] levels of evidence!

        [1] Separate (though sometimes overlapping) entity from conspiracy theorists and further more from those who see things and sometimes think "the official story doesn't add up"

    4. Doctor Syntax Silver badge

      Re: Hypocritical

      For example?

    5. m4r35n357 Bronze badge

      Re: Hypocritical

      Sounds like slander to me . . . are you well represented?

      1. This post has been deleted by its author

        1. Alan Brown Silver badge

          Re: Hypocritical

          The law doesnt differentiate libel or slander - it's all defamation.

          And contrary to popular belief even in advanced civilised societies, truth is NOT an absolute defence against defamation. (Posting about Fred's expunged criminal record or that Karen was a prostitute 30 years ago being two obvious examples upheld in UK and USA courts despite the latter having plenty of evidence to back it (Karen left that behind long ago and is a respectable lady now) the former being part of warnings related to "Fred" engaging in the same behaviour that got him convicted 10 years ago)

          In less enlightened countries *ahem*France*ahem* you can even be found guilty of defamation of a public figure by way of posting a particularly robust opinion or criticism.

          Which all means this law is mostly used by those with deep pockets to gag those with less deep ones and only occasionally for the purposes for which it's apparently intended - and even then if the defamer has deep pockets the defamee seldom prevails.

          1. Hans 1

            Re: Hypocritical

            "Karen left that behind long ago and is a respectable lady now"

            Wow, ok, so a prostitute is not a respectable lady ? I care to disagree, shit, it is even legal in some parts of the world. In Amsterdam you can legally smoke a joint while having intercourse with a prostitute, provided you are a resident in the country and she allows you to smoke in her company.

            1. imanidiot Silver badge

              Re: Hypocritical

              No, afaik, you can't because sex workers are, well, workers, and are required by law to have a smoke free workplace. (as ridiculous as that sound, a sex worker would have to have a designated smoking spot outside her main "work" area. I doubt there's much inspection for this sort of thing though)

          2. Dal90

            Re: Hypocritical

            >And contrary to popular belief even in advanced civilised societies, truth is NOT an absolute defence against defamation. ... being two obvious examples upheld in UK and USA courts

            I can't speak for the home country of George Orwell, but in the U.S.A. truth is an absolute defense. Hard stop.

            See first paragraph, and the subsequent examples where even statements that are not absolutely true but only substantially true receive the same protection absent of malice.

            http://www.dmlp.org/legal-guide/substantial-truth

          3. Michael Wojcik Silver badge

            Re: Hypocritical

            Posting about Fred's expunged criminal record or that Karen was a prostitute 30 years ago being two obvious examples upheld in UK and USA courts

            Citation for a US action, please.

            Here's a source - and an actual lawyer, at that - who explicitly disagrees with you. "The truth is an absolute defense against a defamation action." He's writing about US law.

            1. Orv Silver badge

              Re: Hypocritical

              Ken White (aka Popehat) has also written at length (and often amusingly) about this topic.

              It's actually extremely hard to win a defamation case in the US. Truth is an absolute defense. Also, the statement has to be such that a reasonable person would conclude it was true -- which means satire and general puffery are protected.

              To give you an idea of how hard it is to win, the guy that Elon Musk publicly called a pedophile sued him for defamation and lost.

              1. Kiwi
                Holmes

                Re: Hypocritical

                To give you an idea of how hard it is to win, the guy that Elon Musk publicly called a pedophile sued him for defamation and lost.

                That's not exactly a fair case now is it? I mean for a start, peadoguy Musk is a hero of the American Public and all-round True American Man whereas the other person is some foreign paedo (he must be coz Musk said so and Musk is so great I think I just did some love-wees in my panties!).

                And the second issue is the quality of affordable judgeslawyers. PedoMusk could afford to buy any number of top judgeslawyers he wanted to, whereas the other guy didn't exactly have equal access to the same number let alone quality of lawyers.

                (Of course, saying PedoMusk purchased judges etc is NOT defamation - I have proof as the case went PedoMusk's way and that would not have happened in a truly fair hearing with unbiased judges!)

                1. Orv Silver badge

                  Re: Hypocritical

                  You MIGHT be overestimating how much affection the average American has for Elon Musk. While his business achievements are popular his personal behavior really is not.

          4. Jaybus

            Re: Hypocritical

            "The law doesnt differentiate libel or slander - it's all defamation."

            Yes, although handled differently in practice because libel is a written falsehood, whereas slander is a spoken falsehood. It is much easier to establish what exactly was conveyed when there is a written document involved, even when the slander occurs on an audio recording.

    6. Len

      Re: Hypocritical

      I don't agree on the thieving part but, yes, all my servers run FreeBSD precisely for this reason, ZFS is amazing.

      1. Alan Brown Silver badge

        Re: Hypocritical

        FreeBSD's network stack isn't quite as good as Linux' at high speeds/loads

        There are a couple of other reasons why people favour ZFS on Linux over FreeBSD or openslowaris, but it's not all just about personal OS preferences.

        As for CDDL - Linus is absolutely correct to avoid this mess until the license is changed. CDDL was specifically written from the ground up by Sun lawyers to be incompatible with any form of GPL because Linux was seen as a clear and present danger to Solaris. BSD blindsided them.

        The current setup is a little messy but as long as ZFS and Linux are kept separated until fused on the enduser box AND as long as hardware isn't traded with ZFS-on-Linux installed then it falls within the licensing limits. (Ie, they can only be combined on the enduser system AFTER the enduser has taken physical and financial possession of the system)

        1. chasil

          Re: Hypocritical

          Antergos actually did all of this Linux/ZFS meshing in their installer. I wrote about that here:

          https://www.linuxjournal.com/content/zfs-linux

          Unfortunately, maintenance for Antergos has apparently ended.

          https://itsfoss.com/antergos-linux-discontinued/

        2. Orv Silver badge

          Re: Hypocritical

          I'd think the overhead of FUSE would outweigh that, in the ZFS case, but I've never benchmarked it -- maybe FUSE is more efficient than I would think.

          FreeBSD's stack is highly consistent and stable, but you're right, my experience is it's not quite as fast. This may partly be interface-dependent, though -- FreeBSD's driver support is not as comprehensive as Linux's.

          As a sysadmin I really like FreeBSD, everything is straightforward and easy to maintain. It has a nice, clean init system, multiple good options for handling package updates, and I've never had a failed upgrade between releases. (RedHat, on the other hand, requires you to reinstall to upgrade.)

          1. HighTension

            Re: Hypocritical

            zfsonlinux isn't FUSE, it's a kernel module. I don't think anyone uses the FUSE implementation any more.

    7. Anonymous Coward
      Anonymous Coward

      Re: Hypocritical

      " need absolutely robust non-rotting fileservers for my customers"

      AWS then

      1. John Brown (no body) Silver badge

        Re: Hypocritical

        "AWS then"

        What makes you think AWS or even Amazon as a whole will still be around in 10 or 20 years? The roadside is littered with "big" names from the past of the IT industry.

    8. Anonymous Coward
      Anonymous Coward

      Re: Hypocritical

      "I need absolutely robust non-rotting fileservers for my customers "

      Ah, you should find an old Netware 6.5 install CD. Years and years of file serving with zero issues.

      1. Tom 7

        Re: Hypocritical

        Well not quite - it was always difficult to remember WTF you did 3 or 4 years ago last time the server re-booted by the new IT manager.

        1. MrReynolds2U

          Re: Hypocritical

          "it was always difficult to remember WTF you did 3 or 4 years ago last time the server re-booted"

          Did you have an unstable server? The only time we ever rebooted a Novell box was when the cleaners unplugged it. We found it under a pile of paperwork in a storage room. Probably still running now...

          1. Orv Silver badge

            Re: Hypocritical

            We had one offline overnight because the cleaning person accidentally pushed a chair against it in a way that held the RESET button down. We fabricated a mollyguard for the button and reliability was restored.

        2. Munchausen's proxy
          Pint

          Re: Hypocritical

          I was always entertained when we did an upgrade and counted all the abends that no one noticed.

    9. steelpillow Silver badge

      Re: Hypocritical

      Linux, quite happy to thieve

      As in:

      "Here, it's a free giveaway"

      "Oh, thanks, I'll have some of that"

      "Waaah! No fair! He stole my free giveaway!"

      Yes, boys and girls, the world is full of all sorts of plonkers. We all just met another one.

      But I am thinking of running a sweepstake on how long it will be before Oracle quote that post in an appeal of an appeal of an appeal of a lawsuit against Linux. I mean, how about suing Linux for not rolling in ZFS? Remember, girls and boys, it doesn't have to have even half a suggestion of a leg before Oracle's lawyers are willing to fling it.

      1. Bronek Kozicki

        Re: Hypocritical

        Oracle has absolutely nothing to do with OpenZFS and that is the only ZFS that matters, because it is used on Linux, BSD, MacOS and OpenIndiana (previously illumos).

        Oracle have their own closed-source ZFS which only f*ked over poor Oracle customers care about and which is basically an orphan anyway. They build their stack on RedHat, which as we know is based on ancient upstream kernels, so they are also largely resilient to changes in upstream.

    10. Tigra 07
      Pint

      Re: Hypocritical

      Steve Ballmer is that you?

      1. Anonymous Coward
        Anonymous Coward

        Re: Hypocritical

        Whoever it is, he's trying to break the downvote counter...

        1. Tigra 07
          Thumb Up

          Re: Hypocritical

          I've never seen so many downvotes on a non-deleted post. Let's hope it stays up for historical reference.

    11. Anonymous Coward
      Anonymous Coward

      Re: Hypocritical

      Nice try McBride.

    12. Lee D Silver badge

      Re: Hypocritical

      "Linux, quite happy to thieve anything from anybody when it's convenient then try to place a viral license on it."

      I think "thieving" ZFS would be extremely convenient. They won't touch it through *principle*, not convenience.

      For reference, many things could have been put into Linux for convenience but literally never made the grade. Go ask PaXTeam / grsecurity. It would have been very convenient to take the work that they wanted the kernel to pick up and just put it in.

      Instead, after over a decade, the Linux people *went out of their way* not to, because it wasn't done by the right procedure, and rolled their own solutions instead. For years, people were trying to get grsecurity into the kernel and with the licensing they could have just slapped it in any time they liked. They didn't. Instead they reinvented the wheel and put their own in.

      And they have literally no choice about the licence. They haven't had for decades. It's GPL-v2-only, now and forever, and cannot possibly change without rewriting huge portions of it all over again. It's a fait accompli, and there's nothing anyone can legally do to change it now because of all the work of dead, uncontactable or unwilling contributors meaning you can't ever change the licence. You can't even GPL-v3 it, it's that tightly bound to the code now.

    13. Denarius

      Re: Hypocritical

      is that you again, the rotting corpse of SCO ?

    14. Anonymous Coward
      Anonymous Coward

      Re: Hypocritical

      Nearly every single post in this thread has down votes despite them being sound posts. Larry, are you lurking? Come on, say hello... We won't bite (but you may end up, inexplicably, at the bottom of a lift shaft with a photocopier tied to you)

    15. Roger Kynaston
      Happy

      Re: Hypocritical

      You are Darl McBride and I claim my £5 ($6.49) at todays rates.

    16. Tom Paine

      Re: Hypocritical

      Ahhhhh thank god, it's been a month since the election and I was jonesing for a good holy war.

  2. Rich 2 Silver badge

    The problem is not Oracle (for once)

    The problem is the GPL. Because any GPL project insists (indeed are required - not literally, but effectively) by the GPL to make any external code they use also GPL.

    The GPL, unlike pretty much any other open source licence I can think of, does not respect anyone else’s licenses - it always wants (is designed) to trample all over them

    1. Anonymous Coward
      Anonymous Coward

      Re: The problem is not Oracle (for once)

      >The problem is the GPL. Because any GPL project insists (indeed are required - not literally, but effectively) by the GPL to make any external code they use also GPL.

      Nope the problem is Oracle, the same wankers who want to copyright APIs so they are not to be trusted.

      1. Rich 2 Silver badge

        Re: The problem is not Oracle (for once)

        So why is it Oracle’s fault? What’s your reasoning?

        As I see it, there is a perfectly usable open sourced version of zfs

        I’m no fan of Oracle, but in this specific case, why are they at fault?

        1. Jamie Jones Silver badge

          Re: The problem is not Oracle (for once)

          Exactly. I have no issue with Linus's response (just going by what's quoted in this article), but for people to then lay the blame on Oracle is wrong, but typically not surprising.

          I'm no Oracle fan, but time and time again, I've seen people diss authors for releasing software under something other than the GPL, often when the license chosen is more free!

          On that subject, I suppose that in the times of Johnson and Trump, it shouldn't be a surprise to hear GPL advocates argue that their license with *more* restrictions than, say, the BSD licenses, somehow makes it *more* free.

          I use ZFS, freely. I'm not restricted. It's in the kernel (FreeBSD) and Oracle don't charge me a penny for it.

          As Rich said, why is Oracle at fault if *your* licensing restriction won't let you use it?

          This isn't a rant against the GPL itself, just against all those in the "GPL cult".

          Hey guys, use the GPL if you want, but don't complain when others don't. The whole thing is a cult with some people - many of who don't even understand what the GPL actually says - witness the number of people who moan when they discover that someone is making money off GPL software they wrote. Again, if you don't want that to happen, don't use the GPL, but stop cheerleading it when you obviously don't agree with its philosophy.

          1. Jamie Jones Silver badge
            Thumb Up

            Re: The problem is not Oracle (for once)

            Thanks to the downvoters for proving my point!

            1. el kabong
              Unhappy

              Pointless? perhaps

              I see, you were trying to prove a point. I'm still trying to get what that point might be but no, no success yet. I fear I'm about to quit.

              OK, I quit.

            2. JakeMS

              Re: The problem is not Oracle (for once)

              @Jamie Downvoted for talking about downvotes. It's an unwritten rule around here.

              1. el kabong
                Unhappy

                @JakeMS, You just broke the unwritten rule too

                As punishment I'm giving you a click on the vertically descending arrow associated to your comment.

                1. JakeMS
                  Facepalm

                  Re: @JakeMS, You just broke the unwritten rule too

                  So I did :-D

                  1. This post has been deleted by its author

              2. Jamie Jones Silver badge
                Happy

                Re: The problem is not Oracle (for once)

                I know... It was my fiendishly devious plan to get my point proven even more!

          2. DuncanLarge Silver badge

            Re: The problem is not Oracle (for once)

            > the license chosen is more free!

            Ah but that is not a benefit. Its not a benefit to leave hour house unprotected is it? Well, it makes it easier for you to get in and out, considering there are no locks but its nicer and safer to add locks to prevent the nasty people getting in.

            Thus the GPL gives you the protection from nasty people taking advantage of you and removing your freedom.

            I'd like to stay free, I'd prefer to infringe on the handcuff owners freedom by locking up their handcuffs so they have no choice but to behave.

            1. FIA Silver badge

              Re: The problem is not Oracle (for once)

              Ah but that is not a benefit.

              No, it's a choice. The benefits or lack of are pretty subjective. (A BSD advocate would argue it's better to have access to a well tested TCP/IP stack that you can use rather than inventing the wheel again, for example).

              Thus the GPL gives you the protection from nasty people taking advantage of you and removing your freedom.

              No it doesn't.

              If your source code and susiquent modifications to it being made available is important to you use the GPL, if they're not then use something else. For some people this is important, for others it isn't. But to assume someone who choses a different licence is wrong or stupid is fairly disengenuous.

            2. rcxb Silver badge

              Re: The problem is not Oracle (for once)

              Thus the GPL gives you the protection from nasty people taking advantage of you and removing your freedom.

              In what way is someone opting not to release their modifications, "removing your freedom"? Your definition of "your freedom" sounds like indentured servitude for everyone else.

            3. Anonymous Coward
              Anonymous Coward

              "Thus the GPL gives you the protection from nasty people taking advantage of you"

              Do you force people entering a public park to let their houses open too?

              And tell it to nasty "people" like Amazon, Google, Facebook.... which can get all the code they want and give back nothing - as they don't "redistribute" - GPL was written only to ensure Stallman got all the code he wanted.

          3. DuncanLarge Silver badge

            Re: The problem is not Oracle (for once)

            > the BSD licenses, somehow makes it *more* free.

            Let me enjoy the extra freedom by removing yours. I can be free, you had your chance. Now beg for permission, beg for updates,, beg for your data.

          4. jilocasin
            Holmes

            Re: The problem is not Oracle (for once)

            I don't think anyone in the *GPL* cult is complaining about ZFS' choice of license. There is nothing in the GPL that prevents a _user_ from using ZFS on their Linux system. Linus has reasonably stated that he isn't allowing anything in the Linux *kernel* that isn't licensed GPLv2. This is doubly so with a litigious company such as Oracle has shown itself to be. The kernel developers have stated publicly and repeatedly that they are under no obligation to maintain kernel level compatibility with anything that interacts with the kernel but isn't licensed/maintained by the kernel team.

            If people want ZFS incorporated into the Linux kernel then it's up to Oracle to relicense it. Otherwise the ZFS devs will have to keep up with the kernel changes themselves.

            1. Anonymous Coward
              Anonymous Coward

              "maintain kernel level compatibility"

              Then keep on complaining some hardware manufacturers don't release drivers for Linux...

          5. Alan Brown Silver badge

            Re: The problem is not Oracle (for once)

            The BSD vs GPL issue is simple.

            Under the BSD license you can take my code, roll it into your proprietary product, hide the source or implement secondary restrictive licenses and sue ME for copyright/licensing breaches on code I helped write.

            Under GPL you must make derived source code available when distributing the product or the license is invalid and your code is ow in breach of MY copyright.

            What BSD makes it "freer" to do is steal code and falsely claim ownership of if. It's EXACTLY that experience which motivated many people to switch to GPL - including myself. There are a number of commercial closed-source solutions floating around on the net containing large unattributed chunks of my work.

            This is also the reason Microsoft used to love BSD licensing and rubbish GPL. It's an effective license to steal from the commons without giving anything back and naively relied on basic honesty that's lacking in most businesses. I've always suspected ethics classes were a front for surgical removal of any that might remain, before cauterisation of the stump.

            1. sbt
              WTF?

              ...and sue ME for copyright/licensing breaches on code I helped write

              No, since you still have a copy of the code you're using with the simple BSD license attached. The BSD license doesn't make it any easier to falsify authorship of code than the GPL does. This is FUD.

              There are plenty of examples out there of vendors incorporating GPL code into devices and shipping without providing access to source. BusyBox, anyone?

              If you want to get paid in code, use GPL. If you want to get paid, use BSD.

              1. Kiwi
                Pint

                Re: ...and sue ME for copyright/licensing breaches on code I helped write

                No, since you still have a copy of the code you're using with the simple BSD license attached. The BSD license doesn't make it any easier to falsify authorship of code than the GPL does. This is FUD.

                Purely out of interest, and neither specifying agreement nor disagreement with your post... How do you defend a claim as a person with effectively no pockets whatsoever, perhaps relying on a "public defender" equivalent whose knowledge of BSD licenses didn't exist whatsoever till you visited in desperation, and do this when you're going up against a company with pockets that wrap around the core of the planet several times and have the best lawyers money can buy on speed dial?

                When their lawyers can shout the judge a meal where your lawyer has fits thinking of the cost of just ONE serving of the cheapest entree - how do you prove that you originated the code and they're the ones trying to mislead the court? (not that I'd ever suggest any judge can be bought with an expensive meal - I fear many of them sell out for far far less :( )

                I may still have a copy of my code, but I'd love to know how to prove it is my code.

                1. sbt
                  Pirate

                  I'd love to know how to prove it is my code

                  Well, if you're being sued, I believe the burden works the other way. The provenance is more relevant than say, the presence of your name in the code (which can be easily modified). But if you've previously published it, backed it up, kept it in a SCS, these all create a digital trail which strengthens your defence.

                  An opponent trying to assert primary authorship would have to produce similar evidence that somehow pre-dates or contradicts yours, including maybe inducing some other coder to perjure themselves and claim they wrote it first. It's an unlikely scenario that I haven't ever seen; in any case, the particular license the code is published with makes no difference here (if the claim is about copyright authorship, not license compliance).

                  If you're using other folks' code, you want to be damn sure you've got the correct license with it and comply. I've seen a couple of repos online with GPL files replaced with MIT (possibly not deliberate, e.g. where code bases were combined, but still wrong). "See LICENSE" in the code makes this change too easy; every file should name the explicit license and version that applies.

                  1. Kiwi
                    Pint

                    Re: I'd love to know how to prove it is my code

                    But if you've previously published it, backed it up, kept it in a SCS, these all create a digital trail which strengthens your defence.

                    One of the very sad things with legal cases is that so very often "evidence" is irrelevant. You have absolute irrefutable proof that you're right? Sad news, their lawyer is better than yours and can convince the judge that said evidence should never be shown to the courts, thus it's never brought up in court.

                    They have proof you're right? Again, their 'much better lawyer than you could afford' will show them every little loophole to prevent such evidence being found during 'discovery' and of course tell them all sorts of legal and illegal delaying tactics to make sure you cannot get that evidence. And if somehow you do get it, again they'll make sure it doesn't come out in court (bloody hell I sound like Charles writing this! THE HORROR!)

                    Where it's a couple of evenly matched lawyers you have a chance, and in some cases you can have a good chance especially where there is some publicity. But most devs aren't in that boat when it comes to large companies 'stealing' code.

                    And sometimes people give code for free by accident. I used to have a Hotmail address pre-MS purchase. For some crazy reason (since I didn't do such things at that time) I read MS's license when they took over and discovered interesting clauses such as their having copyright over all material sent via HM, their being allowed to read and use anything in there and so on. So lets say I'd used that to send you code - how could I stand against MS in court and argue that I owned copyright when the EULA gave it to MS?

                    It is because of that I've been careful with what images/video/code etc I share online. I may license you to use my stuff but I want to retain full copyright. But Google, LinkedIn etc claim all sorts of rights over your IP. Send code to someone over gmail? Guess who owns it. Put your company logo on LI? Guess who owns it.

                    Be wary, there are some very nasty license terms out there, and some so buried in masses of legaleese and in seemingly irrelevant places you may easily miss them (I think the LI one was well buried somewhere else, almost like they intended you to miss how you were giving them all sorts of rights over your IP)

                    Have a nice night and sleep well. One of us has to, as you've reminded me of the horrors embedded in corporate licenses and Idon't think I'll sleep for a week :(

                    1. sbt
                      Boffin

                      Sleep well, you're fine.

                      On the legal case question, I'll just add that if you're truly a minnow, the big fish won't bother you. If you're big enough to notice, get a proper lawyer. Start a legal defense fund with the first dollar your code makes.

                      On the EULA point, I read these too. I've never been a Hotmail user, but I would be extremely surprised if the license actually assigned copyrights in material transmitted. I expect you've misread this and the relevant clauses permit them to read the material, to check for viruses, gather marketing information, etc. Normally EULAs for services like Linked In provide for a Use licence grant so that the service can store user generated material, and transmit it to Web users, etc. It's a necessary protection for them, but it doesn't diminish your rights to own, sell or otherwise deal in the same material; you retain the copyright.

                      There are more egregious examples of Big Web exploiting orphaned works and ripping off content makers (e.g. YouTube's policy on automated infringement detection being conditional on agreeing to licence works).

                      1. Orv Silver badge

                        Re: Sleep well, you're fine.

                        Indeed, transferring all the rights to a copyrighted work is actually fairly involved and has to be done quite deliberately -- even paying someone for something doesn't make it a work for hire and transfer copyright, unless there's a contract in place stating it does.

                      2. Kiwi
                        Pint

                        Re: Sleep well, you're fine.

                        Sorry it took me a bit to get back to this one :)

                        On the legal case question, I'll just add that if you're truly a minnow, the big fish won't bother you. If you're big enough to notice, get a proper lawyer. Start a legal defense fund with the first dollar your code makes.

                        Thing is... My code might make me $20,000 (although I have one project that did me much better than that for a while, but that combined software and hardware), out of which I might have to pay the bills. A large corporation might use some of that code in a way that makes or saves them considerably more, or in a way that is worth expending serious $$$ to defend.

                        Normally EULAs for services like Linked In provide for a Use licence grant so that the service can store user generated material, and transmit it to Web users, etc. It's a necessary protection for them, but it doesn't diminish your rights to own, sell or otherwise deal in the same material; you retain the copyright.

                        LI now requires JS to run and given their past exploits (including illegal email access - I must check where the class action on that got to!) I'm not willing to trust their JS.

                        In the past their EULA, like Google's (google+ anyway), gave them a perpetual license to use your IP in any way they saw fit including making derivitave works, selling it etc etc etc. Once you had placed your logo on their system you gave them a right to use it as they saw fit, you could not revoke that right and yes, it basically gave them ownership. You could not remove it, and they could make derivative works from and sell your IP any way they saw fit. If you posted a photo they liked then yes, they could sell it to whomever they wanted and alter it any way they wanted.

                        In contrast, Facebook's license was that you would allow them to use the work and alter it as necessary for displaying it where appropriate (basically if you gave them a 2400x2400 photo they could reduce the size to display on smaller screens/lessen bandwidth use etc). If you deleted the content, you understood it could remain in backup systems for a period of time till those backups were deleted (I can't remember if they gave a time frame or not, may have been "a few months"). You owned the work and could delete it, they wouldn't directly use it to make money from it (aside from of course selling your data and displaying ads at you)

                        We did have get some legal advice to confirm this interpretation (paid for by having a close friend in the right firm)

                        I have been involved in one photographic case which wasn't too hard a win but still looked very hairy at one point. We could provide evidence that the stolen picture was part of a series that made up a larger panorama, and still had the originals. At one point we were being accused of making false claims about having been at the location and at the time, but several of us were taking pictures that day (a group hike) and 3 of us had near identical pics of the scene (shifted by us standing a few feet apart/clouds moving slightly in the seconds between our shots). One of us (not me) made a small panorama that was posted online, and later it turned up in someone else's advertising material without permission.

                        It was a massive headache and I'd hate to go through it against a richer opponent with more at stake. In the end we got costs plus a reasonable compensation rate had the person sold the picture, but it didn't really compensate us.

                        Oh and the evidence from the cameras of the rest of us, including ones showing her taking the shot? (2 of us had her in frame while taking similar shots - one of my favourite pics with her on a hilltop overlooking a valley and clearly enjoying the view without a care in the world). Their lawyer fought hard to have them rejected as evidence - tried to tell the judge how easy it was (~15 years back) for us to make fake pictures using Photoshop and some 30-odd available pictures were clearly fakes. Thankfully the judge was also into photography and had some clues about the effort that would've been required to make that many convincing pics, as well as the "why would so many people go to so much effort for so little gain?" question.

                        Perhaps my experience has helped with some 'confirmation bias' reading similar cases (and not so similar ones), but my experience is that the person with the least cash has the least chance. And they can often more easily drag things out - you still have to pay your lawyer appearance fees and you can be ready to go time and again but they can argue one of their 40 lawyers had a slight sniffle this morning and their testimony/expertise" is absolutely critical and can they please have another postponement "thank you your honour, see you at your favourite restaurant tonight!". Same game played by prosecutors in some criminal cases, try to run you out of money by delays, last minute 'evidence' which isn't at all but takes your lawyer time to examine and realise it really was irrelevant etc.

                        You should try it one day. You'll soon realise in court that "absolutely incontrovertible evidence' is often anything but :(

            2. HandleAlreadyTaken

              Re: The problem is not Oracle (for once)

              >What BSD makes it "freer" to do is steal code and falsely claim ownership of if.

              Bullshit. Utter bullshit.

              If developer A releases some code under a BSD license, nobody can claim ownership over that code. Developer B can modify or extend A's code, and may choose not to release his changes to all and sundry, but he can't claim ownership over A's code. A third party, C, can still use the original code released by A with no restrictions. What B gets is the freedom to choose how to release his code.

              GPL advocates like you are the choosing beggar subset of third parties, bellyaching because A didn't force B to give them his (B's) code as well. To obfuscate the issue, they take a page from Orwell's book and redefine "freedom" to mean "more restrictions" and "free" as in "taking away choices".

              You could perhaps argue that the GPL promotes sharing, or that it creates a more vibrant ecosystem - and that may be true. The kind of word games you're trying instead doesn't hold though.

            3. Orv Silver badge

              Re: The problem is not Oracle (for once)

              I'm genuinely curious if you know of an example of a lawsuit that went this way. It's not a scenario I've heard of before.

              Personally I prefer BSD because while I like having my source code open, I also like to keep the possibility of later monetizing my projects. With a BSD license I could add proprietary features and charge for them, without having to release them for freeloaders to use.

              1. Kiwi
                Boffin

                Re: The problem is not Oracle (for once)

                I'm genuinely curious if you know of an example of a lawsuit that went this way. It's not a scenario I've heard of before.

                I believe El Reg has posted stories on this but I'd have to do some digging to find them, and I could be wrong. I've been involved in a photographic case but that's not exactly the same thing (though the other side did kinda try).

                You could look up Bain, Ellis, Lundy, Dotkom (or was it spelt with a "C"? apols if so), Pora, Dougherty (the DNA evidence that proved his innocence was withheld form his counsel and the jury IIRC) for a few cases in NZ, no doubt you can find other cases closer to you. These are of course criminal cases so not quite the same thing, but they all involve important evidence being withheld from the jury and sometimes even the counsel (IOW the prosecution knew they had proof of innocence but withheld it or prevented it being presented). There's likely lots to choose from in your area. Dig some yourself, if you wish to take the risk. It might change your views on humanity. Could be what turns you from a trusting carefree individual to a bitter jaded cynic.

                Unfortunately the press would be inclined to report who won and what the case was won on, but very little on the suppression of evidence. It is a wondrous thing you really have to see to believe - and you really really do not want to see it in action.

            4. jtaylor

              Re: The problem is not Oracle (for once)

              "Under the BSD license you can take my code, roll it into your proprietary product, hide the source or implement secondary restrictive licenses and sue ME for copyright/licensing breaches on code I helped write."

              Huh? Which BSD license did you read? Here's an example from FreeBSD: https://www.freebsd.org/copyright/license.html

              Copyright 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved.

              Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

              Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

              Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.

              All advertising materials mentioning features or use of this software must display the following acknowledgement:

              This product includes software developed by the University of California, Berkeley and its contributors.

        2. Doctor Syntax Silver badge

          Re: The problem is not Oracle (for once)

          "So why is it Oracle’s fault?"

          Or at least Sun's...

          Back in the day there were suggestions that the CDDL licence was deliberately designed to be incompatible with GPL: https://en.wikipedia.org/wiki/Common_Development_and_Distribution_License

          1. Anonymous Coward
            Anonymous Coward

            Re: The problem is not Oracle (for once)

            > Back in the day there were suggestions that the CDDL licence was deliberately designed to be incompatible with GPL: https://en.wikipedia.org/wiki/Common_Development_and_Distribution_License

            Yes, CDDL was intentionally written by Sun Microsystems to be incompatible with GPL. It was a demand - proffered under threat of group resignation - made by certain members of the Solaris Kernel Group.

            Both Sun and Oracle could have easily dual-licensed ZFS under CDDL and GPL. Both of them chose not to.

            It's quite entertaining to read Sun's official - and pretzel-shaped - ${EXPLANATIONS} at the time as to why CDDL was designed to be incompatible with GPL.

            In final analysis, thumbs-up to Linus Torvalds on his decision about ZFS. It's simply not worth the trouble.

            1. sbt
              Stop

              CDDL was intentionally written by Sun Microsystems to be incompatible with GPL

              It was a demand - proffered under threat of group resignation - made by certain members of the Solaris Kernel Group.

              Just to note that that claim is disputed by Simon Phipps (Sun's Chief Open Source Officer at the time).

              The proliferation of FOSS licences and incompatibilities between them is a significant harm to FOSS, even if individual 'freedoms' are protected. There are always trade-offs between the individual and the group in societies and economies.

              1. Anonymous Coward
                Stop

                Re: CDDL was intentionally written by Sun Microsystems to be incompatible with GPL

                > Just to note that that claim is disputed by Simon Phipps (Sun's Chief Open Source Officer at the time).

                Of course it is. Simon Phipps was very busy speaking out both sides of his mouth at the time.

                1. sbt
                  Meh

                  Speaking out both sides

                  This is all a bit he said, she said, combined with a theological dispute. There are credible claims on both sides as well as possible motives to paint the other side in a bad light. I was just pointing out it's not the smoking gun assumed here.

                  1. Anonymous Coward
                    Stop

                    Re: Speaking out both sides

                    > This is all a bit he said, she said, combined with a theological dispute.

                    No, it is not. In case you haven't noticed, nobody gives, or ever gave, a rat's ass about the CDDL. It was a one-trick-pony, useless license that was only ever used by Sun Microsystems.

                    As to the he said, she said bits, keep your opinions to yourself. You don't know the facts. You weren't there.

                    1. sbt
                      Facepalm

                      You don't know the facts.

                      You weren't there.

                      This is the Internet. How do you know where I was?

                      Stick to what you can prove. I've at least provided a quote.

                      1. Anonymous Coward
                        Stop

                        Re: You don't know the facts.

                        > How do you know where I was?

                        I don't know where you were at the time. I know where you were not. That is, you were not present during the CDDL discussions inside Sun and the ensuing debacle. You do not know why Danese Cooper said what she said, or the history behind it, or what happened afterwards.

                        There is a slight possibility that you were, indeed, present during these discussions. If that's the case, you are deliberately misrepresenting the facts.

                        I don't know which one is worse.

                        And if you want me to tell you how do I know all that, take a very wild guess.

                        1. sbt
                          Thumb Down

                          A very wild guess

                          I know where you were not. That is, you were not present during the CDDL discussions inside Sun...

                          Certain of that, you seem. Yet...

                          There is a slight possibility that you were, indeed, present during these discussions.

                          Come on, you can't have it both ways. Unless you're prepared to de-cloak and identify yourself as a known Solaris contributor or Sun employee, let's stick to public statements by the people involved. Here's another one contradicting Danese, this time from Jörg Schilling.

                          I'm not misrepresenting. I'm saying there's a disagreement about the reasons behind ZFS being open-sourced with the CDDL. It's not a slam-dunk, as you claim, based on the public statements provided. For what it's worth, I think Linus' caution is reasonable.

                          1. Anonymous Coward
                            Stop

                            Re: A very wild guess

                            > Here's another one contradicting Danese, this time from Jörg Schilling.

                            Jörg Schilling never worked at Sun.

                            Did you?

                            1. sbt
                              Boffin

                              Schilling never worked at Sun

                              I never claimed he did. But since you wouldn't accept the statements of Sun employees like Simon Phipps as proof of a disagreement, I offered a third-party source instead, with at least some claim to contact with the Solaris contributor group.

                              Working at Sun seems irrelevant, since not even Sun employees agree on this. Well, at least Danese vs. the rest that I could find.

                              If you are prepared to accept Sun employees' views on this, look at the rest of the thread I originally linked containing Simon's comment; there are more Sun employees than just Simon disputing Danese's version of events. However, I didn't spot any, not one, agreeing with her claims about the reasons for CDDL being chosen rather than GPL.

                              It certainly demonstrates there's a disagreement; my original contention, QEFD. Frankly the more these comments go on, the more I'm convinced the facts support the Phipps position altogether.

                              1. Anonymous Coward
                                Stop

                                Re: Schilling never worked at Sun

                                > If you are prepared to accept Sun employees' views on this [ ... ]

                                As a rule, I don't accept views. I accept facts. And I refuse to accept lies presented as facts. And that is precisely what Simon Phipps was doing.

                                If you are happy with Simon Phipps' narrative about the CDDL, that's fine with me. You can believe whatever you wish. However, I can tell you - from direct experience - that Phipps' narrative is very far removed from the truth. It's the sanitized for public consumption version of the facts. Danese's version is the correct one.

                                Lastly, I'm very happy that you proved to yourself that you were right. Meeting that burden of proof must have been very challenging.

                                1. sbt
                                  Holmes

                                  I don't accept views. I accept facts.

                                  I'll accept a view as evidence of someone's opinion. So I accept that it is your opinion that Danese's version is the correct one, and Simon's is false. You've acknowledged Simon and Danese have different narratives. You must therefore also accept my initial claim there's a disagreement.

                                  But unless you're willing to offer some facts (and evidence for same) and not mere assertions, I'm going with my follow-up, as well. I think the Phipps narrative is more credible in explaining Sun's decision on the CDDL license, and the motive Danese claims about the CDDL being selected for deliberate incompatibility with GPL don't make sense, because the expectation at the time (e.g. as explained by Bryan Cantrill) that CDDL code would be accepted in Linux. Without another corroborating source in support, I find it hard to accept Danese's version as the more likely one.

                                  Facts, please. Otherwise, we'll be going around in circles.

                                  1. Anonymous Coward
                                    Stop

                                    Re: I don't accept views. I accept facts.

                                    > Facts, please. Otherwise, we'll be going around in circles.

                                    I already gave you the facts.

                                    It is not my opinion that Danese's version of the CDDL story is the true one. That is a fact.

                                    You have no way of determining whether that is a fact, or not, because you did not work at Sun at the time -- and I take it you never worked at Sun -- and you have no clue what went on inside. Your only reference points are external mailing list messages.

                                    As I've already said: you are free to believe whatever you wish to believe. Everyone is entitled to their own opinions, and beliefs, but not to their own facts. Just because you believe something to be true, it does not mean it is.

                                    1. sbt
                                      Facepalm

                                      You keep using that word...

                                      Your claim to have the facts isn't credible as rebuttal without independent evidence. Without evidence, your claim is mere opinion.

                                      Circle complete; cheerio!

                                      1. Anonymous Coward
                                        Stop

                                        Re: You keep using that word...

                                        > Without evidence, your claim is mere opinion.

                                        https://web.archive.org/web/20110722120048/http://caesar.acc.umu.se/pub/debian-meetings/2006/debconf6/theora-small/2006-05-14/tower/OpenSolaris_Java_and_Debian-Simon_Phipps__Alvaro_Lopez_Ortega.ogg

                                        At first you can watch Simon Phipps swoon about how open OpenSolaris was going to be. External developers getting commit privileges. Accepting contributions from anyone, solely based on merit. Etc, etc, etc. Sounds great. Too bad none of it was true, nor was there ever any intent on Sun's part to ever make it true.

                                        And then, you can listen to the same Simon Phipps wave his way around the choice of venue clause in the CDDL - in response to a question from Debian.

                                        He tries to handwave it away as a no-op. It's a no-op until you're a developer from Denmark and you get sued in Santa Clara Superior Court because you - unknowingly - infringed on a patent. Then it becomes a very real op, and it costs a lot of money. Patent lawsuits - or threats of patent lawsuits - were commonplace in 2006.

                                        That's the guy you're offering as proof.

                                        Danese starts speaking at 24:42. Danese's explanation of Sun's deliberate choice of CDDL precisely because it is incompatible with the GPL starts at 27:26.

                                        Funny thing is, no immediate rebuttal from Phipps about the intentional GPL incompatibility. The denials only started after Danese left Sun.

                                        Believe what you wish.

        3. eldakka

          Re: The problem is not Oracle (for once)

          So why is it Oracle’s fault? What’s your reasoning?

          As I see it, there is a perfectly usable open sourced version of zfs

          I’m no fan of Oracle, but in this specific case, why are they at fault?

          Because OpenZFS is still under the CDDL as it is a derivative - a fork - of Sun's (now Oracle's) original ZFS released under the CDDL. Since it is a derivative, I believe (OpenZFS devs would have to chime in here?) it uses some of the same APIs.

          Therefore:

          1) It is under the incompatible CDDL, therefore can't be included in GPL'ed code as it stands. And CDDL doesn't allow you to make a fork and change the license. A completely new clean-room from-scratch implementation would have to be done.

          2) Oracle has a history of suing over APIs. Since the current US precedent1 is that APIs can be copyrighted, the Linux kernel project is rightly worried that if ZFS came under the auspices of them, it would paint a big target on their backs for Oracle to sue.

          3) A clean-room implementation would have to do away with the existing APIs and come up with their own, otherwise see 2.

          1: Currently on appeal to SCOTUS, to be heard sometime this year (probably).

          1. Bronek Kozicki
            Facepalm

            Re: The problem is not Oracle (for once)

            OpenZFS was originally part of OpenSolaris in 2005, and the ZOL (ZFS on Linux) project started in 2008. You can clearly see "Sun Microsystems, Inc." in ZOL license and, since it was forked before the Sun acquisition in Jan 2010, Oracle lawyers can do absolutely nothing about it.

            1) That's correct, the intent was to make it impossible for Linux to drive ZFS readmap etc.

            2) Does not apply where there is no Oracle IP which could be sued upon

            3) It might be useful endeavour, but would be also very expensive and would risk divergence with ZFS on Linux and on other OSs (I mean MacOS, OpenIndiana and BSD). Unless some shared governance was established in a way that Linux kernel developers would respect, but they do not really have a great track record for respecting other operating systems needs, which is something that OpenZFS developers care a lot about because the filesystem is used by three other operating systems.

            The other way to look at this problem is this:

            1) GPL is an open source license, and any work derived from GPL must be under GPL license (which makes it "viral")

            2) CDDL is an open source license, and any work derived from CDDL must be under CDDL license and only this license (which also make it "viral", but also incompatible with GPL)

            OpenZFS developers: "please respect our needs, we need these APIs for our out-of-tree open source project"

            Linux developers: "we won't because that's an extra burden for us, please move your project into Linux kernel tree if you want special treatment"

            OpenZFS devopers: "we cannot, the license is incompatible and cannot be changed"

            Linux developers: "something, something, something, Oracle something, something, lawyers something"

            Me personaly: "WTF?! you should have learned over the past decade that Oracle has nothing to do with OpenZFS!"

            Everyone else: "Oracle! Oracle!"

            Me personally: going mad and starting to write long diatribe

            There is such a thing as ZFS file store appliances from Oracle; it is based on ancient Linux kernels and Oracle developers who maintain it (assuming there are still some left) couldn't care less if some APIs are taken from them in the upstream kernel because they are not going to use it anyway. So taking APIs away in the upstream only harms open source OpenZFS developers. Then claiming this on Oracle is really disingenuous and simply disrespectful to OpenZFS, which is a proper open source project with its own governance, where multiple operating systems are involved and which is entirely independent of Oracle.

            To me this feels like blackmail where OpenZFS is held hostage in a stand off between Linux and Oracle to hand the ownership of ZFS because everyone knows by now that BTRFS is shit not that great. Of course Oracle does not care, and OpenZFS developers end up with a gun to their head for no fault of theirs.

      2. bombastic bob Silver badge
        Devil

        Re: The problem is not Oracle (for once)

        "the problem is Oracle, the same wankers who want to copyright APIs so they are not to be trusted."

        Mostly true; however, GPL is unfortunately incompatible with a LOT of other kinds of licenses.

        Oracle _could_ dual-license their open source code if they wanted to. [I do that] Something like "You can use this code with EITHER a GPLv2 or [whatever license]" then include text of both and let the end-user choose which one. Once GPL'd it remains as such and becomes part of Linux, eternally patched and forked like every other chunk o' code that's been contributed to it.

        I think that's what Linus wants to do. But it's hard for him to express this without the use of profanity...

        1. Doctor Syntax Silver badge

          Re: The problem is not Oracle (for once)

          I think that's what Linus wants to do. But it's hard for him to express this without the use of profanity... do this without contacting and getting the approval of all past kernel contributors including those who have experienced their own kernel panic and gone to that great dump partition in the sky.

          FTFY

        2. Graham Cobb Silver badge

          Re: The problem is not Oracle (for once)

          Bob has hit the nail on the head - this isn't about freedom, it is about licences. Linux uses GPL, and the people who have contributed to Linux contributed on the basis that their contribution was going to be used in the way defined by the GPL. I have personally contributed to several projects on exactly that basis.

          Oracle are free to create ZFS under any licence they wish. But if they want it to be part of Linux they have to license it under (a licence compatible with) GPL. They could do that - and I can see why Linus would want it to be accompanied by a clear written promise that they really, really mean it is licensed and can be distributed that way.

          If they don't want it part of Linux that is fine as well, of course. People have choices of what they want to use.

          This must be the first time I have ever agreed with Bob :-)

          1. sbt
            Alert

            To dual-license, wouldn't they need...

            ...agreement from every single ZFS contributor, including Oracle, but not just them? This is usually a stumbling block for re-licensing code once it's been out in the wild and picked up lines from various sources.

            1. Gerhard Mack

              Re: To dual-license, wouldn't they need...

              Unless Oracle was assigned the copyrights, that is correct. Relicencing is even less likely for Linux though, since there are far more contributors and some of them are no longer alive.

          2. Kiwi
            Pint

            Re: The problem is not Oracle (for once)

            This must be the first time I have ever agreed with Bob

            It is a rare and sometimes somewhat horrifying experience, but it is possible for you to heal and lead a largely normal life afterwards.

            The shaking usually goes away after a year or two...

            --> I hear tankerloads of these may help :)

          3. Anonymous Coward
            Devil

            This must be the first time I have ever agreed with Bob :-)

            I seriously looked for a point to downvote Bob’s contribution. I couldn’t find one, but chose to not upvote it.

      3. Daniel B.

        Re: The problem is not Oracle (for once)

        The problem is ... both.

        The GPL is designed to act like a virus, infecting everything it touches.

        Oracle is known for being awful on these kind of things, even if you don't count the whole Java API fiasco.

        And then the CDDL itself was purposefully made incompatible with the GPL.

        Oracle also closed sourced ZFS at some point so there are now two branches: propietary ZFS by Oracle and OpenZFS by BSD. AFAIK OpenZFS is still stuck on the CDDL license.

    2. alain williams Silver badge

      Re: The problem is not Oracle (for once)

      It is not a GPL problem - it is that some licenses are not compatible with others.

      Whoever wrote the code gets to choose the license and thus what is/is-not possible to do with their code. If you do not like the license then you are at liberty to write it all from scratch and release it under some other license.

      1. Len

        Re: The problem is not Oracle (for once)

        But that the GPL is not compatible with the CDDL is a licensing problem not caused by the CDDL. One could therefore argue it is GPL problem.

        1. jilocasin
          Headmaster

          Re: The problem is not Oracle (for once)

          It's as much *caused* by the GPL as by the CDDL. It's all a matter of where you are standing.

          If you are on ZFS island, then the problem is that the *GPL* isn't compatible with the CDDL that's the problem.

          If you are on Linux land, then the problem is that the *CDDL* isn't compatible with the GPL that's the problem.

          So in the end, what do you want to do? Run ZFS on Linux, or Linux on ZFS? Neither is going to happen without a compatible license. Linux can't really change license at this point, so the ball's in Oracle's court.

          1. Lars Johansson
            Linux

            Re: The problem is not Oracle (for once)

            Rubbish!

            The GPL predates the CDDL by a about 15 years. Any problems are clearly caused by the latter being deliberately incompatible with the former.

          2. FIA Silver badge

            Re: The problem is not Oracle (for once)

            ...Linux can't really change license at this point, so the ball's in Oracle's court.

            It isn't though, Oracle closed sourced their ZFS implementation in 2010, the ZFS in question is OpenZFS which contains contributions from people other than Sun.

            Oracle could open source their implementation of ZFS under the GPL, but not OpenZFS, without the explicit aproval of all concerned.

            1. jilocasin
              Headmaster

              Re: The problem is not Oracle (for once)

              While that is true, OpenZFS is based upon the code originally licensed under CCDL by (now) Oracle. So unless the entire OpenZFS code base has been completely rewritten in a clean-room without infringing on any existing ZFS patents, they can't relicense, or even dual license it without Oracle's blessing.

              So you just can't get around Oracle's involvement in the process.

        2. alain williams Silver badge

          Re: The problem is not Oracle (for once)

          Software authors chose the licenses that they did because they wanted the protections/liberties that those licenses provide. Different licenses are not like two pieces of software that you can make play together with a clever bit of interface code.

          Trying to blame one or the other is a bit like blaming a motor-car or a bicycle because you cannot make a composite vehicle; the two have very different purposes.

          1. John Robson Silver badge
            Joke

            Re: The problem is not Oracle (for once)

            "Trying to blame one or the other is a bit like blaming a motor-car or a bicycle because you cannot make a composite vehicle; the two have very different purposes."

            Have you seen these new fangled motorbikes ? ;)

        3. Alan Brown Silver badge

          Re: The problem is not Oracle (for once)

          GPL existed long before CDDL

          Sun released other code under BSD or GPL licenses

          It's been carefully and thoroughly explained by many people many times - several of them with firsthahd knowledge of the events in question - why CDDL is deliverately incompatible with GPL and why that was done.

          As for "Blame Oracle" - the very first thing they did after acquiring Sun was to cancel ALL the open source licensing on Solaris-related products. OpenZFS (and opensolaris) is forked from the LAST available CDDL ZFS version - more than a decade past.

          Oracle have _no interest whatsoever_ in reopening ZFS under any kind of license, let alone changing legacy licensing. It's been clear for a very long time they'd prefer to invalidate CDDL without a replacement, but thankfully they can't (yet) do that retroactively until they grease the right palms in various legislatures.

      2. vtcodger Silver badge

        Re: The problem is not Oracle (for once)

        "If you do not like the license then you are at liberty to write it all from scratch and release it under some other license."

        Unless, of course, it's a API or some other entity that Oracle feels possessive about

        1. alain williams Silver badge

          Re: The problem is not Oracle (for once)

          Unless, of course, it's a API or some other entity that Oracle feels possessive about

          Since we are talking about Licenses and Linux - you should probably have written SCO rather than Oracle.

          Me: looks around warily and heads off to wash my mouth out and find some garlic.

      3. bombastic bob Silver badge
        Devil

        Re: The problem is not Oracle (for once)

        well I favor convincing the author to ALSO have a GPL version of it. dual licensing is probably the best way if you want your code to be TRULY "open" source.

        MIT or BSD licenses merely require a copyright statement, can ship closed.

        GPL demands code to be 'eternally open', no additional requirements (like a copyright).

        So you do this: license under EITHER an MIT or GPL license, end-user's choice. Once GPL'd that fork will remain GPLd, but the original can be used ANY way desired, provided you include the copyright in your docs/license/whatever. Best of both worlds. It's what _I_ do.

        1. Alan Brown Silver badge

          Re: The problem is not Oracle (for once)

          "GPL demands code to be 'eternally open', no additional requirements (like a copyright)."

          Bzzzt. Incorrect.

          GPL is fully copyright aware. The license conditions are such that if breached then the license is void and what results is a simple copyright violation - with inherent criminal penalties in the USA if there's more than $50k involved.

          This is why the very few cases challenging GPL have collapsed. You can comply with the license by publishing the derived source code, or you can face potential jail time for criminal copyright infringements. Once judges point this out the arguments from lawyers dry up rapidly...

          This is like the case in the UK where water companies tried to claim ownership of all water falling from the sky and prosecuted people collecting rainwater. When walked through "their" water (and associated liabilities) causing property damage through flooding etc, the prosecution lawyers asked for a hasty adjournment, then dropped all cases in less than an hour - they'd realised they might prevail on water ownership but in doing so, open themselves up to hundreds of billions of dollars of liabilities and well-heeled insurance companies claiming their repair costs back.... (it's called a pyhrric victory)

          1. anonymous boring coward Silver badge

            Re: The problem is not Oracle (for once)

            "This is like the case in the UK where water companies tried to claim ownership of all water falling from the sky and prosecuted people collecting rainwater."

            Oh, dear oh dear! This actually happened?

            1. Wilseus

              Re: The problem is not Oracle (for once)

              I've not been able to find any evidence of that. In some parts of the USA such as Colorado, this already applies however...

            2. Anonymous Coward
              Anonymous Coward

              Re: The problem is not Oracle (for once)

              Not in the UK as far as I can tell from Googling, but there are plenty examples of disputes over "Rain ownership" in the US of A

    3. Flocke Kroes Silver badge

      Re: The problem is Oracle (again)

      Open source ≠ Free software

      Open source only means you can read the source. There may be wonderful permissions with an open source license - or not. There may be horrible requirements with an open source license - or not. There is a huge selection of open source licenses to choose from. Some of the most popular ones are compatible with the GPL.

      Free software is about preserving the four freedoms for end users:

      0: The freedom to run the program for any purpose.

      1: The freedom to study how the program works, and change it to make it do what you wish.

      2: The freedom to redistribute and make copies so you can help your neighbour.

      3: The freedom to improve the program, and release your improvements (and modified versions in general) to the public, so that the whole community benefits.

      You can mix Gnu GPL and CDDL version 1 source code in a project for your own use but you cannot distribute the the result to others. Gnu GPL requires that the four freedoms are preserved in the distributed work but CDDL version 1 requires allowing others to take those freedoms away.

      Sun were aware that CDDL and GPL were incompatible. They deliberately chose that when creating CDDL. They created ZFS so they get to pick the license. Likewise many programmers chose GPL for their own works or chose to contribute to GPL projects.

      If you are unhappy with that you are welcome to create your own kernel and distribute it with the either the CDDL or something compatible so you can use ZFS. Ideally you should be able to create something compatible with ZFS and distribute it with the GPL or one of the many compatible licenses so it can be used with Linux. Oracle may decide to sue you whether that is legal or not.

      The benefit of GPL is the huge amount of high quality software available to link into your project. The cost is that if you do, you cannot take the four freedoms from others. You get to decide if the benefit is worth the cost. If that feels like being trampled on to you, imagine a big smile on my face while I go STOMP STOMP STOMP.

      1. Rich 2 Silver badge

        Re: The problem is Oracle (again)

        I really didn’t want to get into this, but a couple of observations;-

        “ Gnu GPL requires that the four freedoms are preserved in the distributed work but CDDL version 1 requires allowing others to take those freedoms away”

        I’m sorry, but that is one of the the most twisted arguments I’ve ever heard (that is repeated time and again by gpl people - replaces “CDDL” with any non-gpl licence you like).

        “The benefit of GPL is the huge amount of high quality software available to link into your project”

        Of course, there is also 100 times more crappy quality software out there as well

        And “ high quality software available to link into your project” is not restricted to just GPL-licensed stuff. There’s plenty of other non-GPL software out there as well, and most of that doesn’t cause problems (no Borg assimilation required) when it comes to linking in and distributing your derivative work

        1. Adair Silver badge

          Re: The problem is Oracle (again)

          So what's your complaint? The GPL is aimed to achieve particular practical and philosophical ends. If you don't like it don't use it.. If some software you like uses it, which stops you doing the thing you want to do, then write your own code, by all means, but don't whine because someone else has chosen a particular, perfectly legitimate, licence. Likewise, for those who find that some other licence prevents them doing what they want to do, whether under the GPL or anything else - don't cry like a spoilt brat.

          Let's just be thankful that all software isn't (yet) owned by Oracle, MS, Amazon, Google or anybody else, and that people are still allowed to release their software under whatever licence they see fit, whether it be a selfish money grubbing one, an open hearted generous one, etc.

          It's still the freedom thing, which includes the freedom to be mortally offended that somebody, somewhere isn't doing what we want.

        2. NotBob

          Re: The problem is Oracle (again)

          What's twisted about that argument? It's perfectly accurate.

      2. heyrick Silver badge

        Re: The problem is Oracle (again)

        "Gnu GPL requires that the four freedoms are preserved in the distributed work"

        On paper, yes.

        In practice, not so much. As any number of ADSL routers, IPcams, and plenty of other IoT tat will easily demonstrate.

        1. SImon Hobson Bronze badge

          Re: The problem is Oracle (again)

          So life isn't perfect ?

          Enforcement action is taken over GPL infringements (I'm assuming you are referring to the number of devices that use GNU/Linux without complying to the letter with the GPL) but it needs several things to be in place - the two key ones being that someone points the issue out to a copyright holder, and that holder is not already fully committed with other actions. BTW - how many of these masses of non-conforming devices you seem to know about have you reported to the/a copyright holder ?

          But I would counter that for many devices (more than I've actually looked at the licence status), they come with instructions on where to find the source to various bits of FOSS code they are using. This is quite common with routers for example. It shouldn't be hard to do provided you keep control of the development process and the licences for the programs you use.

    4. DuncanLarge Silver badge

      Re: The problem is not Oracle (for once)

      > to make any external code they use also GPL.

      No, it merely needs to be compatible with the GPL.

    5. steelpillow Silver badge

      Re: The problem is not Oracle (for once)

      The problem is neither Oracle nor the GPL. Here we have two incompatible licenses in a world where incompatible licenses and legal actions are the norm. The one player wants to do a deal with the other, in order to avoid legal action.

      ,,,........................................................................................ ..

      oh, sorry, dropped off for a minute there. Its so normal the real problem is I can't fecking keep awake to finish m

    6. Temmokan

      Re: The problem is not Oracle (for once)

      "You will be GPL'ed. Resistance is futile."

      You don't like it, you don't use it, that simple. There's nothing wrong with GPL per se - personally, I prefer to have possibility to see the source of code that runs on my device(s), when it's possible.

  3. Belperite

    So basically nothing has changed. ZFS development will continue, albeit with workarounds for things which are now using EXPORT_GPL. It will be Ubuntu which may prod Larry's lawyers into action as they're looking to include it in their installer.

    1. Anonymous Coward
      Anonymous Coward

      The article is about the Linux kernel.

      Anyone can include it in their distro along with a bucket full of other licenses e.g. non-free.

    2. Alan Brown Silver badge

      "It will be Ubuntu which may prod Larry's lawyers into action as they're looking to include it in their installer."

      If the installer downloads the modules or compiles the sources to link into the kernel on the fly then that's perfectly legal and acceptable. Oracle attempting to stomp on it will come off second best in such a stylish.

  4. This post has been deleted by its author

    1. Anonymous Coward
      Anonymous Coward

      Re: Torvalds declared: "Don't use ZFS. It's that simple."

      Well... systemd aside, the same problem occurs on Linux as it does Windows, software. With Linux I know where to look for whatever software (does Nvidia have propietary drivers for BSD?). With that said, if you want a file server, BSD (my little NAS is BSD for that reason).

      1. BinkyTheMagicPaperclip Silver badge

        Re: Torvalds declared: "Don't use ZFS. It's that simple."

        Nvidia ships proprietary binary drivers for FreeBSD. You can also use open source drivers for AMD/Intel.

        NetBSD supports the Linux DRM kernel interface so it can run recent nouveau drivers (plus AMD, Intel)

        OpenBSD supports the Linux DRM kernel interface but does not include nouveau support, it's earlier levels of support for AMD and Intel.

        DragonflyBSD supports DRM for AMD and Intel only it looks like - never used.

        1. Anonymous Coward
          Anonymous Coward

          Re: Torvalds declared: "Don't use ZFS. It's that simple."

          >NetBSD supports the Linux DRM kernel interface so it can run recent nouveau drivers

          Nouveau drivers are for masochists, I appreciate their work without technical document support but they are just too flaky, lacking many features and slow.

          1. Alan Brown Silver badge

            Re: Torvalds declared: "Don't use ZFS. It's that simple."

            "Nouveau drivers are for masochists"

            There's a simple answer as explained by several of the gurus on #nouveau in the usual open source IRC network:

            "Don't buy Nvidia products and expect them to work on Linux. Even if the drivers support your device NOW that can stop at any time without warning. Nouveau is a best-effort project and Nvidia have consistently been less than helpful whilst providing constantly moving targets for their largely undocumented APIs. Nvidia simply don't care about Linux support as it's such a tiny area of their marketplace"

            In short - and this is the advice from the Nouveau team themselves - if you want it to simply work, buy AMD graphics cards/GPUs

            1. Bronek Kozicki

              Re: Torvalds declared: "Don't use ZFS. It's that simple."

              The funny thing about Nvidia on Linux is that they do care about their CUDA libraries working well on Linux, because that's the platform of choice for the majority of machine learning (especially in the cloud). And these CUDA libraries work only on regular Nvidia drivers.

            2. Anonymous Coward
              Anonymous Coward

              Re: Torvalds declared: "Don't use ZFS. It's that simple."

              >Nvidia simply don't care about Linux support as it's such a tiny area of their marketplace

              They do and Aaron Plattner is very helpful, I've had far more issues with AMD especially how they screwed people over with sub HD5000 and FGLRX support.

    2. Paul Crawford Silver badge

      Re: Torvalds declared: "Don't use ZFS. It's that simple."

      Torvalds comment seem to miss a key point: you don't usually chose ZFS for performance but for (1) data integrity, and (2) low-overhead snapshots (another data integrity plus if faced with user mistakes or user-level ransom-ware).

      BtrFS promises similar, but to my limited knowledge is not nearly as good as ZFS for now (also ironically it was started by Oracle to have a ZFS alternative).

      1. katrinab Silver badge

        Re: Torvalds declared: "Don't use ZFS. It's that simple."

        It is fast enough to saturate my gigabit network connection, and that is it running on 7 y/o desktop-grade hardware. Also, I haven't had a single instance of data corruption in 10 years of using it.

        At some point, I will upgrade to 10 gigabit, but first, I will need a new server with sufficient pci-e lanes.

        1. werdsmith Silver badge

          Re: Torvalds declared: "Don't use ZFS. It's that simple."

          Also, I haven't had a single instance of data corruption in 10 years of using it.

          How rare is file system data corruption? I can’t remember having file system data corruption in decades of using any OS.

          1. katrinab Silver badge

            Re: Torvalds declared: "Don't use ZFS. It's that simple."

            I've had a couple of instances on ext2/3/4 and NTFS, and loads of instances on FAT.

          2. Anonymous Coward
            Anonymous Coward

            Re: Torvalds declared: "Don't use ZFS. It's that simple."

            How rare is file system data corruption?

            If you've never had a hard drive crash and suffered data loss as a result, then you haven't had enough hard drives. Anyone who builds arrays with ZFS, Btrfs, or similar is bound to have seen it happen, which is why they're taking steps to mitigate it.

            1. John Brown (no body) Silver badge

              Re: Torvalds declared: "Don't use ZFS. It's that simple."

              I suspect he's talking about corruption caused *by* the filesystem or other software interacting with the filesystem rather than hardware failures.

          3. Alan Brown Silver badge

            Re: Torvalds declared: "Don't use ZFS. It's that simple."

            "How rare is file system data corruption"

            Statistically about one misread sector per 4.8TiB read (Undetected disk ecc error) - ZFS is the ONLY file system which can not only pick this up but it writes the correction back to the drive on the fly.

            The ZFS philosophy is simple - "All disks are crap. Deal with it"

            In practical terms I've seen ZFS filesystems run without data error for 10+ years, rolling disk drive replacements in that time and total unplanned failures too. It just hiccups momentarily and carries on when a drive goes bad.

            1. This post has been deleted by its author

          4. Carpet Deal 'em

            Re: Torvalds declared: "Don't use ZFS. It's that simple."

            It happened to me multiple times in fairly short succession when I used Reiser3(that distro's default). When I switched to one using ext3, the corruption stopped.

            1. Orv Silver badge

              Re: Torvalds declared: "Don't use ZFS. It's that simple."

              I had corruption issues with Reiser, but I later found that machine had a defective disk controller, so I can't really blame Reiser for it. Reiser certainly broke more spectacularly than other filesystems, though.

      2. Alan Brown Silver badge

        Re: Torvalds declared: "Don't use ZFS. It's that simple."

        "BtrFS promises similar, but to my limited knowledge is not nearly as good as ZFS "

        After several total loss incidents during testing of BtrFS that ZFS simply shrugged off, I think you're highly optimistic.

        Everyone who cared about data integrity jumped ship in the development arena years ago, so the prospects of BtrFS improving are slim to negligible. The better mousetrap already exists.

    3. hmv

      Re: Torvalds declared: "Don't use ZFS. It's that simple."

      Or just run Linux. Despite all the toys being thrown from pram to pram, ZFS on Linux works fine (between various servers I have around 100Tbytes of available storage).

      Not that I'm bashing *BSD; that's my escape plan if systemd gets annoying enough.

    4. bombastic bob Silver badge
      Devil

      Re: Torvalds declared: "Don't use ZFS. It's that simple."

      you DO have a good point...

      /me has been using ZFS with FreeBSD for several years, since 8.x even, and it saved my ass a couple of times when hard drives started going bad after a few years of use, and I found out early because of ZFS, and preserved all files when moving to a new hard drive in both cases.

  5. tiggity Silver badge

    I like ZFS

    Found it good back in the day on Solaris & still find its features useful.

    Given the history of Oracle in throwing their lawyers weight around I can understand Torvalds not wanting anything to do with zfs in the kernel - why risk it, they are far more aggressive than Sun were over alleged IP use (BTW F. Off on copyrighting APIs Oracle)

    Torvalds is taking sensible precautions, but irrespective of that its not hard to find a flavour of Unix that supports zfs if you want to use it so (as long as there are different Unixes out there) it's no problem.

    Disclosure I run FreeBSD for more "serious stuff", but also have an old laptop that rather than chuck I run Mint and use it for occasional web browsing, media watching etc. Horses for courses.

  6. heyrick Silver badge

    Strange

    A user of a licence that only plays well with itself is getting upset because something has been released with a more permissive licence.

    1. cornetman Silver badge

      Re: Strange

      I think it was made pretty plain in the article that Linus was for more concerned about the ambiguity of the situation with ZFSs licensing situation in the light of Oracle's notoriously litigious reputation.

      There are plenty of free software licenses that are compatible with the GPL.

  7. Long John Silver
    Pirate

    Innovation and 'rights' cannot co-exist

    The 'disagreement' reported here is but part of a general escalation of disputation over copyright and patents across the range of cultural output. For instance, consider bitter disputes over interpretation of 'fair use' in materials uploaded to YouTube and similar publishing platforms; think on the recent EU effort to 'protect' supposed 'rights' of commercial online publishers by restricting how others may link to 'content' of the former; also ponder the efforts of bodies like the Premier League wishing to stamp out a black market in streaming which naturally arose in response to the League's restrictive distribution and price-gouging.

    The foregoing, and much more too, represent deep malaise such as is retarding cultural advance. The flagship argument offered by holders of 'rights' is based upon supposed need to protect creative individuals'/teams' entitlement to draw money from their efforts; reality is of 'rights' being commodities tradable by owners which most often are conglomerate concerns, these seeking 'rent' at arbitrarily legal-monopoly determined prices - this regarding copyright almost in perpetuity (e.g. creator's life time plus 70 years).

    Inimical effects on advance arise from provisions in law preventing unauthorised 'derivation' from works still in copyright. Intent being to protect income streams arising from the work derived from. This restriction is a necessary consequence of the concept of copyright. Exceptions and relaxations (e.g. 'fair use') enshrined in law are impossible to define absolutely and thus have given rise to a plethora of litigation. Persons of creative aptitude must constantly look over their shoulders lest their expression of ideas somehow intrudes on the 'rights' of another, some such persisting for a century and more.

    Given that hardly anything, if at all, is created without attachment to cultural context, i.e. that which came before, derivation becomes bedrock for all cultural advance. One area in which derivation is welcomed is presentation of ideas/findings in academic literature. The more that others acknowledge and derive from an academic work the greater becomes the reputation of its author(s). Reputation is the currency of academic success. The second greatest sin in academia (the first being confabulation) is plagiarism. Dispute over activity by Sci-Hub and Lib-Gen focusses on right of distribution rather than ownership of ideas embedded in 'content'. This separation of threads within copyright is helpful for understanding issues in the general application of copyright, notably in popular culture, where the two threads have been intertwined into a single strand.

    Current ferment about so-called 'intellectual property' was brought about by introduction of digital encoding for 'content' and by inception of the Internet for mass use. The former incontrovertibly established that the 'message' has separate existence to that of the 'medium' in which it is embedded. For example, printed books and vinyl records are physical media; they are objects tradable according to conventional market-economics because they possess scarcity of supply; not so a digital sequence representing the 'content': it is indefinitely reproducible at full quality for negligible cost; moreover, it can be sent anywhere on the world via the Internet by using simple equipment to be found in most households; supply and demand cannot apply and thus neither can price-discovery (in a market not 'fixed' through monopoly powers). Put thusly, digital sequences carry no intrinsic monetary worth (despite cost of initial production), yet cultural worth, measured on scales to which money-men are not privy, could be immense.

    The concept of 'intellectual property' rests on shifting sand. It seeks to protect creative activity but results in stifling it. That suggests a form of reductio ad absurdum. Further suggestion of lack of fitness for purpose is evinced from increasingly complicated, nay impenetrable, law seeking to bolster a time-expired idea. To be borne in mind is law and morality are not coincident circles in a Venn diagram. Their overlap shifts according to general societal diktat. Physical property rights, some nowadays worthy of deep reappraisal, do not carry forward into the falsely analogous rights claimed for 'intellectual property'. Thus, individuals may choose to disobey copyright restrictions with clear conscience. Unfortunately for businesses and agencies objecting to the reach of copyright and patents they remain easy prey to parasitical lawyers.

    Copyright and patents exist solely through consent within and between nations. Nation 'A' may decide to block access by nation 'B' to manufactured widgets: the stuff of trade wars and sanctions. Nation 'A' cannot block access to published ideas and those embedded in widgets (patents). Now that 'intellectual property' is established as restrictive, price-gouging, and destructive, 'third world' nations may soon abandon all pretence at obeying copyright and patents (especially for pharmaceuticals). These perhaps soon to be followed by other nations and trading blocks not in thrall to the USA.

    Take away copyright and patents, what is left to motivate innovation? The only tradable commodity related to supposed 'intellectual property' is the imagination and skill to produce new cultural artefacts (ranging from pharmaceuticals to recorded caterwauling of a 'pop star'). In that competitive market reputation is all. Reputation can be protected under extant law should one person attempt to impersonate another in order to 'steal' money-making capability. Reputation depends, in part, upon attribution from others when they re-issue, modify, extend, and otherwise derive from an existing work. 'Entitlement to attribution' replaces copyright and patents. It is 'moral' entitlement few would object to and enforceable by people at large, rather than through law, through shunning blatant plagiarists.

    Bear in mind, digital sequences have no monetary value. However, individuals/teams producing work others admire can raise funding for further works through voluntary patronage (donation, subscription, crowd-funding, etc.) and sale of associated 'added value' physical products and services; success being predicated upon reputation accrued from previous works; as reputation grows then so might income. People seeking to make livelihood from digitally representable cultural artefacts have no 'right' to do so; they first must gain reputation; also take note that most of the greatest contributions to culture (e.g. in science, mathematics, philosophy, and 'serious' music) were not, nor intended to be, vast money-spinners: present day cultural expectations of revenue generation are distorted by the tawdry rentier economics underlying distribution of popular 'content'.

    The Internet places distribution in the hands of creators should they wish. Perhaps they might form distribution co-operatives. Cottage industry supporting roles could arise; these would have no 'rights' over material passing through their hands', they would receive money for services rendered. Collapse of the rentier economy for culture would impact solely upon the behemoth distribution/publishing houses (and upon some talentless caterwaulers manufactured according to the wiles of marketing), these populated by people whose only acquaintance with originality being through 'creative accounting'.

    Huge sums of individual and national discretionary disposable incomes will be released upon demise of current middleman distributors. No longer shall it be syphoned off to 'intellectual property' monopoly dependent nations like the USA or to tax havens. Removal of distributor stranglehold (and cynical shaping in cahoots with the advertising industry) on popular culture shall enable many more people to dip toes into producing 'content' and earning respect and inflow of money from admirers. Although bland culture shall continue to have a following, the post-copyright regimen, based upon a truly competitive market, will encourage risk taking e.g. compare the independent film industry and 'content' displayed at film festivals with output from Hollywood.

    The above is largely generic. Yet impact upon computer/Internet related enterprise will be immense. Software, both published code and that decompiled/disassembled by users) shall be freely available for use, modification, and extension by anybody under discipline imposed by attribution (more or less as now applied by many open source producers). Monoliths like Microsoft shall go the way of the Dodo to be replaced by numerous cottage industries. Rentier economics shall be superseded by competition among the skilled and without need of the present thick overlay of corporate and conglomerate structure.

    Cotton weavers were displaced from their cottages by 18/19th century technological advance; those protesting, i.e. Luddites, were condemned as impeding progress. The tables are now turning. Large aggregates no longer are the best source of impetus for intellectual ferment and innovation (for instance Microsoft is stuck in the rut of Windows and Office). The mantle returns to cottages. Modern Luddites are to be found in computer behemoths, the recorded entertainment industries (plus streaming outfits), and companies like Elsevier living off borrowed time; instead of smashing equipment these seek to prop-up, ever more ineffectually, their defunct business models via buying legislators and imposing increasingly knotted 'intellectual property' law to protect themselves against competition. Gordian knots just beg to be sliced asunder.

    -----

    Released under the Creative Commons Attribution 4.0 international license.

    1. Mike Moyle

      Re: Innovation and 'rights' cannot co-exist

      Obligatory "The Oatmeal" post:

      https://theoatmeal.com/comics/exposure

    2. First Light

      Re: Innovation and 'rights' cannot co-exist

      Especially when the rights exist IN PERPETUITY, which is the case with some patents.

    3. Doctor Syntax Silver badge

      Re: Innovation and 'rights' cannot co-exist

      "Released under the Creative Commons Attribution 4.0 international license."

      Oh, the irony. It's only because you have copyrights in what you wrote that you have any right to attach that statement.

  8. a_yank_lurker

    Leisure Suit Larry Might Strike

    In a nutshell Linus is complaining about Larry's Minions and their antics not the technical merits of ZFS. To protect the vanilla Linux kernel he will not include the ZFS code in it because the Minions tend to sue whether they a valid claim or not. This is perfectly sensible decision until the Minions clarify the situation.

  9. Bronek Kozicki

    I use ZFS with Linux - and so should you

    There is a reason why ZFS license is incompatible with GPL, and that reason is to prevent ZFS being subsumed into Linux kernel. The incompatibility is in very superficial terms - both are open source licenses, where the users of the software are granted full rights (in simplistic terms) to the sources of the code they use.

    The incompatible license means that ZFS will not be driven solely by Linux kernel developers (like all other parts of the kernel) which is good news for ZFS users on BSD, MacOS and OpenIndiana (a.k.a. illumos). This is actually a very important property of any file system, and ZFS aside there is no other enterprise-grade filesystem with such guarantee (obviously, FAT does not count).

    It would be good if Linux kernel developers accepted that there are filesystems which need to work on many operating systems, including Linux. Sadly, the attitude which Linus is displaying now is similar to the one which is currently driving systemd development - "I know best and I do not care what you have to say".

    1. anonymous boring coward Silver badge

      Re: I use ZFS with Linux - and so should you

      Before agreeing with you, I would have to know:

      1) What kernel change is being introduced, and why, that makes ZFS incompatible?

      2) Why Oracle can't change ZFS to work with these changes (as is usually done when kernel changes are done)?

      Granted, internal kernel interface changes shouldn't be made lightly, but presumable Linus has good reasons sometimes to do so. (Unless Linux is a finished project, not to be changed ever again.)

      1. Anonymous Coward
        Anonymous Coward

        Re: I use ZFS with Linux - and so should you

        1) Ask Linus

        2) The open source version of ZFS is not developed by Oracle. Anyway, while code can be adapted, the issue is if users find they systems no longer work suddenly or they can't install a security patch until a fixed version of ZFS becomes available - that's a file system, you can't accept that.

        Think if Microsoft felt free to broke the drivers interface every time it releases a kernel patch...

      2. Bronek Kozicki

        Re: I use ZFS with Linux - and so should you

        1) The thread start at when one developer asks for a pair of functions kernel_fpu_begin/end not to be change exports to GPL-only. This pair can be used to enable SIMD computations for code running in the kernel mode and is used (among others) by ZFS on Linux to enable the acceleration of checksumming and compression. The kernel developers then basically respond "fuck off, we do not care if ZFS is open source too, because it is owned by evil Oracle" (which it is not, but hey). At the end of discussion you can see one of ZFS on Linux developers responding we will just use our SPL as we always did

        2) as explained before. Not a big deal, because ZFS on Linux has SPL project specifically for the purpose of re-implementing everything that ZFS needs from the kernel, with a GPL license and exported for ZFS to use. It is a hassle, and hopefully one day Greg, Linus and the rest will understand that they are not shooting in the Oracle's leg but their own (and other open source developers).

        I have a feeling that Linux kernel developers would happily remove the GPL-only limitation on an exported function for a closed source project like Nvidia drivers, because Linux Foundation members need CUDA working on Linux. One has to pay the bills, right?

        1. Bronek Kozicki

          Re: I use ZFS with Linux - and so should you

          ... oh wait, that thread is one year old. I guess the Linux kernel developers like to repeat the exercise annually; I wonder what they disabled this time.

        2. jilocasin
          Linux

          Re: I use ZFS with Linux - and so should you

          That's a bit harsh, no?

          It's not a matter of open source it's a matter of GPL. The kernel is licensed GPLv2. Everything that touches the kernel directly must also be licensed GPLv2. Linus has stated before that Linux is under no obligation to maintain compatibility with code that directly interacts with the kernel but is not maintained by the kernel team.

          To quote Greg Kroah-Hartman ( https://lkml.org/lkml/2019/1/15/305):

          "If you want to interact directly with Linux kernel

          code in kernel-space, then you have to abide by it's license, which is

          GPLv2". That's it. If you wish to use open source code by another

          license, wonderful, I'm not telling you what you can, and can not do,

          but please, do not violate the license of the code I contributed under

          GPLv2.

          So, no 'fuck off' and NVIDIA would, and has, gotten the same answer. License your code GPLv2, or in the case of hardware/drivers, open the documentation so that interested users can write drivers themselves. Otherwise it's on you to keep up with any changes in the kernel.

          1. Bronek Kozicki

            Re: I use ZFS with Linux - and so should you

            Right, i feel that apologies are due - sorry about that hyperbole. The fact of the matter remains that ZFS is not GPL and will never be, because it cannot be done - the dozens of contributors who added to OpenZFS code will not change their license for many reasons, one of them could be that, perhaps, not all of them can be contacted or are even alive. One the other hand, the Oracle flavor of ZFS could be perhaps re-licensed, but is irrelevant because almost no-one uses that and is not compatible with the flavour of ZFS used everywhere else outside of Oracle's walled garden - that is on Linux, MacOS and BSD.

            I get very annoyed when people invite Oracle to discussion on ZFS because it just feels disrespectful to dozens of people who contributed and continue to contribute to OpenZFS. Especially if the people in question should really know better and are as exposed as Linus Torvalds.

            Also I feel really strongly that Linux should recognise that it is a good thing that there is a common enterprise grade filesystem which works well on three major platforms and made some effort to not make working on this filesystem harder than it needs to be. This filesystem will be always outside of Linux tree because otherwise the users from other operating systems would be badly disadvantaged, and as a result it would eventually cease to be a common filesystem recognised and supported by other OSs.

  10. Bronek Kozicki

    Linus either is a little confused, or ...

    ... he actually means that branch of ZFS which is owned by Oracle and very few people pay any attention to it. Which would be strange to mention, really - because it is irrelevant.

    The development of ZFS has been happening since 2010 in illumos, and later in OpenZFS project - Oracle has no say in it.

  11. Denarius

    why is this a big issue ?

    can someone explain why ZFS is better than bttfs ? I used ZFS in Solaris and it is a great filesystem. On fly snapshots, mirroring, easy server migrations, update rollback. Also memory hungry. All the things that the linux filesystem writers are doing with GPL filesystems. My servers and PC bttfs does not seem memory hungry either. With on the fly block checking now in bttfs why is there a need for ZFS in Linux outside of Solaris shops ?

    1. Bronek Kozicki

      Re: why is this a big issue ?

      There are very few filesystems with bitrot prevention (i.e. checksumming and built-in erasure encoding and/or mirror copy). ZFS, btrfs and that's all I've heard about. Bitrot prevention matters because the probability of data loss, due to corruption of data-at-rest, increases with the size of the data. I've learned it the hard way, when some of the pictures I took long ago in RAW format were corrupted and had to be thrown out.

      There are few other enterprise-y features in ZFS that I like (nested filesystems, volumes, snapshots, send/receive, compression, since recently also native encryption), but the above is the big one. ZFS has been doing it much longer than btrfs, the codebase is tested against multiple different operating systems and is being used by the largest data stores on the planet (CERN, among others). It is also supported by multiple distributed filesystems such as Lustre and Gluster.

    2. jms222

      Re: why is this a big issue ?

      Because ZFS is stable and BTRFS (originally from Oracle but not Sun) isn't. Simple as that.

      (In case you say "look at Synology" they use Linux MD underneath as they daren't use BTRFS for that layer.)

  12. Sanguma
    Coat

    APIs, header files, definitions, 1+1=2

    The point of Linus' comments, asw I see it, is that Oracle owns the pre-forked ZFS APIs, header files, whatnot. As Oracle owns the Java programming language APIs. If Oracle can screw up Java so thoroughly by launching stupid law-suits against one of the Java language's greatest users, what's it going to do once OpenZFS gets embedded into Linux? The *BSD and OpenSolaris OSes are minor players in comparison to Linux, which is now one of the biggest OSes in the market (rightly or wrongly, I couldn't care less: I could do without some of the later additions, but it works for me.).

    In theory someone could take the data definitions and data structures and write a compatible file system clean-room from scratch, but that's not going to help much if even data defintions and data structures are litigated as "copyrightable" and a judge is stupid and ignorant enough to agree. Of course, it'll sink the *BSd and Opensolaris, but does anyone think Oracle cares?

    1. Anonymous Coward
      Anonymous Coward

      Re: APIs, header files, definitions, 1+1=2

      Just wait AWS clones fully Linux, Postgres, Git, etc. APIs without releasing a line of source code, then adds some proprietary features on top of it to ensure lock-in, and you will change your mind about copyright....

      1. Sanguma

        Re: APIs, header files, definitions, 1+1=2

        You mean, like Unix was cloned more than once (Coherent, Minix, a number of others) before Linus Torvalds cloned it and set out on the path to world domination with Linux? You mean, like relational databases use ODBC as their standardized API, and their own specialized APIs for more system-oriented utilities?

        And anyone who was watching the Microsoft versus DOJ circus in the late nineties knows that lock-in only occurs when you already have a major presence in the marketplace with any given item. AWS would need to convert at least 75% of the GIT users to their putative proprietary GIT-clone before being able to lock anyone in.

        I think you should rethink your arguments.

    2. Tom 7

      Re: APIs, header files, definitions, 1+1=2

      if Oracle win the court case to get to stop others playing with the Java API then, within a year I would expect j2c to grow by a couple of orders of magnitude and open source Java development to be reserved to Oracle PDQ.

      I've played with it on a couple of old Java projects and been surprised with the results - though I have to admit that could be because I know a reasonable amount of what's going on in both languages.

  13. anonymous boring coward Silver badge

    Looks like Linus said all that needs to be said, really. And with no swearing!

  14. Steve Davies 3 Silver badge
    Pint

    Keep on trucking Linus

    Given Oracle's stance on API Copyrights staying the hell away from anything Oracle has touched in the last 20 years seems like a good idea.

    Oracle is Evil and worse than Google.

    If I met Linus I'd buy him a beer.

  15. Anonymous Coward
    Anonymous Coward

    That’s nothing - I tried to use reiserfs

    But it’s murder to get it installed.

    1. Anonymous Coward
      Anonymous Coward

      Re: That’s nothing - I tried to use reiserfs

      Problem with reiserfs is that when it goes wrong it buries all the evidence.

      1. Anonymous Coward
        Anonymous Coward

        Re: That’s nothing - I tried to use reiserfs

        The following command should help recover all you need:

        % pat io excavate

  16. eLJaybud

    Well it would be nice if someone could revoke Oracle's use of the gpl due to their practices and litigious nature. Do they disclose all their code as open source that is connected with Linux?

    1. teknopaul

      When you say connected I presume you mean compiled including gpl parts. I don't think Oracle do much kernel Dev they resell redhat patches.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like