[citations desperately needed] [[for BTRFS more feature rich]]
OK I looked at the latest list. There was some nice stuff added to ZFS so I'll retract. But: File Change logs and Extents are still BTRFS only.
So, you're saying that it therefore makes more sense for Oracle to back the multi-year development of a completely new filesystem that is attempting to mimic a subset of the features of ZFS, rather than to... merely dual-license ZFS, which would allow ZFS to be included in Linux tomorrow?
Well ZFS is a module for most Linuxes. The Linux kernel supports variable modules. So that's not a big deal. That being said releasing ZFS under the GPL as well would be good as well as releasing the BTRFS.
That's irrelevant because in your example it's the minimum lag achievable with any protocol.
No it isn't. You don't make round trips you download client side and make sure to stay way ahead. Which is what MMORPGs do for example. That is the alternative.
Urelated to this, 83ms lag is acceptable for all purposes other than games and graphics/CAD editors
No it isn't. It isn't acceptable for almost any touch based application it is too confusing for the end user. I don't find that sort of lag acceptable in business productivity applications.
- web browsers, for example, have higher latency,
Yes and web browsers have spent 20 years pushing complex code client side to emulate low latency situations.
Desktops don't have touch interfaces.
Some don't. 4-6% of laptops sold 4Q2012 did. That number is going up.
Desktops don't have voice interfaces.
Yes they do. They are migrating over from phones. Apple had to introduce several iOS voice interfaces to OSX and Microsoft has had them for years.
If that was true, no remote access systems would ever be usable, as 10ms latency is almost never available over the Internet.
See above.
In reality, I have perfectly usable X11 without wrappers over 25-30ms connection between my home and work.
Which is likely usable with a mouse. Try it with a touch based application where the applications follows your finger (i.e. more than just pushing buttons) like a paint application.
and no protocol, existing or possible, would satisfy your requirements.
I've given several that work fine. Java applets, Flash, Javascript, thick client.... Wayland is using something like RDP which is working the latency problem by making sure that most activities don't require a round trip.
Agree with what you wrote. I might make it the blessed 4 to include selling "non GPLed version" to cover the Qt, MySQL case rather than bundling that as a service, but... its fine either way. Regardless any attempt to monazite the GPL code directly is a disaster, we agree. Companies often don't realize when they buy this sort of IP that a few puny developers run off and fork and suddenly there multi-million dollar asset is worth 0. They aren't used to having to keep engineers happy as a critical business process.
As an aside, the PDF reader isn't GPLed it is closed source but free. The.pdf spec is fully open and public. I should also mention reasonably well written. Anyone is free to create their own.pdf tools from the spec, and there are lots of open source tools. They never published the spec for Flash, but they had competing business models there (Flash specific issue that's a rabbit hole as far as this discussion). And I should mention.pdf isn't hated by open source people. TeX which is the basis of GNU documentation (texinfo) and is heavily used by the Unix community now uses pdfTeX as the engine (i.e. TeX code that compiles to.pdf) and most of the extensions of the last decade are on top of.pdf like Omega / XeTeX.
companies that get bought for insane amounts of money usually end up getting turned into a mess by the buyer if they can't see a quick enough ROI and as companies like Red Hat have shown you have to be in it for the long haul with the GPL, you can't just flip companies for quick cash it just doesn't work that way.
I would agree with that. I don't think there is any quick profit making scheme with GPL that I know of. RedHat BTW has been a company that has been able to acquire a turn a nice profit on acquisitions.
"Stallman poisioned the well by using propaganda techniques to make people into GPL zealots
That's very different than the lawsuit. And I think no question that Stallman, had a huge impact on making people believe in the GPL and changing people's mind. But if you look at the LAMP stack: GPL: Linux, MySQL BSD/MIT style: Apache, Perl
So at least by the early 1990s were it the case that the BSD/MIT licenses were far superior there were ways people could have seen it. Looking back 18 years later... Perl hit roadblocks to the point that the "P" became PHP, which is also BSD and everyone else exploded with success.
(I think this is a very partial truth, myself, and that there are structural differences that are much more important.)
That's my opinion that the structural differences are what mattered. Now it is possible that without the lawsuit Linux never meaningfully exists, i.e. it is just a kernel from the Minix community for 386 capabilities and never goes beyond being a hobbyist kernel... i.e. Linux stays at the LINIX (Linus' Minix stage).
"The Linux community had a structure that was better able to scale development and evangelization".
I'd agree with that one.
There was very little focus on recruiting in the BSD community - there was a "put it out there and let them use it if they want to"
Exactly! That's a huge differences from the Linux community which was very evangelistic. There were no BSD installfests. The BSD community was ambivalent about attracted youth in the 1990s. The 14 year old who installed Linux in 1994, is a 32 year old system administrator today who knows how to configure Sys V and doesn't know BSD style admin. BTW this is still true. My 13 year old daughter's "boyfriend" talks about various Linux gimmicks used to attract people all the time. He's excited and enthusiastic about Linux, no one from the BSD community wants to deal with a 13 year old who can program a little Python and so figures he's an OS expert.
Right now far and away the most popular desktop Unix is a BSD, OSX. That is the system for most computer enthusiasts today, not Windows. Why would the BSD community not work closely with the Darwin community to make the transition smooth? They still don't recruit.
Overall, it is very hard to estimate which of these effects is larger - what I personally have seen (from watching the BSD community from the inside and the Linux community from the outside) would make me estimate the positive effect of freedom on company contributions to be much larger than the positive effect of restrictions, but people on the opposite side tend to consistently make the opposite estimation.
I think both sides see each other's progress. The Linux community does follow XFree86 / X.org, Apache, Perl / PHP / Ruby / Python, SSH, Postgres.... The BSD community follows Gnome, KDE, GCC, MySQL... If I had to guess the reason for the difference I think it has more to do with how they value contributions. The BSD community wants quality code, and Linux wants innovative code. A lot of time GPL contributors are company B getting code from A who was working on something entirely different and donating to C which B doesn't care about. The way that for example Apple with GCC wrote a huge chunk of the code that the Sony Playstation used for their game compilers. The BSDs would be very antsy about something like that since obviously Apple has no intention of maintaining code for Sony.
This is a very chaotic process that I don't think the BSDs would be comfortable with. So they tend to value long term contributors and devalue people who jump in for a few years dump lots of complex code, do an integration get it partially working and leave. Conversely Linux people, because they jump from distribution to distribution don't experience continuity and thus devalue it.
As for user friendliness: I switched from Linux to BSD back i
If it is any consolation Oracle has had problems with companies like Peoplesoft as well. They don't do well with merges it ain't just the GPL. If I had to guess I'd assume that primarily Oracle doesn't care about making the money directly. I think they were buying Sun's customer base a huge percentage of whom were Sun/Oracle so that they could switch them gradually to x86/Oracle rather than potentially losing them to x86/Postgres or AIX/DB2. I think the complexity of Sun caught them off guard.
But no question Oracle is a terrible fit for open source culture. I don' think they really understood that products would fork within weeks of them being jerkish and once forked regaining control would be difficult. Oracle also bought MySQL and Berkley DB. In the case of Berkley they kept the old developers on, and in the case of MySQL they got forks.
An example of a company that does work really quite well with open source acquisitions is Adobe. Adobe gets the idea of open sourcing one part of a product and selling surrounding products because so many of their core products (pdf, flash, shockwave...) had free readers. They aren't doing your blessed 3 either they are in the: give away part and keep the rest. They use the GPLed (and other open source and or free) product as advertising for their non GPLed components.
As far as the blessed 3 BTW both Trolltech and MySQL were good counter examples. They both made money by giving away the GPLed version and selling the non-GPLed version. Both very successful. That model seems to work quite well for component software. When Nokia moved to the LGPL, which is liberal enough that people were perfectly happy with Qt's free version then it became essentially a worthless company. Digia paid $5m to land all kinds of Qt consulting gigs (selling services / support).
Apple doesn't have a filesystem strategy today. They have had this problem since OSX 10.5. Steve Jobs might have an ego about getting punked, but he wasn't crazy.
Apple supposedly had problems with their internal drives, where it mattered. As for shallow understanding of ZFS they hired several of the world's leading experts on it for their port.
Wow, you haven't deployed either of them at scale, have you?
Nope, parroting other's opinions.
Since Oracle is still selling ZFS, it would be hard to see why they would put significant resources into creating a free alternative at this point.
AFAIK BTRFS is better for Oracle and more feature rich. Ultimately it is not to Oracle's advantage that Z-OS, I-OS (the IBM one) and AIX are the OSes for very large storage. That's how they could lose to DB2 / Netezza.
Postgres is a comparable to about Oracle 8 in terms of features, there are still very good reasons to use Oracle. Besides my point was that even if the database business does die they have many many more products now.
I'm not being dense I don't think Google was being a jerk. Google was doing what they thought was best. Google doesn't work for Sun. Just because Sun wants something doesn't mean Google is being for not doing it. I'm sure Microsoft wasn't fond of Open Office, that doesn't mean that Sun was being a jerk for not canceling the project because Sun objected.
I'm hard pressed to see what the difference is between what Google did when they created Davlik and what Sun did when they made Java from Oak.
Stallman has defended right to fork long after XEmacs. At the time what he said was:
The long delay in releasing Emacs 19 is the FSF's fault. (In some sense, therefore, mine.) While it's regrettable that there are multiple versions, I can't blame people for filling the gap that the FSF left. One of the goals of the copyleft is to allow people to do this--so that one central maintainer's lapse does not hold back the rest of the community.
ZFS is open source. Turns out it sucks on inexpensive (x86 quality) hardware, in ways that are unfixable even by smart people. Apple lost years proving that.
Their are penalties for frivolous lawsuits. The problem is there is a huge range between bad lawsuits and frivolous. The bar is high for frivolous.
As an American I'd like to see more suits have to pass a quick summary judgement on viability. And much stronger restrictions on changing the initial fillings.
We just had a test of this on a major GPL company. Trolltech was sold to Nokia for $153m. Their product was GPL / commercial and profitable, though $150m was grossly overpaying. Nokia LGPLed it which killed the 2 distribution model and thus Trolltech's way to make money on software. When Nokia sold Trolltech to Digia I think it was about $5m total.
That's a myth from the BSD community. Absolutely the LAMP stack wasn't the BAMP stack. Possibly you can argue that Linux was ahead in 1994 because of the lawsuit but what about the 18 years since then?
There were niches like embedded where BSD was well established that they lost to Linux. BSD lost to Linux because:
a) They didn't care much about appealing to Windows power users who became the base of Linux. They focused on recruiting from the smaller Unix community. b) They didn't have the GPL so they never got the industrial cooperation from hardware players that Linux got, contrary to their theories about licensing. c) Their product is too damn hard to use, even today twenty years later.
The freedom to run the program means the freedom for any kind of person or organization to use it on any kind of computer system, for any kind of overall job and purpose, without being required to communicate about it with the developer or any other specific entity. In this freedom, it is the user's purpose that matters, not the developer's purpose; you as a user are free to run the program for your purposes, and if you distribute it to someone else, she is then free to run it for her purposes, but you are not entitled to impose your purposes on her.
Sun had a legitimate right to trademark the name "java" and control what could be called Java. But davlik is "java like" I don't see where Sun has any rights over a Java like language.
83 ms is totally unacceptable. I think we talked about this before and Microsoft in their UI training is finding that when it comes to touch more than 1ms is detectable by humans. Having worked on voice systems 83 ms is very confusing for humans, though still easily understandable. In terms of response to input the this should be about 10 ms on a mouse and 1 ms on a touchscreen.
GTA uses the direct rendering subsystem. It isn't meaningfully using X11.
There are plenty of examples of running benchmarking apps that can do similar OpenGL calls in OSX vs. X11 or Windows and I haven't seen X11 win one of those in a decade.
Excluding the last item, nothing in your list is impossible or more difficult with Wayland. Wayland doesn't help, Wayland doesn't hurt.
For the last item Wayland is using something like RDP. There are some Wayland hooks but fundamentally the core of this will be KDE and Gnome for the next few years.
No you can't retrofit it. Fundamentally the Linux community is going to have to choose whether to design applications for Wayland in which case they won't run on X11, or whether to design applications for X11 in which case they won't take advantage of Wayland's features.
Slashdot Mirror
[citations desperately needed] [[for BTRFS more feature rich]]
OK I looked at the latest list. There was some nice stuff added to ZFS so I'll retract. But: File Change logs and Extents are still BTRFS only.
So, you're saying that it therefore makes more sense for Oracle to back the multi-year development of a completely new filesystem that is attempting to mimic a subset of the features of ZFS, rather than to... merely dual-license ZFS, which would allow ZFS to be included in Linux tomorrow?
Well ZFS is a module for most Linuxes. The Linux kernel supports variable modules. So that's not a big deal. That being said releasing ZFS under the GPL as well would be good as well as releasing the BTRFS.
That's irrelevant because in your example it's the minimum lag achievable with any protocol.
No it isn't. You don't make round trips you download client side and make sure to stay way ahead. Which is what MMORPGs do for example. That is the alternative.
Urelated to this, 83ms lag is acceptable for all purposes other than games and graphics/CAD editors
No it isn't. It isn't acceptable for almost any touch based application it is too confusing for the end user. I don't find that sort of lag acceptable in business productivity applications.
- web browsers, for example, have higher latency,
Yes and web browsers have spent 20 years pushing complex code client side to emulate low latency situations.
Desktops don't have touch interfaces.
Some don't. 4-6% of laptops sold 4Q2012 did. That number is going up.
Desktops don't have voice interfaces.
Yes they do. They are migrating over from phones. Apple had to introduce several iOS voice interfaces to OSX and Microsoft has had them for years.
If that was true, no remote access systems would ever be usable, as 10ms latency is almost never available over the Internet.
See above.
In reality, I have perfectly usable X11 without wrappers over 25-30ms connection between my home and work.
Which is likely usable with a mouse. Try it with a touch based application where the applications follows your finger (i.e. more than just pushing buttons) like a paint application.
and no protocol, existing or possible, would satisfy your requirements.
I've given several that work fine. Java applets, Flash, Javascript, thick client .... Wayland is using something like RDP which is working the latency problem by making sure that most activities don't require a round trip.
Agree with what you wrote. I might make it the blessed 4 to include selling "non GPLed version" to cover the Qt, MySQL case rather than bundling that as a service, but... its fine either way. Regardless any attempt to monazite the GPL code directly is a disaster, we agree. Companies often don't realize when they buy this sort of IP that a few puny developers run off and fork and suddenly there multi-million dollar asset is worth 0. They aren't used to having to keep engineers happy as a critical business process.
As an aside, the PDF reader isn't GPLed it is closed source but free. The .pdf spec is fully open and public. I should also mention reasonably well written. Anyone is free to create their own .pdf tools from the spec, and there are lots of open source tools. They never published the spec for Flash, but they had competing business models there (Flash specific issue that's a rabbit hole as far as this discussion). And I should mention .pdf isn't hated by open source people. TeX which is the basis of GNU documentation (texinfo) and is heavily used by the Unix community now uses pdfTeX as the engine (i.e. TeX code that compiles to .pdf) and most of the extensions of the last decade are on top of .pdf like Omega / XeTeX.
companies that get bought for insane amounts of money usually end up getting turned into a mess by the buyer if they can't see a quick enough ROI and as companies like Red Hat have shown you have to be in it for the long haul with the GPL, you can't just flip companies for quick cash it just doesn't work that way.
I would agree with that. I don't think there is any quick profit making scheme with GPL that I know of. RedHat BTW has been a company that has been able to acquire a turn a nice profit on acquisitions.
Well written counter argument.
"Stallman poisioned the well by using propaganda techniques to make people into GPL zealots
That's very different than the lawsuit. And I think no question that Stallman, had a huge impact on making people believe in the GPL and changing people's mind. But if you look at the LAMP stack:
GPL: Linux, MySQL
BSD/MIT style: Apache, Perl
So at least by the early 1990s were it the case that the BSD/MIT licenses were far superior there were ways people could have seen it. Looking back 18 years later... Perl hit roadblocks to the point that the "P" became PHP, which is also BSD and everyone else exploded with success.
(I think this is a very partial truth, myself, and that there are structural differences that are much more important.)
That's my opinion that the structural differences are what mattered. Now it is possible that without the lawsuit Linux never meaningfully exists, i.e. it is just a kernel from the Minix community for 386 capabilities and never goes beyond being a hobbyist kernel ... i.e. Linux stays at the LINIX (Linus' Minix stage).
"The Linux community had a structure that was better able to scale development and evangelization".
I'd agree with that one.
There was very little focus on recruiting in the BSD community - there was a "put it out there and let them use it if they want to"
Exactly! That's a huge differences from the Linux community which was very evangelistic. There were no BSD installfests. The BSD community was ambivalent about attracted youth in the 1990s. The 14 year old who installed Linux in 1994, is a 32 year old system administrator today who knows how to configure Sys V and doesn't know BSD style admin. BTW this is still true. My 13 year old daughter's "boyfriend" talks about various Linux gimmicks used to attract people all the time. He's excited and enthusiastic about Linux, no one from the BSD community wants to deal with a 13 year old who can program a little Python and so figures he's an OS expert.
Right now far and away the most popular desktop Unix is a BSD, OSX. That is the system for most computer enthusiasts today, not Windows. Why would the BSD community not work closely with the Darwin community to make the transition smooth? They still don't recruit.
Overall, it is very hard to estimate which of these effects is larger - what I personally have seen (from watching the BSD community from the inside and the Linux community from the outside) would make me estimate the positive effect of freedom on company contributions to be much larger than the positive effect of restrictions, but people on the opposite side tend to consistently make the opposite estimation.
I think both sides see each other's progress. The Linux community does follow XFree86 / X.org, Apache, Perl / PHP / Ruby / Python, SSH, Postgres.... The BSD community follows Gnome, KDE, GCC, MySQL... If I had to guess the reason for the difference I think it has more to do with how they value contributions. The BSD community wants quality code, and Linux wants innovative code. A lot of time GPL contributors are company B getting code from A who was working on something entirely different and donating to C which B doesn't care about. The way that for example Apple with GCC wrote a huge chunk of the code that the Sony Playstation used for their game compilers. The BSDs would be very antsy about something like that since obviously Apple has no intention of maintaining code for Sony.
This is a very chaotic process that I don't think the BSDs would be comfortable with. So they tend to value long term contributors and devalue people who jump in for a few years dump lots of complex code, do an integration get it partially working and leave. Conversely Linux people, because they jump from distribution to distribution don't experience continuity and thus devalue it.
As for user friendliness: I switched from Linux to BSD back i
If it is any consolation Oracle has had problems with companies like Peoplesoft as well. They don't do well with merges it ain't just the GPL. If I had to guess I'd assume that primarily Oracle doesn't care about making the money directly. I think they were buying Sun's customer base a huge percentage of whom were Sun/Oracle so that they could switch them gradually to x86/Oracle rather than potentially losing them to x86/Postgres or AIX/DB2. I think the complexity of Sun caught them off guard.
But no question Oracle is a terrible fit for open source culture. I don' think they really understood that products would fork within weeks of them being jerkish and once forked regaining control would be difficult. Oracle also bought MySQL and Berkley DB. In the case of Berkley they kept the old developers on, and in the case of MySQL they got forks.
An example of a company that does work really quite well with open source acquisitions is Adobe. Adobe gets the idea of open sourcing one part of a product and selling surrounding products because so many of their core products (pdf, flash, shockwave...) had free readers. They aren't doing your blessed 3 either they are in the: give away part and keep the rest. They use the GPLed (and other open source and or free) product as advertising for their non GPLed components.
As far as the blessed 3 BTW both Trolltech and MySQL were good counter examples. They both made money by giving away the GPLed version and selling the non-GPLed version. Both very successful. That model seems to work quite well for component software. When Nokia moved to the LGPL, which is liberal enough that people were perfectly happy with Qt's free version then it became essentially a worthless company. Digia paid $5m to land all kinds of Qt consulting gigs (selling services / support).
No. There is nothing like Peoplesoft or Siebel. Oracle financials is excellent. I would never describe those as poor.
Apple doesn't have a filesystem strategy today. They have had this problem since OSX 10.5. Steve Jobs might have an ego about getting punked, but he wasn't crazy.
Apple supposedly had problems with their internal drives, where it mattered. As for shallow understanding of ZFS they hired several of the world's leading experts on it for their port.
Wow, you haven't deployed either of them at scale, have you?
Nope, parroting other's opinions.
Since Oracle is still selling ZFS, it would be hard to see why they would put significant resources into creating a free alternative at this point.
AFAIK BTRFS is better for Oracle and more feature rich. Ultimately it is not to Oracle's advantage that Z-OS, I-OS (the IBM one) and AIX are the OSes for very large storage. That's how they could lose to DB2 / Netezza.
Postgres is a comparable to about Oracle 8 in terms of features, there are still very good reasons to use Oracle. Besides my point was that even if the database business does die they have many many more products now.
R was a research application it wasn't commercial. Oracle was very very different. There wasn't any code in common.
I'm not being dense I don't think Google was being a jerk. Google was doing what they thought was best. Google doesn't work for Sun. Just because Sun wants something doesn't mean Google is being for not doing it. I'm sure Microsoft wasn't fond of Open Office, that doesn't mean that Sun was being a jerk for not canceling the project because Sun objected.
I'm hard pressed to see what the difference is between what Google did when they created Davlik and what Sun did when they made Java from Oak.
Stallman has defended right to fork long after XEmacs. At the time what he said was:
Just in case anyone believes this:
Oracle founded 1977
DB2 first version 1983
ZFS is open source. Turns out it sucks on inexpensive (x86 quality) hardware, in ways that are unfixable even by smart people. Apple lost years proving that.
But http://zfsonlinux.org/
But Oracle's BTRFS plays the same role and is even better.
What is competitive with Oracle for large relational databases other than DB2? Oracle financials, Peoplesoft, Siebel, JD Edwards...
I think they are fine for now.
Their are penalties for frivolous lawsuits. The problem is there is a huge range between bad lawsuits and frivolous. The bar is high for frivolous.
As an American I'd like to see more suits have to pass a quick summary judgement on viability. And much stronger restrictions on changing the initial fillings.
We just had a test of this on a major GPL company. Trolltech was sold to Nokia for $153m. Their product was GPL / commercial and profitable, though $150m was grossly overpaying. Nokia LGPLed it which killed the 2 distribution model and thus Trolltech's way to make money on software. When Nokia sold Trolltech to Digia I think it was about $5m total.
That's a myth from the BSD community. Absolutely the LAMP stack wasn't the BAMP stack. Possibly you can argue that Linux was ahead in 1994 because of the lawsuit but what about the 18 years since then?
There were niches like embedded where BSD was well established that they lost to Linux. BSD lost to Linux because:
a) They didn't care much about appealing to Windows power users who became the base of Linux. They focused on recruiting from the smaller Unix community.
b) They didn't have the GPL so they never got the industrial cooperation from hardware players that Linux got, contrary to their theories about licensing.
c) Their product is too damn hard to use, even today twenty years later.
Let me quote RMS on that one
Sun had a legitimate right to trademark the name "java" and control what could be called Java. But davlik is "java like" I don't see where Sun has any rights over a Java like language.
83 ms is totally unacceptable. I think we talked about this before and Microsoft in their UI training is finding that when it comes to touch more than 1ms is detectable by humans. Having worked on voice systems 83 ms is very confusing for humans, though still easily understandable. In terms of response to input the this should be about 10 ms on a mouse and 1 ms on a touchscreen.
GTA uses the direct rendering subsystem. It isn't meaningfully using X11.
There are plenty of examples of running benchmarking apps that can do similar OpenGL calls in OSX vs. X11 or Windows and I haven't seen X11 win one of those in a decade.
Excluding the last item, nothing in your list is impossible or more difficult with Wayland. Wayland doesn't help, Wayland doesn't hurt.
For the last item Wayland is using something like RDP. There are some Wayland hooks but fundamentally the core of this will be KDE and Gnome for the next few years.
No you can't retrofit it. Fundamentally the Linux community is going to have to choose whether to design applications for Wayland in which case they won't run on X11, or whether to design applications for X11 in which case they won't take advantage of Wayland's features.