Yes probably. I'd assume Gnome 4 works on X11 but worse and support is dropped entirely or almost entirely by Gnome 5. Similarly KDE 6 works better on Wayland with KDE 8 dropping support for X11.
People who want to use X11 applications in 2030 are going to have run legacy applications almost exclusively.
Why don't you try considering how people are actually using their computers these days. They want access to their apps from the "cloud". The success of Google Docs, Office365, Photoshop in the cloud,
All those push data to a local rendering system. Precisely the opposite of network transparency. Your own examples prove how bad the X approach is for today's applications.
.[RDP] untenable across a slow or latent WAN link.
You have that backwards. RDP is far more usable across a high latency link than X because round trips are minimized. Network to work requires very low latencies that are impossible to achieve over WANS. And thus X11 has all sorts of work arounds which are part of today's codebase which can't really work because of fundamental contradictions in the design that go down to the core.
a) CPU is on the same bus as the display b) CPU is on a low latency LAN with the display c) CPU is on a high latency WAN with the display
Case (a) is the most common followed by case (c). X was optimized for case (b). Because that was the most common 30 years ago. Case (b) should still be covered but the system should be designed around (a) with strong support for (c).
A respectful comment from the anti-Wayland side! Well done.
Anyway the reality is that Wayland guys wanted to get a prototype working to make sure there weren't any problems with remote... i.e. in theory a remote system could meet Wayland's "every frame is perfect". They don't actually care if in reality the frames are even visible for end users right now. That is the need it to work as essentially a unit test they don't need it to work right now for end user.
They do agree remote usage is an important feature and it is something they intend to add. Remember that applications, especially those that are often used remotely, are likely to be somewhat delayed in how long before they switch to Wayland only versions. Also VNC works much better with Wayland. So in their mind they have years to fix the remote problem. They do agree it is something on their todo list.
They implement a far better remote usage system They didn't implement network transparency. If you don't understand the distinction maybe you should assume that the developers who work on X11, which is the Wayland team, know a bit more than you do about how X11 works
Actually no. You gain remote access. Wayland has something like Windows RDB. That hasn't been true in years. As for X emulation... you can run an X11 server on Wayland same as you do on Windows/Aero or OSX/Aqua/Quartz. This one has never been true.
1) Developers being able to address new problems more easily because they have a more sane codebase 2) Better local video in a variety of ways. For example media coherence (i.e. video and sound won't get out of sync) 3) Better handling of latency so better remote desktop over Wans. 4) 32,768 pixel limit (not a problem yet but with 5k displays becoming more standard we are getting close to having a problem) 5) Validated Windows trees so windows won't get lost as often
We were talking about both. GP had said which induced response to what you are quoting, "In the 1980's Unix nearly DIED on the scattered hardware base. That was the biggest cause of all it's problems at the time." He was clearly talking abut big box Unixes in the Workstation era. Which is btw what I was talking about as well the "1980s workstation market"
I get that. But the GPL was part of a social movement that used copyright law to advance its aims. Most licenses are used to make someone some profit on IP.
uh, no, we're not dipping into the public purse to clean up your mess, but we will be pursuing your former stakeholders to accomplish the same.
Actually that almost never happens by design. One of the crucial distinctions between a corporation and a partnership is that a corporation is a legal person and thus doesn't incur liability for its investors. What you want would require no corporations and just partnerships, or very strict limits on partnerships... more like 18th century corporate law.
Re:Yea- we need the GPL or we won't get sources
on
On Being Pro-GPL
·
· Score: 1
I'd agree. If that works then RMS's arguments about software suddenly apply to hardware.
Let me just point out the OSX machines are essentially variants on NeXT. NeXT released their first model spring of 1987 and was working hard on their stuff in 1985. Spring 1987 is probably around the time Vector would have released their 2nd version of the their cool stuff on the drawing board when customers would have had to transition.
Now if you are going to be transitioning in 1987... and the first fully object oriented graphic operating system which has concepts like the web (in primitive form) and email native is out how well does Vector's enhanced CP/M do? A very primitive version of OSX is where the bar would have been for them.
_____
You are right about the 1983 Commodore price war comment, that is what I meant. Commodore is a more complex situation since in the 1990s they were the low end and the grey box makers wanted to go down market. I suspect they could have gone 3rd world, and aimed for computers for $50. Something like the One Laptop Per Child initiative from a decade later . Or they could have been a player like Palm, Danger or BlackBerry were with early handhelds and heck that market now is much bigger than PCs.
I don't know what Commodore was thinking about during the last days. But given that SGI didn't make it...
Let me point out most of the Unixes went down in the flames in the mid 1990s, not the 1980s, when the workstation and server market consolidated (around NT and Linux). Unix thrived on diversity.
If there had been ten quirky brands of hardware environments to target, there wouldn't have been the critical mass of a hardware install base for the Linux hackers to focus on. The fact that there was an 'industry standard hardware base' in the form of the x86 clone motherboard made it easy for Linux to get a base to run on.
I don't think you are right here in a more general sense. Obviously something like the Linux kernel is a bad choice. But imagine the Linux toolset (i.e. the rest of the operating system) with at worse a slightly more hardware diverse kernel: NetBSD, QNX... Nothing much has to change. Again we know Linux can thrive in that environment, it does now and UNIX before it did, which I'll address below.
Many IT operations had old boxes that management had abandoned, for maverick techs to install Linux on. Where it Just Worked
Sure. And YellowDog Linux "just worked" on PPC hardware. You pick a distribution for your hardware. People would be used to that since that's what they were doing anyway.
In the 1980's Unix nearly DIED on the scattered hardware base. That was the biggest cause of all it's problems at the time.
Died? Huh? That was the highpoint for commercial Linux. Commercial Linux in the 1980s was displacing other systems that had evolved in hardware monocultures because it handled the rapidly evolving workstation / server hardware. Your history is just backwards here.
Now certainly the success of Unix led to some problems that Open Systems needed to address, but that was a problem with success.
I agree with your first point about convergence. The Microsoft / Intel / Western Digital Standard for platform (IBM essentially) was a huge deal for hardware standardization.
But I'm not sure about your second point. Linux exists comfortably on dozens of hardwares. Linux is a diverse ecosystem not a monoculture. Why would Linux (or something like it) have not done better if there were 10 quirky brands about? Heck that's pretty much the 1980s workstation market and there is where Unix pretty much evolved from.
I think Vector is a perfect example of disruptive technology. When they first choose CP/M it was the right choice. They couldn't switch too early from CP/M to a more advanced OS because that would be a downgrade. When more advanced OSes could run the same software they were caught hopelessly behind. Fundamentally Vector didn't own any technology that was unique to it. They had a mostly generic CP/M box with a few tweaks. One of the last CP/M manufacturers. I don't know the system but it wouldn't shock me if they made one of the best business CP/M boxes out there, that's what it takes to become a dominant player. But let's forget that Commodore launched a major price war in 1983 and was driving CP/M machines down to 0. I don't know if Vector could have survived Commodore even if there was no disruption.
Anyway If you look at competitors from the same time period like SGI, HP or Sun you can see there were possibilities up market. Or your example of Commodore, with Amiga or Apple with Mac. Others were able to transition. It clearly was possible to make a box that wasn't an IBM that was successful. They just didn't do much of anything to get there in time.
May 1982 Sun-1 Workstation Jan 1983 Apple Lisa Jan 1984 Apple Mac Early 1985 SGI Iris 1000 July 1985 Amiga 1000
And their engineers had stuff in 1985 on the drawing board which was definitely better than CP/M and moving towards DOS? This wasn't just a failure of executive management. They deserved to die.
There are orders of magnitude more embedded devices than laptops and desktops. There are also more servers than desktop or laptops. So you can't make an appeal to quantity and then argue for desktops and laptops. Relative to those two categories the ones you are talking about are niche.
There are also far more Android tablets which appear to be replacing much of the consumer laptop / desktop space. There are more smartphones and even the Apple this year will outsell PCs Android having crossed years ago.
As for people in the developing world even excluding smartphones and tables given sales data and longer longevity about 20% are running Macs. In terms of profits Macs take about 85-92% for many years. So I'm not sure what the burn is. Windows is the dominant player in a large niche which is rapidly shrinking being successfully attacked from both above and below.
There were until recently niches were OS/2 thrived. You can always define things narrowly, but using most broad based metrics not constructed specifically to be a niche Windows is not dominant.
Apple's changes to GCC were released, but mostly lived in a separate branch in GCC's svn.
Apple was aggressively merging until November 2005 when their LLVM changes got blocked. June 2005 they had already announced the switch to Intel. I think we are disagreeing on facts here. Yes a lot of their changes weren't merged in the end but that was because the GCC developers didn't like the direction they were going. Early on they got merged. When really matters.
I'm not sure how second contradicts what I wrote.
As for Sony I was talking PS3 they used Apple's early work on GCC.
Not a great example. WebKit was originally released as a bunch of code dumps that were basically impossible to merge with upstream KHTML. It was the promise of outside developer interest (after many years of developing it with no community involvement), not the license, that caused them to release it as a proper open source project (i.e. with a public revision control system). Oh, and the Apple-original bits of WebKit (i.e. those not derived from KHTML) are BSD licensed...
How does that disprove anything? The question is cooperation. Webkit clearly has lots of people working with it. That's what GPL does:
KHTML existed and was open Apple used KHTML and developed Webkit. Webkit existed and was open Webkit attracted contribution
Under a BDS license for KHTML Webkit is never open source.
I disagree the GPL does nothing, I agree with your case.
Another common case than the one you cite is this. A, B, C and D are all building software for some secondary goal. For example Apple computer was working on GCC's optimizations for PPC because they wanted Mac software including OSX to run faster on PPC hardware, not because they wanted to get into the complier business. The GPL forced Apple to release all their changes to GCC even though they may have seen these as a competitive advantages and didn't release similar code bases when the code was BSD. This GCC release allowed Sony to leverage them on their game design system. Apple was forced to cooperate with Sony even though Apple made had no desire to. Another example involving Apple is the release of Webkit because of the KHTML code BTW.
That's certifying that the hardware's embedded software is free not that the hardware itself is free.
Re:Yea- we need the GPL or we won't get sources
on
On Being Pro-GPL
·
· Score: 2
Device driver yes. RMS wasn't frustrated that the Xerox 9700 itself wasn't free.
Re:Few people understand the economics
on
On Being Pro-GPL
·
· Score: 1
I'd say copyright assignment. Relicensing is dangerous because it makes standing complex. With relicensing you are just another license holder equal to the license violator. How do you prove damages or have the right to settle violations if you don't have copyright?
Re:Free software locks down users?
on
On Being Pro-GPL
·
· Score: 1
Cisco systems paid damages (a donation) to the FSF for Linksys violations. NuSphere lost a bunch to MySQL. I could probably come up with more examples.
You are misunderstanding the quote. The quote from the original article was about how it is getting to hard to create a platform from scratch and the with rise of free platforms most people creating proprietary software are starting with a free (BSD usually) licensed system and extending it as a commercial product.
Re:The geek on the lecture circuit.
on
On Being Pro-GPL
·
· Score: 1
I agree with your point entirely ethically.
However... I'm just going to nitpick. The hardware definition of open platform is a combination of interoperability, portability, and open software standards that allow you to have supplier choice. It has nothing to do with you personally being one of those vendors. The relative ease of being able to join the Android ecosystem (i.e. it becomes practical at a few hundred thousand units) makes it an open platform regardless of how locked down each of those units is individually. You may not like the definition but that is the definition.
I'd also like to mention three area you forgot which is incredibly open.
1) Clouds especially IaaS. Those are running lightweight OSes designed to freely run all sorts of client OSes. That's a huge area of freedom replacing the big box proprietary systems. 2) Moreover this server based technology scales down nicely to both server and desktop. So that while you may not be able to freely run different base OSes you can freely run different client OSes and those are free distributed on free VM managers. 3) Finally on phones Android itself is constructed to allow for different environments, Google Play is replaceable or combinable. You can have Android.NET (if Microsoft wanted it which they are playing with) and Xiaomi's business model is based on this.
A writes a GPL application. B uses it and writes something on top. B gives a modified binary to C with no notice to A. C distributes it wildly including to D. How do D or A force B to release his changes? Remember C doesn't have the source changes.
Hardware drivers are not a problem of software. Those are given away for free mostly. The problem is the open source aspect not the monetization aspect. Your 2 year windows wouldn't solve the driver the problem. As for games... I'm not sure what incentive game writers have for ever doing GPL. They are in the one off commercial software business.
Slashdot Mirror
Yes probably. I'd assume Gnome 4 works on X11 but worse and support is dropped entirely or almost entirely by Gnome 5. Similarly KDE 6 works better on Wayland with KDE 8 dropping support for X11.
People who want to use X11 applications in 2030 are going to have run legacy applications almost exclusively.
All those push data to a local rendering system. Precisely the opposite of network transparency. Your own examples prove how bad the X approach is for today's applications.
You have that backwards. RDP is far more usable across a high latency link than X because round trips are minimized. Network to work requires very low latencies that are impossible to achieve over WANS. And thus X11 has all sorts of work arounds which are part of today's codebase which can't really work because of fundamental contradictions in the design that go down to the core.
Get a grip. There are 3 use cases
a) CPU is on the same bus as the display
b) CPU is on a low latency LAN with the display
c) CPU is on a high latency WAN with the display
Case (a) is the most common followed by case (c). X was optimized for case (b). Because that was the most common 30 years ago. Case (b) should still be covered but the system should be designed around (a) with strong support for (c).
A respectful comment from the anti-Wayland side! Well done.
Anyway the reality is that Wayland guys wanted to get a prototype working to make sure there weren't any problems with remote... i.e. in theory a remote system could meet Wayland's "every frame is perfect". They don't actually care if in reality the frames are even visible for end users right now. That is the need it to work as essentially a unit test they don't need it to work right now for end user.
They do agree remote usage is an important feature and it is something they intend to add. Remember that applications, especially those that are often used remotely, are likely to be somewhat delayed in how long before they switch to Wayland only versions. Also VNC works much better with Wayland. So in their mind they have years to fix the remote problem. They do agree it is something on their todo list.
They implement a far better remote usage system They didn't implement network transparency. If you don't understand the distinction maybe you should assume that the developers who work on X11, which is the Wayland team, know a bit more than you do about how X11 works
LAN or WAN?
Actually no. You gain remote access. Wayland has something like Windows RDB. That hasn't been true in years.
As for X emulation... you can run an X11 server on Wayland same as you do on Windows/Aero or OSX/Aqua/Quartz. This one has never been true.
1) Developers being able to address new problems more easily because they have a more sane codebase
2) Better local video in a variety of ways. For example media coherence (i.e. video and sound won't get out of sync)
3) Better handling of latency so better remote desktop over Wans.
4) 32,768 pixel limit (not a problem yet but with 5k displays becoming more standard we are getting close to having a problem)
5) Validated Windows trees so windows won't get lost as often
We were talking about both. GP had said which induced response to what you are quoting, "In the 1980's Unix nearly DIED on the scattered hardware base. That was the biggest cause of all it's problems at the time." He was clearly talking abut big box Unixes in the Workstation era. Which is btw what I was talking about as well the "1980s workstation market"
I get that. But the GPL was part of a social movement that used copyright law to advance its aims. Most licenses are used to make someone some profit on IP.
Actually that almost never happens by design. One of the crucial distinctions between a corporation and a partnership is that a corporation is a legal person and thus doesn't incur liability for its investors. What you want would require no corporations and just partnerships, or very strict limits on partnerships... more like 18th century corporate law.
I'd agree. If that works then RMS's arguments about software suddenly apply to hardware.
Let me just point out the OSX machines are essentially variants on NeXT. NeXT released their first model spring of 1987 and was working hard on their stuff in 1985. Spring 1987 is probably around the time Vector would have released their 2nd version of the their cool stuff on the drawing board when customers would have had to transition.
Now if you are going to be transitioning in 1987... and the first fully object oriented graphic operating system which has concepts like the web (in primitive form) and email native is out how well does Vector's enhanced CP/M do? A very primitive version of OSX is where the bar would have been for them.
_____
You are right about the 1983 Commodore price war comment, that is what I meant. Commodore is a more complex situation since in the 1990s they were the low end and the grey box makers wanted to go down market. I suspect they could have gone 3rd world, and aimed for computers for $50. Something like the One Laptop Per Child initiative from a decade later . Or they could have been a player like Palm, Danger or BlackBerry were with early handhelds and heck that market now is much bigger than PCs.
I don't know what Commodore was thinking about during the last days. But given that SGI didn't make it...
Let me point out most of the Unixes went down in the flames in the mid 1990s, not the 1980s, when the workstation and server market consolidated (around NT and Linux). Unix thrived on diversity.
I don't think you are right here in a more general sense. Obviously something like the Linux kernel is a bad choice. But imagine the Linux toolset (i.e. the rest of the operating system) with at worse a slightly more hardware diverse kernel: NetBSD, QNX... Nothing much has to change. Again we know Linux can thrive in that environment, it does now and UNIX before it did, which I'll address below.
Sure. And YellowDog Linux "just worked" on PPC hardware. You pick a distribution for your hardware. People would be used to that since that's what they were doing anyway.
Died? Huh? That was the highpoint for commercial Linux. Commercial Linux in the 1980s was displacing other systems that had evolved in hardware monocultures because it handled the rapidly evolving workstation / server hardware. Your history is just backwards here.
Now certainly the success of Unix led to some problems that Open Systems needed to address, but that was a problem with success.
I agree with your first point about convergence. The Microsoft / Intel / Western Digital Standard for platform (IBM essentially) was a huge deal for hardware standardization.
But I'm not sure about your second point. Linux exists comfortably on dozens of hardwares. Linux is a diverse ecosystem not a monoculture. Why would Linux (or something like it) have not done better if there were 10 quirky brands about? Heck that's pretty much the 1980s workstation market and there is where Unix pretty much evolved from.
Mod up for that link.
I think Vector is a perfect example of disruptive technology. When they first choose CP/M it was the right choice. They couldn't switch too early from CP/M to a more advanced OS because that would be a downgrade. When more advanced OSes could run the same software they were caught hopelessly behind. Fundamentally Vector didn't own any technology that was unique to it. They had a mostly generic CP/M box with a few tweaks. One of the last CP/M manufacturers. I don't know the system but it wouldn't shock me if they made one of the best business CP/M boxes out there, that's what it takes to become a dominant player. But let's forget that Commodore launched a major price war in 1983 and was driving CP/M machines down to 0. I don't know if Vector could have survived Commodore even if there was no disruption.
Anyway If you look at competitors from the same time period like SGI, HP or Sun you can see there were possibilities up market. Or your example of Commodore, with Amiga or Apple with Mac. Others were able to transition. It clearly was possible to make a box that wasn't an IBM that was successful. They just didn't do much of anything to get there in time.
May 1982 Sun-1 Workstation
Jan 1983 Apple Lisa
Jan 1984 Apple Mac
Early 1985 SGI Iris 1000
July 1985 Amiga 1000
And their engineers had stuff in 1985 on the drawing board which was definitely better than CP/M and moving towards DOS? This wasn't just a failure of executive management. They deserved to die.
There are orders of magnitude more embedded devices than laptops and desktops. There are also more servers than desktop or laptops. So you can't make an appeal to quantity and then argue for desktops and laptops. Relative to those two categories the ones you are talking about are niche.
There are also far more Android tablets which appear to be replacing much of the consumer laptop / desktop space. There are more smartphones and even the Apple this year will outsell PCs Android having crossed years ago.
As for people in the developing world even excluding smartphones and tables given sales data and longer longevity about 20% are running Macs. In terms of profits Macs take about 85-92% for many years. So I'm not sure what the burn is. Windows is the dominant player in a large niche which is rapidly shrinking being successfully attacked from both above and below.
There were until recently niches were OS/2 thrived. You can always define things narrowly, but using most broad based metrics not constructed specifically to be a niche Windows is not dominant.
Apple was aggressively merging until November 2005 when their LLVM changes got blocked. June 2005 they had already announced the switch to Intel. I think we are disagreeing on facts here. Yes a lot of their changes weren't merged in the end but that was because the GCC developers didn't like the direction they were going. Early on they got merged. When really matters.
I'm not sure how second contradicts what I wrote.
As for Sony I was talking PS3 they used Apple's early work on GCC.
How does that disprove anything? The question is cooperation. Webkit clearly has lots of people working with it. That's what GPL does:
KHTML existed and was open
Apple used KHTML and developed Webkit.
Webkit existed and was open
Webkit attracted contribution
Under a BDS license for KHTML Webkit is never open source.
I disagree the GPL does nothing, I agree with your case.
Another common case than the one you cite is this. A, B, C and D are all building software for some secondary goal. For example Apple computer was working on GCC's optimizations for PPC because they wanted Mac software including OSX to run faster on PPC hardware, not because they wanted to get into the complier business. The GPL forced Apple to release all their changes to GCC even though they may have seen these as a competitive advantages and didn't release similar code bases when the code was BSD. This GCC release allowed Sony to leverage them on their game design system. Apple was forced to cooperate with Sony even though Apple made had no desire to. Another example involving Apple is the release of Webkit because of the KHTML code BTW.
That's certifying that the hardware's embedded software is free not that the hardware itself is free.
Device driver yes. RMS wasn't frustrated that the Xerox 9700 itself wasn't free.
I'd say copyright assignment. Relicensing is dangerous because it makes standing complex. With relicensing you are just another license holder equal to the license violator. How do you prove damages or have the right to settle violations if you don't have copyright?
Cisco systems paid damages (a donation) to the FSF for Linksys violations.
NuSphere lost a bunch to MySQL.
I could probably come up with more examples.
You are misunderstanding the quote. The quote from the original article was about how it is getting to hard to create a platform from scratch and the with rise of free platforms most people creating proprietary software are starting with a free (BSD usually) licensed system and extending it as a commercial product.
I agree with your point entirely ethically.
However... I'm just going to nitpick. The hardware definition of open platform is a combination of interoperability, portability, and open software standards that allow you to have supplier choice. It has nothing to do with you personally being one of those vendors. The relative ease of being able to join the Android ecosystem (i.e. it becomes practical at a few hundred thousand units) makes it an open platform regardless of how locked down each of those units is individually. You may not like the definition but that is the definition.
I'd also like to mention three area you forgot which is incredibly open.
1) Clouds especially IaaS. Those are running lightweight OSes designed to freely run all sorts of client OSes. That's a huge area of freedom replacing the big box proprietary systems. .NET (if Microsoft wanted it which they are playing with) and Xiaomi's business model is based on this.
2) Moreover this server based technology scales down nicely to both server and desktop. So that while you may not be able to freely run different base OSes you can freely run different client OSes and those are free distributed on free VM managers.
3) Finally on phones Android itself is constructed to allow for different environments, Google Play is replaceable or combinable. You can have Android
That's hard to enforce on people redistributing.
A writes a GPL application. B uses it and writes something on top. B gives a modified binary to C with no notice to A. C distributes it wildly including to D. How do D or A force B to release his changes? Remember C doesn't have the source changes.
Hardware drivers are not a problem of software. Those are given away for free mostly. The problem is the open source aspect not the monetization aspect. Your 2 year windows wouldn't solve the driver the problem. As for games ... I'm not sure what incentive game writers have for ever doing GPL. They are in the one off commercial software business.