Re:I think we'll see more specialized OSs
on
A New Kind of OS
·
· Score: 3, Informative
I agree we're going to see a lot more customized forks in the world of GNU/Linux, but I disagree that most of them are going to be full operating systems. Instead we're seeing a common core with customized faceplates on it. For example what the Ubuntu folks are doing with Ubuntu/Kubuntu/Xubuntu/Edubuntu. Behind the scenes it is all one OS, but with different faceplates changing how it appears to the user. Debian are doing the same thing.
Linux the kernel and GNU/Linux the operating system both have a bunch of things that OpenSolaris doesn't:
1) Vendor Neutrality. OpenSolaris is closely associated with Sun, but no company has a stranglehold over Linux.
2) Portability. Linux has been ported to an amazing array of hardware. Ubuntu runs on more architectures than OpenSolaris, even though they dropped most of the archs supported by Debian.
3) Scalability. Linux scales up to supercomputers and mainframes (where Solaris also has a respectable track record), and just as important, Linux scales down to wristwatches, cell phones, videorecorders, and all sorts of embedded processors.
4) Momentum and mindshare. There is a huge community committed to and invested in Linux. That hasn't yet happened with OpenSolaris.
Why is proprietary okay for controlling an MRI scanner? The software controlling the scan sequence and image reconstruction are at the heart of the system; they are arguably even more important than the coil and magnet designs. MRI is so flexible precisely because only by changing software one can change which physical properties of the subject are imaged. There is a lot of research still going on into new MRI sequences and reconstruction algorithms.
Getting low-level access to our imaging systems is an ongoing problem for many of us doing research in radiology. There are solutions--the device manufacturers do realize we impact their bottom line--but I would much rather have full source unencumbered by NDA than the status quo.
He [RMS] happens to have an incredible drive to accomplish what any idiot could have seen was impossible when he started.
Absolutely true. But the thing is, Stallman has succeeded in accomplishing some of those "impossible" things. That's why he is a genius but you and I are just idiots.
It is not true that people only use rings 0 and 3 on the x86 architecture. For example, the Xen virtual machine monitor, runs in ring 0, has the the guest OS kernels--e.g. Linux--in ring 1, and leaves ring 3 for userland.
What can you do with Linux that you couldn't do with SVr4 in 1992? Freely share the source with all your friends and customers without fear of lawsuit and include pre-installed binaries on hardware without paying royalties.
GPL licensing is the single most important feature distinguishing Linux from proprietary kernels such as UNIX and other free kernels like the BSDs. The GPL's copyleft provision and the dual-licensing opportunity it creates are why companies like IBM and SGI have contributed subsystems like JFS and XFS to Linux. They wouldn't have shared the same code under a BSD license. Linus has said that choosing the GPL is the best decision he made in the early days.
I want to be able to search the browser cache, since that's where pages I've recently visited can be found. Sure, I can grep the directory, but this really should be integrated into the browser.
As you say, the majority of Solaris users don't care. Those of us who do care about good support for cheap hardware have already left Sun and are using GNU/Linux. We're the reason that Sun has been hemorrhaging market share.
Sun like to talk about how Solaris scales up to big iron. But scalability goes both ways. Linux has been so successful because it doesn't just scale up to the high end, it also scales down to the low end. At work we use Linux across our cheap old desktops, our beefy servers, and our 100 CPU compute cluster. Having the same OS everywhere is a major savings in admin and developer time. Sun can't compete, because they refuse to support the cheap hardware.
Sun would gain access to Linux's device drivers if they chose to use the GPL. MS Windows and Linux are the only kernels with extensive hardware support; if you can't piggyback on at least one of those two sets of drivers, there is a very large class of hardware you don't run on. Sun don't have the manpower to rewrite all those drivers.
Given that Slowaris x86's biggest weakness is hardware support, yeah there is something bad about Sun not choosing the GPL. But it is bad for Sun and their users, not those of us already in the GNU/Linux camp.
From Rosen's viewpoint, remember he's a lawyer, having free software developers be dragged into court is a win, for they are likely to hire him as counsel.
Which better gauge do you propose? Imperfect data is much better than no data and n=5000 is a plenty big sample size from which to make statistically significant inferences.
If you'd like your machines counted apt-get install popularity-contest
The Debian popularity contest project is an attempt to measure which configurations of hardware and packages are most common among Debian users. One use for the data is figuring out which packages belong on the install CD since everybody uses them. Less popular packages go on the supplemental CDs.
If one ignores the large chunk (over a quarter) of machines flagged unknown architecture, about 4% of popcon users are on non-i386 machines.
Re:The Cassandra effect and Public discourse (long
on
A New Ice Age?
·
· Score: 3, Interesting
I agree with you that the public discourse about science (and medicine) is dominated by pepole who are appalingly naive and misinformed. Alas, mathematics educators are busy fighting about calculus reform instead of figuring out how to teach probability/statistics in high school.
I have a question about the nature of climate. You say:
Climate, defined as the long-term average behavior of the atmsophere, ocean and ice, shows a lot of natural variability in response to perturbations in forcing.
What about 1/f noise? Are we studying a stationary random process? Is it even legitimate from a mathematical-modeling viewpoint to talk about long-term average behavior? The prediction you make about 2304 is reasonable, but hardly long-term by geological standards.
Most info files are written using texinfo which outputs various formats including Info, PDF and HTML. The hardcopy manuals sold by
GNU Press
come from the same source file as the softcopy documentation on your computer.
(Texinfo is several years older than either PDF or HTML. If web browsers had already existed the GNU project probably wouldn't have invented info browsers.)
The kernel developers have clearly said what the boundaries are. On one side the bus interface and the other the system call interface. Everything in between is the kernel. There deliberately is no boundary between kernel and drivers, and Linus has repeatedly said he will not reconsider that decision since the technical reasons for it are still valid.
You're confused about the GPL. The GPL places
no limitations on how you may use software.
In particular, the GPL places no restrictions on "commercial
use" (whatever that means). The GPL only affects you if you
redistribute GPLed software. For futher
details see the license itself, or the GPL FAQ.
By the way, the crime of "piracy" involves hijacking a
ship. (Hijacking a plane is legally "air piracy".)
Illegally copying software is properly refered to as
"copyright infringement".
In other words, was there any redeeming factors
Debian had over any other distribution?
I can't speak for Klaus, but I do know why if I were
trying to do something similar I'd start with Debian:
Source code is available for everything in
Debian, including the utilities used to produce the official
distribution CDs.
Debian is a multi-lingual distribution, with an active
translation team. Knoppix defaults to German.
It is legal to redistribute everything in both Debian
proper and in the non-free supplement.
With 11 (and counting) supported architectures Debian
is the most portable GNU/Linux distribution. If someone
wants a non-x86 CD much difficult trailblazing is already
there.
Anybody with technical chops can be a Debian developer.
(Klaus maintains the cloop-src and cloop-utils
packages.)
Sid (Debian "Unstable") is updated daily, making it
easy to get the latest packages.
Sure, other distros have most of these things. But
Debian is the only one with all of them.
It looks to me like the community should be happy with
anything that contributes to the cause, which I had always
assumed was presenting a viable alternative to M$.
You have assumed incorrectly. Most of us in
the free
software movement are interested in having an OS system
we are free to use, study, share, and improve. Like M$,
Lindows tramples on those freedoms. Lindows is not
contributing to the cause, they're detracting from it.
57 percent of respondents never or seldom pay for copyrighted works they download.
Almost everything on the web is copyrighted. When you click on a link your browser downloads it in order to display it to you. 100% of web surfers never or seldom pay for the copyrighted web pages they read.
(There are a few specialty markets, e.g. academic journals, where copyrighted web content is available by subscription only. But most of the web is gratis to all.)
You'll make a lot of new friends. People you didn't meet in high
school and aren't likely to meet later in life. Some will be people
who are similar to you and people who are very different from you.
(Especially if you live in a dorm. And if you don't live in a dorm
you're missing out on one of the most important parts of college.)
You'll learn a lot. Many of the important lessons won't be in
the classroom. They'll come from all the new friends you make and
from the folks your dorm who have nothing in common with you.
You have time to study something (your major) in depth.
You get to learn "fun stuff". Learn a language. Learn about art
or how to play an instrument. Learn some neat science. Read some
poetry. Join the radio station. Theater. Whatever your fancy is
you'll be able to do some of it.
If you're lucky you'll learn to think.
If you're very lucky you'll learn how to write. You should
certainly take some classes that try to teach writing. They're not
necessarily in the English department; I learned a lot about writing
in history classes.
What to major in: whatever you want. Don't waste your time in an MIS
department. Major in one of the liberal arts, since those are more
likely to teach you how to think and how to write. (And for the
record, mathematics and physics are both liberal arts, as are music,
history, English, etc.)
Steve Case, chairman of AOL is an investing genius. AOL stock was wildly overvalued at the time of the merger; the price would have tanked with our without the merger. Case used internet bubble speculation to buy Time Warner, converting soon-to-be-worthless paper into a valuable asset. Without the Time-Warner side of the company AOL would be worth a lot less today. (The Time-Warner side might be better off today without the merger, but it was definitely good for AOL shareholders.)
Slashdot Mirror
I agree we're going to see a lot more customized forks in the world of GNU/Linux, but I disagree that most of them are going to be full operating systems. Instead we're seeing a common core with customized faceplates on it. For example what the Ubuntu folks are doing with Ubuntu/Kubuntu/Xubuntu/Edubuntu. Behind the scenes it is all one OS, but with different faceplates changing how it appears to the user. Debian are doing the same thing.
You're probably running Gentoo on an x86 and only worried about the Linux kernel, so the limitations of 3.4 don't affect you. Debian switched to 4.1 in part because it fixes problems with the hppa and m68k ports (and the nascent Debian ports to HURD and the FreeBSD kernel), and in part because it has new warnings pointing out code that isn't 64-bit clean.
Linux the kernel and GNU/Linux the operating system both have a bunch of things that OpenSolaris doesn't:
1) Vendor Neutrality. OpenSolaris is closely associated with Sun, but no company has a stranglehold over Linux.
2) Portability. Linux has been ported to an amazing array of hardware. Ubuntu runs on more architectures than OpenSolaris, even though they dropped most of the archs supported by Debian.
3) Scalability. Linux scales up to supercomputers and mainframes (where Solaris also has a respectable track record), and just as important, Linux scales down to wristwatches, cell phones, videorecorders, and all sorts of embedded processors.
4) Momentum and mindshare. There is a huge community committed to and invested in Linux. That hasn't yet happened with OpenSolaris.
Ricoh copiers run NetBSD. Jim Barton has written about licensing issues in the Tivo; they picked Linux over BSD because it was technically superior.
Why is proprietary okay for controlling an MRI scanner? The software controlling the scan sequence and image reconstruction are at the heart of the system; they are arguably even more important than the coil and magnet designs. MRI is so flexible precisely because only by changing software one can change which physical properties of the subject are imaged. There is a lot of research still going on into new MRI sequences and reconstruction algorithms. Getting low-level access to our imaging systems is an ongoing problem for many of us doing research in radiology. There are solutions--the device manufacturers do realize we impact their bottom line--but I would much rather have full source unencumbered by NDA than the status quo.
What's the license like, is Indy open source?
Thanks,
--Andre
He [RMS] happens to have an incredible drive to accomplish what any idiot could have seen was impossible when he started. Absolutely true. But the thing is, Stallman has succeeded in accomplishing some of those "impossible" things. That's why he is a genius but you and I are just idiots.
It is not true that people only use rings 0 and 3 on the x86 architecture. For example, the Xen virtual machine monitor, runs in ring 0, has the the guest OS kernels--e.g. Linux--in ring 1, and leaves ring 3 for userland.
What can you do with Linux that you couldn't do with SVr4 in 1992? Freely share the source with all your friends and customers without fear of lawsuit and include pre-installed binaries on hardware without paying royalties. GPL licensing is the single most important feature distinguishing Linux from proprietary kernels such as UNIX and other free kernels like the BSDs. The GPL's copyleft provision and the dual-licensing opportunity it creates are why companies like IBM and SGI have contributed subsystems like JFS and XFS to Linux. They wouldn't have shared the same code under a BSD license. Linus has said that choosing the GPL is the best decision he made in the early days.
I hope they do a better job with CVS then when they botched implementing NTP
I want to be able to search the browser cache, since that's where pages I've recently visited can be found. Sure, I can grep the directory, but this really should be integrated into the browser.
As you say, the majority of Solaris users don't care. Those of us who do care about good support for cheap hardware have already left Sun and are using GNU/Linux. We're the reason that Sun has been hemorrhaging market share.
Sun like to talk about how Solaris scales up to big iron. But scalability goes both ways. Linux has been so successful because it doesn't just scale up to the high end, it also scales down to the low end. At work we use Linux across our cheap old desktops, our beefy servers, and our 100 CPU compute cluster. Having the same OS everywhere is a major savings in admin and developer time. Sun can't compete, because they refuse to support the cheap hardware.
Sun would gain access to Linux's device drivers if they chose to use the GPL. MS Windows and Linux are the only kernels with extensive hardware support; if you can't piggyback on at least one of those two sets of drivers, there is a very large class of hardware you don't run on. Sun don't have the manpower to rewrite all those drivers.
Given that Slowaris x86's biggest weakness is hardware support, yeah there is something bad about Sun not choosing the GPL. But it is bad for Sun and their users, not those of us already in the GNU/Linux camp.
From Rosen's viewpoint, remember he's a lawyer, having free software developers be dragged into court is a win, for they are likely to hire him as counsel.
Which better gauge do you propose? Imperfect data is much better than no data and n=5000 is a plenty big sample size from which to make statistically significant inferences. If you'd like your machines counted apt-get install popularity-contest
The Debian popularity contest project is an attempt to measure which configurations of hardware and packages are most common among Debian users. One use for the data is figuring out which packages belong on the install CD since everybody uses them. Less popular packages go on the supplemental CDs. If one ignores the large chunk (over a quarter) of machines flagged unknown architecture, about 4% of popcon users are on non-i386 machines.
I have a question about the nature of climate. You say:
What about 1/f noise? Are we studying a stationary random process? Is it even legitimate from a mathematical-modeling viewpoint to talk about long-term average behavior? The prediction you make about 2304 is reasonable, but hardly long-term by geological standards.
(Texinfo is several years older than either PDF or HTML. If web browsers had already existed the GNU project probably wouldn't have invented info browsers.)
The kernel developers have clearly said what the boundaries are. On one side the bus interface and the other the system call interface. Everything in between is the kernel. There deliberately is no boundary between kernel and drivers, and Linus has repeatedly said he will not reconsider that decision since the technical reasons for it are still valid.
You're confused about the GPL. The GPL places no limitations on how you may use software. In particular, the GPL places no restrictions on "commercial use" (whatever that means). The GPL only affects you if you redistribute GPLed software. For futher details see the license itself, or the GPL FAQ.
By the way, the crime of "piracy" involves hijacking a ship. (Hijacking a plane is legally "air piracy".) Illegally copying software is properly refered to as "copyright infringement".
I can't speak for Klaus, but I do know why if I were trying to do something similar I'd start with Debian:
Sure, other distros have most of these things. But Debian is the only one with all of them.
In post 4387579 XP_fetchbeer wrote:
You have assumed incorrectly. Most of us in the free software movement are interested in having an OS system we are free to use, study, share, and improve. Like M$, Lindows tramples on those freedoms. Lindows is not contributing to the cause, they're detracting from it.
Almost everything on the web is copyrighted. When you click on a link your browser downloads it in order to display it to you. 100% of web surfers never or seldom pay for the copyrighted web pages they read.
(There are a few specialty markets, e.g. academic journals, where copyrighted web content is available by subscription only. But most of the web is gratis to all.)
What to major in: whatever you want. Don't waste your time in an MIS department. Major in one of the liberal arts, since those are more likely to teach you how to think and how to write. (And for the record, mathematics and physics are both liberal arts, as are music, history, English, etc.)
Steve Case, chairman of AOL is an investing genius. AOL stock was wildly overvalued at the time of the merger; the price would have tanked with our without the merger. Case used internet bubble speculation to buy Time Warner, converting soon-to-be-worthless paper into a valuable asset. Without the Time-Warner side of the company AOL would be worth a lot less today. (The Time-Warner side might be better off today without the merger, but it was definitely good for AOL shareholders.)