- Erlang didn't get less than a second of downtime in a year, an application written in Erlang got less than a second of downtime in a year. I bet people clever enough to write such an application in Erlang could have written it in another language. Would it have been more difficult? Probably. But just because you use Erlang doesn't mean that your application is going to magically never going to have downtime - you're still going to have to work hard at it.
- Erlang is not necessarily the right choice for "high-end multi-core multi-system clustered application development". Erlang is not fast at math, and if you have a clustered application that computes fluid dynamics or cracks RC5, you'll probably keep writing it in C or C++ or Fortran because they do math fast. Don't believe me? The n-body benchmark over at the Computer Language Benchmarks Game is all double-float arithmatic, and the Erlang version takes almost six times as long as the version in Common Lisp and almost eight times as long as the version in Fortran.
I'm not explicitly granting permission for people to find me; I'm letting a select group of people try to get in touch with me. I don't give my number out to just anyone, and even if I do give it to you, I'm not always going to choose to pick up the phone when you call. I do have my phone configured to give location information to emergency services, but not to anyone else.
"The package system. A huddle of shell scripts without a strict API. Can be forgiven since nothing essential depends on it. A big plus sign for Linux."
pkgsrc works, but it's nowhere near as nice as apt. There's a Debian port to use the FreeBSD kernel (http://wiki.debian.org/Debian_GNU/kFreeBSD) that looks promising - the more operating systems that get something like apt, the better.
FreeBSD has ZFS. My understanding is while ZFS is a good filesystem, it isn't without issues. It doesn't work well on 32-bit architectures because of the memory requirements, isn't reliable enough to host a swap partition, and can't be used as a boot partition when part of a pool. Here's FreeBSD's rundown of known problems: http://wiki.freebsd.org/ZFSKnownProblems.
On the other hand, the new filesystems in the Linux kernel - ext4 and btrfs - are taking the lessons learned from ZFS. I'm excited about next-generation filesystems, and I don't think ZFS is the only way to go.
"I wonder, too... does Mr. Stallman's PC have a proprietary BIOS, or did he write that code, too?"
Yes, he uses a free BIOS.
Stallman is pragmatic, it's just that his line between pragmatic and unrealistic isn't quite where you (or I) would draw it. Before the Linux kernel came along, the GNU project had existed almost a decade - he surely used a computer then in order to write (or help write) the first versions of GCC, emacs, and a host of other essential free software programs. Now he doesn't have to, so he won't go back, and I'm sure he jumped onto a free kernel as quickly as it became remotely practical to use. Remember, GNU was working on a kernel of its own before Linux came on the scene. Again, this is not the pragmatism you or I might have chosen, but it is pragmatism. Same thing with FreeBIOS - he was in the news a year or two ago about switching the FSF over to computers that could run a completely free software BIOS. He's also been one of the many campaigning for free firmware, and the FSF has made that an important issue recently. CPU microcode? Haven't heard him talk about that yet, but I'm sure it will come.
Stallman has been working 30 years on getting a completely free system, and he's lead the way for a lot of us that use a kinda-sorta free system and are better off for it. There's a discussion to be had here about the intersection of Javascript, free software, software as a service, and cloud computing, but that sort of snarky comment doesn't help. He's been right on so many other issues - copyright, patents, free software - when he says something is a problem, we ought to at least listen.
Debian's bug tracker deals with this well. It understands many different types of bug trackers - IIRC, among them launchpad, bugzilla, trac. If a bug is opened with Debian, it can be forwarded upstream, and when it's resolved, Debian's bug tracker will mark it as 'resolved upstream', and it can be closed when a package with the fix is uploaded.
"Before RMS spoke about it most of you were for Cloud Computing now you are against it. You're a bunch of sheep."
Exactly what I thought! What sheep, changing their mind after being convinced by a well-reasoned argument from someone with a track record for making good predictions!
"Mozilla's beef is with Debian or anybody else messing around with code or the settings and still trying to palm it off as Mozilla Firefox. People are still free to branch the code and call it anything they like, which is just what Debian has done."
What if the Linux kernel developers did the same thing - the only way you can call it Linux is to distribute an official kernel.org release with no custom patches? KDE? GNOME? vim? emacs? every single application you use? It would really suck to install software, because there would be no name you could use to identify what your distribution was shipping without infringing on someone else's trademark.
Sure, trademark law allows Mozilla to do this, but it's been custom in the free software community *not* to do this for over 25 years.
The GPL even includes a specific provision for Mozilla's worry of "someone will patch it to break it and our reputation will be tarnished" - if you distribute a modified copy, "the work must carry prominent notices stating that you modified it, and giving a relevant date".
Xubuntu's performance targeting appears limited to choice of desktop environment, which was a small component of what these benchmarks tested. The big performance increases the article talks about were in databases, compilers, encryption, memory access, and audio/video encoding/decoding, none of which really have much to do with the desktop environment.
"The obvious conclusion, supported by lots of data for those inclined to look, is that big mergers always increase efficiency and hence reduce prices for the consumer."
Counterpoint: cell phone carriers and text messages.
From the groklaw link: "Instead of a EULA, the new page you get on install is a notices page with no "I agree" requirement, along with a link to an optional services agreement, and instructions there on how to avoid having to accept the services, if you don't want them."
Let me get this straight. There's a popup window with legalese that includes an agreement that you have to figure out how to opt out of? So it's like a EULA, but they just assume you agree, and the "I Agree" button has been renamed "Next"?
"The Firefox EULA outlines some quite important issues, not least of which is that it doesn't ship with a warranty."
Why is Firefox so special or important that it makes me confirm a EULA? And why, after these several decades since the Free Software Movement started, has no other major piece of free software done something similar? It's not like the Free Software Foundation is still working out the basics of licensing or anything.
I have 1,804 packages installed on my Debian system. I don't know _any_ of those packages that don't disclaim warranty to the maximum extent provided by law. It's in/usr/share/doc/packagename/copyright, for me to read as I please. Since it's Debian, and I get software from main, I know that anything I get from there places no restrictions on my use of the software, and that I only need to check it if I intend on modifying or distributing the software.
I'm glad Debian did away with Firefox and provides a free, rebranded version so I don't have to put up with that crap.
"Linux is not a GNU OS -- much less "The GNU OS". It is an OS that uses GNU utilities."
GNU/Linux is a GNU OS that uses the Linux kernel and GNU userland (including "utilities" like glibc). There are other GNU OSs that use other kernels with the GNU userland; for example, GNU/kFreeBSD and, yes, GNU/Hurd. There are not other OSs that use the Linux kernel and non-GNU userland.
The best analogy I have is trucks and engines. Say you have a truck made/assembled/whatever by Dodge that uses a Hemi engine as one of its components. The GNU/Linux camp would describe the truck as a Dodge truck with a Hemi engine. The Linux camp would describe the truck as a Hemi engine with some Dodge parts.
It's all a point of view thing, but from where I sit (interacting bash, gcc, ls, glibc, and all the other fine pieces of the GNU userland), it looks more like GNU than Linux.
"Have any free formats ever been taken to court and won, proving their status as truly free? Or are they 'under the radar' at the moment, not worth testing in court because they've not reached critical mass yet?"
With the number of software patents around, do you think that one court case could declare something as "truly free" and not infringing on any patent? Ten cases? One hundreded? What's with the requirement that someone get sued and win to declare something patent-free or usable, anyway?
Xiph has been dilligent about searching for patents, and that's pretty good. Additionally, Vorbis has seen significant adoption in some industries - the game industry comes to mind - so it's not just some random format that's been around for a while but mostly unused.
Hopefully they're doing SSL, so even if the DNS is spoofed, all transmissions need to go through a connection with an Apple SSL certificate, so it won't be an issue.
"I can download the things as many times as I want."
You may download the games as many times as you want until they are no longer available for download. What happens if they go away, or decide not to provide the service anymore? Sure, Steam is big, but so is Yahoo, and they just closed one of their online stores that promised something similar.
"Which we could then encase in leak proof containers and dump them in a subduction zone."
Two problems:
- Leak proof? Hah.
- Subduction zone? The material that's subducted feeds volcanoes. You don't have to guard against earthquakes, just volcanoes that spew radioactive ash high into the atmosphere where it travels far and wide.
Slashdot Mirror
Airfare search is hard. Really hard. The guy most responsible for ITA's (now Google's) flight search engine wrote up a presentation:
http://www.demarcken.org/carl/papers/ITA-software-travel-complexity/img0.html
See in particular "Some complexity results": http://www.demarcken.org/carl/papers/ITA-software-travel-complexity/img24.html
The article specifically addresses random access across high latency. Search for "latency", or just read the article.
When I looked at the release notes sent out by email, I saw this under "New functionality":
"httpd(8) can now serve files larger than 2GB in size."
I'm very surprised by this.
- Erlang didn't get less than a second of downtime in a year, an application written in Erlang got less than a second of downtime in a year. I bet people clever enough to write such an application in Erlang could have written it in another language. Would it have been more difficult? Probably. But just because you use Erlang doesn't mean that your application is going to magically never going to have downtime - you're still going to have to work hard at it.
- Erlang is not necessarily the right choice for "high-end multi-core multi-system clustered application development". Erlang is not fast at math, and if you have a clustered application that computes fluid dynamics or cracks RC5, you'll probably keep writing it in C or C++ or Fortran because they do math fast. Don't believe me? The n-body benchmark over at the Computer Language Benchmarks Game is all double-float arithmatic, and the Erlang version takes almost six times as long as the version in Common Lisp and almost eight times as long as the version in Fortran.
I'm not explicitly granting permission for people to find me; I'm letting a select group of people try to get in touch with me. I don't give my number out to just anyone, and even if I do give it to you, I'm not always going to choose to pick up the phone when you call. I do have my phone configured to give location information to emergency services, but not to anyone else.
"The package system. A huddle of shell scripts without a strict API. Can be forgiven since nothing essential depends on it. A big plus sign for Linux."
pkgsrc works, but it's nowhere near as nice as apt. There's a Debian port to use the FreeBSD kernel (http://wiki.debian.org/Debian_GNU/kFreeBSD) that looks promising - the more operating systems that get something like apt, the better.
FreeBSD has ZFS. My understanding is while ZFS is a good filesystem, it isn't without issues. It doesn't work well on 32-bit architectures because of the memory requirements, isn't reliable enough to host a swap partition, and can't be used as a boot partition when part of a pool. Here's FreeBSD's rundown of known problems: http://wiki.freebsd.org/ZFSKnownProblems.
On the other hand, the new filesystems in the Linux kernel - ext4 and btrfs - are taking the lessons learned from ZFS. I'm excited about next-generation filesystems, and I don't think ZFS is the only way to go.
"I wonder, too ... does Mr. Stallman's PC have a proprietary BIOS, or did he write that code, too?"
Yes, he uses a free BIOS.
Stallman is pragmatic, it's just that his line between pragmatic and unrealistic isn't quite where you (or I) would draw it. Before the Linux kernel came along, the GNU project had existed almost a decade - he surely used a computer then in order to write (or help write) the first versions of GCC, emacs, and a host of other essential free software programs. Now he doesn't have to, so he won't go back, and I'm sure he jumped onto a free kernel as quickly as it became remotely practical to use. Remember, GNU was working on a kernel of its own before Linux came on the scene. Again, this is not the pragmatism you or I might have chosen, but it is pragmatism. Same thing with FreeBIOS - he was in the news a year or two ago about switching the FSF over to computers that could run a completely free software BIOS. He's also been one of the many campaigning for free firmware, and the FSF has made that an important issue recently. CPU microcode? Haven't heard him talk about that yet, but I'm sure it will come.
Stallman has been working 30 years on getting a completely free system, and he's lead the way for a lot of us that use a kinda-sorta free system and are better off for it. There's a discussion to be had here about the intersection of Javascript, free software, software as a service, and cloud computing, but that sort of snarky comment doesn't help. He's been right on so many other issues - copyright, patents, free software - when he says something is a problem, we ought to at least listen.
Debian's bug tracker deals with this well. It understands many different types of bug trackers - IIRC, among them launchpad, bugzilla, trac. If a bug is opened with Debian, it can be forwarded upstream, and when it's resolved, Debian's bug tracker will mark it as 'resolved upstream', and it can be closed when a package with the fix is uploaded.
"Before RMS spoke about it most of you were for Cloud Computing now you are against it. You're a bunch of sheep."
Exactly what I thought! What sheep, changing their mind after being convinced by a well-reasoned argument from someone with a track record for making good predictions!
"Mozilla's beef is with Debian or anybody else messing around with code or the settings and still trying to palm it off as Mozilla Firefox. People are still free to branch the code and call it anything they like, which is just what Debian has done."
What if the Linux kernel developers did the same thing - the only way you can call it Linux is to distribute an official kernel.org release with no custom patches? KDE? GNOME? vim? emacs? every single application you use? It would really suck to install software, because there would be no name you could use to identify what your distribution was shipping without infringing on someone else's trademark.
Sure, trademark law allows Mozilla to do this, but it's been custom in the free software community *not* to do this for over 25 years.
The GPL even includes a specific provision for Mozilla's worry of "someone will patch it to break it and our reputation will be tarnished" - if you distribute a modified copy, "the work must carry prominent notices stating that you modified it, and giving a relevant date".
Xubuntu's performance targeting appears limited to choice of desktop environment, which was a small component of what these benchmarks tested. The big performance increases the article talks about were in databases, compilers, encryption, memory access, and audio/video encoding/decoding, none of which really have much to do with the desktop environment.
"The obvious conclusion, supported by lots of data for those inclined to look, is that big mergers always increase efficiency and hence reduce prices for the consumer."
Counterpoint: cell phone carriers and text messages.
Great article. I'd like to see more of this on Slashdot.
From the groklaw link: "Instead of a EULA, the new page you get on install is a notices page with no "I agree" requirement, along with a link to an optional services agreement, and instructions there on how to avoid having to accept the services, if you don't want them."
Let me get this straight. There's a popup window with legalese that includes an agreement that you have to figure out how to opt out of? So it's like a EULA, but they just assume you agree, and the "I Agree" button has been renamed "Next"?
I don't see how this is significantly different.
Thanks, FSF! I appreciate the work you do to promote free software, and this is another great example.
"The Firefox EULA outlines some quite important issues, not least of which is that it doesn't ship with a warranty."
Why is Firefox so special or important that it makes me confirm a EULA? And why, after these several decades since the Free Software Movement started, has no other major piece of free software done something similar? It's not like the Free Software Foundation is still working out the basics of licensing or anything.
I have 1,804 packages installed on my Debian system. I don't know _any_ of those packages that don't disclaim warranty to the maximum extent provided by law. It's in /usr/share/doc/packagename/copyright, for me to read as I please. Since it's Debian, and I get software from main, I know that anything I get from there places no restrictions on my use of the software, and that I only need to check it if I intend on modifying or distributing the software.
I'm glad Debian did away with Firefox and provides a free, rebranded version so I don't have to put up with that crap.
"Linux is not a GNU OS -- much less "The GNU OS". It is an OS that uses GNU utilities."
GNU/Linux is a GNU OS that uses the Linux kernel and GNU userland (including "utilities" like glibc). There are other GNU OSs that use other kernels with the GNU userland; for example, GNU/kFreeBSD and, yes, GNU/Hurd. There are not other OSs that use the Linux kernel and non-GNU userland.
The best analogy I have is trucks and engines. Say you have a truck made/assembled/whatever by Dodge that uses a Hemi engine as one of its components. The GNU/Linux camp would describe the truck as a Dodge truck with a Hemi engine. The Linux camp would describe the truck as a Hemi engine with some Dodge parts.
It's all a point of view thing, but from where I sit (interacting bash, gcc, ls, glibc, and all the other fine pieces of the GNU userland), it looks more like GNU than Linux.
M-x doctor
I am the psychotherapist. Please, describe your problems. Each time you are finished talking, type RET twice.
I'm a stressed astronaut.
Why do you say you are a stressed astronaut?
Because I'm in space and I'm feeling stressed.
Is it because you are in space and you are feeling stressed that you came to me?
Yes.
I see... Well, what makes you believe this is so?
*opens airlock*
"Have any free formats ever been taken to court and won, proving their status as truly free? Or are they 'under the radar' at the moment, not worth testing in court because they've not reached critical mass yet?"
With the number of software patents around, do you think that one court case could declare something as "truly free" and not infringing on any patent? Ten cases? One hundreded? What's with the requirement that someone get sued and win to declare something patent-free or usable, anyway?
Xiph has been dilligent about searching for patents, and that's pretty good. Additionally, Vorbis has seen significant adoption in some industries - the game industry comes to mind - so it's not just some random format that's been around for a while but mostly unused.
Hopefully they're doing SSL, so even if the DNS is spoofed, all transmissions need to go through a connection with an Apple SSL certificate, so it won't be an issue.
"I can download the things as many times as I want."
You may download the games as many times as you want until they are no longer available for download. What happens if they go away, or decide not to provide the service anymore? Sure, Steam is big, but so is Yahoo, and they just closed one of their online stores that promised something similar.
Why do people still use BIND? It has a track record of security vulnerabilities almost as long as Sendmail's.
"Which we could then encase in leak proof containers and dump them in a subduction zone."
Two problems:
- Leak proof? Hah.
- Subduction zone? The material that's subducted feeds volcanoes. You don't have to guard against earthquakes, just volcanoes that spew radioactive ash high into the atmosphere where it travels far and wide.
Maybe SCO can just give them 3,576 licenses for the Linux kernel.