What is most needed by physics students? Here are my thoughts.
First, equation solving. This means numerical integration, root finding, and so on.
Second, spectral methods.
Third, numerical error. What does it mean for a numerical calculation to be ill-conditioned?
If they start out without the skills needed to write a trivial program, the physics department should not have to teach them. Don't people learn that much in high school?
The most important thing is the hardest to teach in a course of limited scope: writing readable, maintainable, debuggable code. Most scientists suck at this.
What language? Something that graphs! Something easy to debug (they don't need to spend their lives chasing one-off indexing errors). Good languages with OSS semi-equivalents are Matlab (Octave) and IDL (GDL). Is a spreadsheet appropriate? I don't know enough to say. If someone learns on a spreadsheet, will they be able to more-or-less understand a piece of code in F90? I also object to forcing people to purchase from Microsoft.
"Red Hat has never, ever waivered in its support of Free, OSS software, and eg. released some proprietary closed source software as part of their distros"
Er, Peter, would you like to see my copy of RHL 4.something with a proprietary X server on it? I believe it was something like MetroX. Netscape 4 was also available as part of RHL for a while (looks like at least 6 and 7).
Not saying this to attack RH; their distros are the only ones I ever use.
I couldn't get freezing to work on my dead drive. The trick that worked was: Let sit on the desk for two months and then try it again. It still made noise, but it worked long enough to find and retrieve the files that weren't backed up.
I really like selinux. The best part about it is this: Whenever something is broken, I uninstall selinux, and then whatever-it-is works again. I wouldn't know how to fix the system if I couldn't uninstall selinux.
(I am not denying that it is important or useful. I just can't understand how to make it work.)
This has been mentioned, but I want to explore it a little more deeply.
The assumption is that any position must be liberal or conservative, but this leads to nonsense. There is a large centrist middle. So what we see is that the largely centrist press looks liberal to extreme conservatives and conservative to extreme liberals. For example, being against abolishing social security or against abolishing the death penalty are not liberal or conservative, they are centrist. Does the media have a centrist bias?
Furthermore, support or criticism of a particular politician is not evidence of bias. It is poor reporting. Name your most or least favorite politician - probably they are in office because of a period of uncritical support by the media and their biggest problem is unfair, inaccurate, negative reporting in the media.
The rendering engines have patented components, so you can't make things look precisely the same.
You can copy over the ttf files. Personally, I always want verdana for web browsing under Linux; verdana is one thing that MS haters give them credit for doing right.
Pick appropriate fonts. To me, most fonts only look good under X at certain sizes (depending on the font design, rendering engine, et cetera, et cetera). I find that the old X bitmap fonts are best in terminals (ymmv). I like helvetica for small labels. Play with whatever settings are available; things look different on CRTs and LCDs, and your settings might be optimized for the wrong type of monitor.
Konqueror has nice facilities for browsing a directory of ttf files.
"Technically savvy users need little more than an IP address and a beer to do the right thing. Hell, our sysadmins consult with me to help figure out how to do things right."
Simple tests: Which one (the user or the IT person) tests the network by using ping and traceroute and which uses a web browser? Which one says "the internet is broken" and which one says "the nameserver has stopped responding to requests"? Which knows how to change the IP address without rebooting?
Questions like these tell us who should be administering the machine. Sometimes it will be the user, sometimes IT.
"Any public interfaces to the kernel should be treated like a library and licensed appropriately"
It depends how you define the public interfaces. Most people would say that the public interfaces are the system calls. And any program under any license can run on Linux and use the system calls.
You have a completely different view of what is a public interface. Who gets to define which interfaces are public? If it is the code authors, then you can say that the kernel authors have decided that the interface in question is not public.
"why do they have to pass a checklist of requirements to release their code freely?"
The problem is that it is not "their code". They claim that the "project implements Windows kernel API and NDIS (Network Driver Interface Specification) API within Linux kernel." That means it is a derivative work of the kernel. That means that they don't own all the code themselves, and that they have to follow the license of the other code they are using. And Linus's interpretation of the GPL and kernel modules is a lot more permissive than what is probably legally correct.
"This is stupid, people are trying to release the code of the project to the community and the restrictive terms of the GPL is preventing them."
This would be different if it were purely their code, but it isn't. This isn't stand-alone code. What they created is a derivative work of the Linux kernel. They used code which they didn't write and they don't own. Your argument is that the people who actually wrote the original kernel code have no right to say how it and its derivative works are used. Legally, you are completely and totally wrong. Essentially, you are advocating an end to copyrights on computer code. That's fine, but it has nothing to do with the GPL.
"I'm not entirely sure how you get the power loss there to be under 2 watts."
Easy. I was in a hurry, and I screwed up the calculation. Thanks for not calling me an idiot even thought it is justified.
Let me try again. 10 m of 12 gauge wire has a resistance of 0.061 ohms. If we say that we don't want to lose more than 5 W at the highest load, that lets us carry 9 A. That also means a half volt drop. Assuming a 12 V system, that means I am losing 10 W (have to count the return too) to power a 108 W device. Hopefully, a well-designed dedicated DC power unit could be 10% more efficient than a PC power supply. But 108 W is really too small for what we want. We can tweak things to get a factor of two here or there, but it is really not adequate.
So we have to go to extremes to make this work in a residential dwelling. First option, put the DC supply next to the computer and run wire for low power devices (chargers, clocks...). Second option, keep all devices low power (green PC, not a server). Third option, run 1" bus bar through the basement, send cable up through the walls for the outlets. Bus bar is probably easier to work with than 1" cable, because for the latter you need to get lugs to make connections, but you can just drill holes in the bus bar.
"Having one big DC supply with lots of outlets in your home is likely to be less energy efficient (because of the large voltage loss over long cable runs at high currents) than having local ones."
Think about what you are saying and try again.
The ArsTechnica article says that the difference between a good power supply and an inefficient one is 10-20% of the total power thrown out. This might be 10-50 W depending on the computer. That is a crap load of power.
Now look at the power lost by sending 10 amps over 10 meters of 16 gauge wire (which is pretty thin wire). It is under 2 watts. Use 12 gauge wire and it is under 1 watt.
So by using a central, high-efficiency AC-to-DC converter you save tens of watts but you lose a watt in distributing the power. Sounds good to me.
Now consider how many really inefficient DC power supplies there are lurking around the house. How about the devices that put the power switch on the low-voltage side of the transformer so they burn watts whenever they are plugged in (I had an HP printer like that)?
I doubt we could ever get manufacturers to agree on a single DC voltage to use, though. I don't want outlets with 15 different plugs covering every DC voltage everyone wants to use.
I don't suppose that I can fault you for stating the blindingly obvious.
Joel's writing may be interesting and insightful, but his enormous blind spots make him anything but authoritative. Take this gem:
A lot of the complexities in these file formats reflect features that are old, complicated, unloved, and rarely used. They're still in the file format for backwards compatibility, and because it doesn't cost anything for Microsoft to leave the code around.
So it "doesn't cost anything" for the company with a reputation for the buggiest, least secure code around to keep a bunch of legacy code in its aps? Especially code that was not written for maintainability? Joel only sees issues from a very particular perspective.
"Thus, you need to specifically support IE in your web pages unless you're ok with leaving the majority of your potential users in the dark (which no sane person would be)."
That's where you make your mistake. Lots of sane people don't care how MSIE renders the web pages they create. These people are not commercial web developers. Two different groups, two different goals. A lot of things that other people do make no sense if you only look at them from your own perspective.
There are other possibilities that are more likely. For example, perhaps the patched kernel is doing something like loading microcode into the processor. The kernel code would be GPLed but the microcode would not be.
Think how much money they could have saved if everyone just voted on the mass of the neutrino instead performing all those expensive experiments. (And this probably would have resulted in the wrong answer, too.)
But I am not talking about truth, I am talking about presentation. The view of the minority should not be suppressed, but that does not mean that it should be presented as equivalent to the consensus view.
Sometimes the consensus view is a house of cards, but usually it isn't. It does a disservice to the reading public to maintain that every issue is seriously disputed.
Is it too much to ask that science journalists (by which I mean anyone who writes an article about a scientific topic for a mainstream magazine or newspaper) be conversant with at least the level of science taught in high school?
Perhaps a general understanding of the electromagnetic spectrum?
And the difference between a watt and a joule, or between 1 rem and 1 rem/hr?
Or that hydrogen is not an energy source for denizens of the surface of the planet earth?
A vague grasp of statistics is too much to ask, I know.
Perhaps most important: If 99.8% of scientists think one thing, and 0.2% think another, this is not a dispute with two equally valid points of view. This is a consensus and a handful of wackjobs which should be presented as such.
I don't understand. The GPL requires that the source code of distributed binaries must be available (with multiple options depending on the details). No part of the MS-PL requires that. The portion that you quoted only refers to the licensing of distributed code. How is that related?
What you raise appears to be a subtle philosophical point with no obvious consequences.
The starting point is different. The GPL adds to the rights that you have under copyright law. It is purely permission to do things that you otherwise could not legally do. A EULA starts with the assumption that copyright holders should be able to control the use of the software. But that difference doesn't appear to affect the operation of the licenses in this case.
There could be subtle legal issues about contracts without compensation, the difference between a contract and a license, and all sorts of things that have never been worked out in court. Legally, the GPL and MS-PL could wind up working very differently, but there is no way to predict now if there are likely to be differences or what they would be.
"Your ideas are interesting but they are also flatly wrong in this case."
Sorry, but you cannot disprove, "The concern is that X could reasonably happen" by saying "X hasn't happened yet." I am saying that the Open Source status of the MS-PL could be abused. Your argument is that the Open Source status of the MS-PL cannot be abused because the MS-PL is currently in use. That makes no sense.
It would be ridiculous to assert that no one using the MS-PL would ever make the source available, and I don't make such a stupid claim.
There are open source projects. There are Open Source licenses. The trouble is that using a certified Open Source license does not necessarily result in a project that is open source by any meaningful definition.
This sounds like nonsense and therefore requires explanation. Pick your favorite BSD. Would you call it open source? Certainly. You can download the source and do almost whatever you want with it. Is the source available because the license requires it or because the authors decided to make it available? The source is available because the authors decided to make it available.
In theory, I could write a piece of code and distribute only binaries under the BSD license, an OSI certified Open Source license, such as the BSDL. It is possible to have Open Source code where no one but the author has the source. How is this open source in any meaningful sense?
Here is the problem. Most of the people submit licenses for certification that are used on genuine open source projects. The MS-PL is submitted in a vacuum - anything using it is hypothetical at this point. OSI knows that a certified Open Source license does not guarantee an open source project, and they depend on the goodwill of the participants. Depending on the goodwill of Microsoft is somewhere between naivete and insanity. They failed to design the license definition to avoid misuse, and they are trying to make up for it by ignoring their own definition.
No one expects lots of software to appear under the BSDL with no source available. On the other hand, it seems reasonable to expect Microsoft to offer MS-PL licensed software with no source and loudly trumpet how much Open Source code they distribute.
Slashdot Mirror
First, equation solving. This means numerical integration, root finding, and so on.
Second, spectral methods.
Third, numerical error. What does it mean for a numerical calculation to be ill-conditioned?
If they start out without the skills needed to write a trivial program, the physics department should not have to teach them. Don't people learn that much in high school?
The most important thing is the hardest to teach in a course of limited scope: writing readable, maintainable, debuggable code. Most scientists suck at this.
What language? Something that graphs! Something easy to debug (they don't need to spend their lives chasing one-off indexing errors). Good languages with OSS semi-equivalents are Matlab (Octave) and IDL (GDL). Is a spreadsheet appropriate? I don't know enough to say. If someone learns on a spreadsheet, will they be able to more-or-less understand a piece of code in F90? I also object to forcing people to purchase from Microsoft.
Er, Peter, would you like to see my copy of RHL 4.something with a proprietary X server on it? I believe it was something like MetroX. Netscape 4 was also available as part of RHL for a while (looks like at least 6 and 7).
Not saying this to attack RH; their distros are the only ones I ever use.
I couldn't get freezing to work on my dead drive. The trick that worked was: Let sit on the desk for two months and then try it again. It still made noise, but it worked long enough to find and retrieve the files that weren't backed up.
(I am not denying that it is important or useful. I just can't understand how to make it work.)
The assumption is that any position must be liberal or conservative, but this leads to nonsense. There is a large centrist middle. So what we see is that the largely centrist press looks liberal to extreme conservatives and conservative to extreme liberals. For example, being against abolishing social security or against abolishing the death penalty are not liberal or conservative, they are centrist. Does the media have a centrist bias?
Furthermore, support or criticism of a particular politician is not evidence of bias. It is poor reporting. Name your most or least favorite politician - probably they are in office because of a period of uncritical support by the media and their biggest problem is unfair, inaccurate, negative reporting in the media.
You can copy over the ttf files. Personally, I always want verdana for web browsing under Linux; verdana is one thing that MS haters give them credit for doing right.
Pick appropriate fonts. To me, most fonts only look good under X at certain sizes (depending on the font design, rendering engine, et cetera, et cetera). I find that the old X bitmap fonts are best in terminals (ymmv). I like helvetica for small labels. Play with whatever settings are available; things look different on CRTs and LCDs, and your settings might be optimized for the wrong type of monitor.
Konqueror has nice facilities for browsing a directory of ttf files.
Simple tests: Which one (the user or the IT person) tests the network by using ping and traceroute and which uses a web browser? Which one says "the internet is broken" and which one says "the nameserver has stopped responding to requests"? Which knows how to change the IP address without rebooting?
Questions like these tell us who should be administering the machine. Sometimes it will be the user, sometimes IT.
You are wrong. It has certainly been settled that the source code is not a published specification.
It depends how you define the public interfaces. Most people would say that the public interfaces are the system calls. And any program under any license can run on Linux and use the system calls.
You have a completely different view of what is a public interface. Who gets to define which interfaces are public? If it is the code authors, then you can say that the kernel authors have decided that the interface in question is not public.
The problem is that it is not "their code". They claim that the "project implements Windows kernel API and NDIS (Network Driver Interface Specification) API within Linux kernel." That means it is a derivative work of the kernel. That means that they don't own all the code themselves, and that they have to follow the license of the other code they are using. And Linus's interpretation of the GPL and kernel modules is a lot more permissive than what is probably legally correct.
This would be different if it were purely their code, but it isn't. This isn't stand-alone code. What they created is a derivative work of the Linux kernel. They used code which they didn't write and they don't own. Your argument is that the people who actually wrote the original kernel code have no right to say how it and its derivative works are used. Legally, you are completely and totally wrong. Essentially, you are advocating an end to copyrights on computer code. That's fine, but it has nothing to do with the GPL.
I think electrical discharge in the dialectric would happen before physical breakdown in some cases.
Lots of stored energy is bad. And the solution is...?
He seems to be conflating open source principles and the Unix philosophy. Perhaps a principle of open source would be, um, opening the source.
Easy. I was in a hurry, and I screwed up the calculation. Thanks for not calling me an idiot even thought it is justified.
Let me try again. 10 m of 12 gauge wire has a resistance of 0.061 ohms. If we say that we don't want to lose more than 5 W at the highest load, that lets us carry 9 A. That also means a half volt drop. Assuming a 12 V system, that means I am losing 10 W (have to count the return too) to power a 108 W device. Hopefully, a well-designed dedicated DC power unit could be 10% more efficient than a PC power supply. But 108 W is really too small for what we want. We can tweak things to get a factor of two here or there, but it is really not adequate.
So we have to go to extremes to make this work in a residential dwelling. First option, put the DC supply next to the computer and run wire for low power devices (chargers, clocks...). Second option, keep all devices low power (green PC, not a server). Third option, run 1" bus bar through the basement, send cable up through the walls for the outlets. Bus bar is probably easier to work with than 1" cable, because for the latter you need to get lugs to make connections, but you can just drill holes in the bus bar.
Think about what you are saying and try again.
The ArsTechnica article says that the difference between a good power supply and an inefficient one is 10-20% of the total power thrown out. This might be 10-50 W depending on the computer. That is a crap load of power.
Now look at the power lost by sending 10 amps over 10 meters of 16 gauge wire (which is pretty thin wire). It is under 2 watts. Use 12 gauge wire and it is under 1 watt.
So by using a central, high-efficiency AC-to-DC converter you save tens of watts but you lose a watt in distributing the power. Sounds good to me.
Now consider how many really inefficient DC power supplies there are lurking around the house. How about the devices that put the power switch on the low-voltage side of the transformer so they burn watts whenever they are plugged in (I had an HP printer like that)?
I doubt we could ever get manufacturers to agree on a single DC voltage to use, though. I don't want outlets with 15 different plugs covering every DC voltage everyone wants to use.
Joel's writing may be interesting and insightful, but his enormous blind spots make him anything but authoritative. Take this gem:
So it "doesn't cost anything" for the company with a reputation for the buggiest, least secure code around to keep a bunch of legacy code in its aps? Especially code that was not written for maintainability? Joel only sees issues from a very particular perspective.That's where you make your mistake. Lots of sane people don't care how MSIE renders the web pages they create. These people are not commercial web developers. Two different groups, two different goals. A lot of things that other people do make no sense if you only look at them from your own perspective.
There are other possibilities that are more likely. For example, perhaps the patched kernel is doing something like loading microcode into the processor. The kernel code would be GPLed but the microcode would not be.
Think how much money they could have saved if everyone just voted on the mass of the neutrino instead performing all those expensive experiments. (And this probably would have resulted in the wrong answer, too.)
But I am not talking about truth, I am talking about presentation. The view of the minority should not be suppressed, but that does not mean that it should be presented as equivalent to the consensus view.
Sometimes the consensus view is a house of cards, but usually it isn't. It does a disservice to the reading public to maintain that every issue is seriously disputed.
Perhaps a general understanding of the electromagnetic spectrum?
And the difference between a watt and a joule, or between 1 rem and 1 rem/hr?
Or that hydrogen is not an energy source for denizens of the surface of the planet earth?
A vague grasp of statistics is too much to ask, I know.
Perhaps most important: If 99.8% of scientists think one thing, and 0.2% think another, this is not a dispute with two equally valid points of view. This is a consensus and a handful of wackjobs which should be presented as such.
I don't understand. The GPL requires that the source code of distributed binaries must be available (with multiple options depending on the details). No part of the MS-PL requires that. The portion that you quoted only refers to the licensing of distributed code. How is that related?
What does the "L" in EULA stand for?
What you raise appears to be a subtle philosophical point with no obvious consequences.
The starting point is different. The GPL adds to the rights that you have under copyright law. It is purely permission to do things that you otherwise could not legally do. A EULA starts with the assumption that copyright holders should be able to control the use of the software. But that difference doesn't appear to affect the operation of the licenses in this case.
There could be subtle legal issues about contracts without compensation, the difference between a contract and a license, and all sorts of things that have never been worked out in court. Legally, the GPL and MS-PL could wind up working very differently, but there is no way to predict now if there are likely to be differences or what they would be.
Sorry, but you cannot disprove, "The concern is that X could reasonably happen" by saying "X hasn't happened yet." I am saying that the Open Source status of the MS-PL could be abused. Your argument is that the Open Source status of the MS-PL cannot be abused because the MS-PL is currently in use. That makes no sense.
It would be ridiculous to assert that no one using the MS-PL would ever make the source available, and I don't make such a stupid claim.
There are open source projects. There are Open Source licenses. The trouble is that using a certified Open Source license does not necessarily result in a project that is open source by any meaningful definition.
This sounds like nonsense and therefore requires explanation. Pick your favorite BSD. Would you call it open source? Certainly. You can download the source and do almost whatever you want with it. Is the source available because the license requires it or because the authors decided to make it available? The source is available because the authors decided to make it available.
In theory, I could write a piece of code and distribute only binaries under the BSD license, an OSI certified Open Source license, such as the BSDL. It is possible to have Open Source code where no one but the author has the source. How is this open source in any meaningful sense?
Here is the problem. Most of the people submit licenses for certification that are used on genuine open source projects. The MS-PL is submitted in a vacuum - anything using it is hypothetical at this point. OSI knows that a certified Open Source license does not guarantee an open source project, and they depend on the goodwill of the participants. Depending on the goodwill of Microsoft is somewhere between naivete and insanity. They failed to design the license definition to avoid misuse, and they are trying to make up for it by ignoring their own definition.
No one expects lots of software to appear under the BSDL with no source available. On the other hand, it seems reasonable to expect Microsoft to offer MS-PL licensed software with no source and loudly trumpet how much Open Source code they distribute.