My impression is that, at least as of the time they were designing it, nobody knew how to make a really reliable gyro. Therefore, they built in lots of redundancy (3 spares!), and figured they would have time to replace the spares before they all ran out. Remember, the telescope was intended to be serviced at regular intervals, for upgrades of instruments if nothing else. Unfortunately, there seems to have been a miscalculation, either in the reliability of the the gyros, the reliability of the space shuttle, or both.
There is something one should remember when looking at space equipment failures: whoever puts it up is severely limited in how thoroughly they can test it. Essentially, anything that gets put up has to be considered an alpha version. Unfortunately, there is no beta. I am astounded that much of the stuff that gets sent up works at all; that it does is a real testament to the care and skill of those who build it.
Yes, an awful lot of free software is developed at educational instutions (by both students and faculty), which can get funded because students can get loans and scolarships. However, these are investments in and contrubutions to the education of students, not the developement of software. If that were the intent, they'd lend (or give) to software companies.
I do think an argument can be made that students and faculty that develop software in an academic setting should license it for the maximum benefit to the academic community. After all, most of the scolarships and grants that help support it are meant to improve education and human knowledge.
The banks, however, are entirely irrelevant to ones moral obligation with respect to ones code. It seems a little like saying the bank that financed your car should get your pay check, because you use it to drive to work. Yes, if nobody could borrow money to buy cars, the economy would suffer, but banks hardly deserve credit for the work done by everyone who borrowed money to buy the car with which they drive to work.
While it would be nice if Linux were easy enough anyone could run it, it isn't disasterous if it is not.
The most important strength (among several) of Linux (and open source software generally) is flexibility. Those who want (or need) the level of flexibility offered by OSS will need the technical ability to take advantage of the flexibility offered; those who need Linux most are those who are able to use it. Furthermore, it is these users who are most able to improve the software. Although convienient, it is not essential that Linux be usable to someone with no computer experience.
I'm not claiming it wouldn't be a very good thing if it were that simple to use. However, it is not essential, and not worth it if it makes it less useful to those that don't need quite that level of simplicity. -Hil
I think there is an important misconception in the article, one that is pretty common in general: that education is about learning information. It is not, at least not in the way normally considered.
In science and engineering courses, the most valuable thing learned is how to solve particular sorts of problems. I am not refering the needed laws and theories directly conveyed; these can generally be looked up as needed. Instead, I refer to the skill one gets (through practice and observation of others) at solving particular sorts of problems generally. I could probably have made it through astronomy grad school knowing only the information I learned in high school. However, very few students (if any at all) just out of high school have anywhere near the problem solving skills necessary to survive graduate education in a hard science.
Can these sorts of skills be taught over a computer network? I doubt it. Where I (and those I TA'd) did most of their learning was in doing problem sets (which you can do anywhere, net or no), lab experience, and tutoring (either from other students, professors, TA's, or professional tutors). This is where the most valuable stuff gets learned in college, not in lectures. Virtual labs are no replacement for real lab experience, and net communications are still too clunky for effective tutoring.
If a student learns the skills properly, any information needed can be learned with minimal effort. (I took a few comp. sci. and EE classes at the undergrad level as a grad student in astro. I went in with less informational knowledge (and probably less talent) than most of the students, and yet did very well relative to the classes as a whole, with less effort. Why? I picked up a lot of programming and math "skills" solving physics problems.)
I suspect the situation for teaching many of the humanities over the net is nearly as bad. The most valuable aspects of these sorts of courses I took were learned in discussion with peers and professors, and in writing papers, and in careful analysis of my papers by others. (The article seems to suggest that these classes are, in general, not worth taking anyway. I disagree. They are less vocationally useful, but have helped me enjoy my life more, and I would never have known what I was missing otherwise.) -Hil
Slashdot Mirror
My impression is that, at least as of the time they were designing it, nobody knew how to make a really reliable gyro. Therefore, they built in lots of redundancy (3 spares!), and figured they would have time to replace the spares before they all ran out. Remember, the telescope was intended to be serviced at regular intervals, for upgrades of instruments if nothing else. Unfortunately, there seems to have been a miscalculation, either in the reliability of the the gyros, the reliability of the space shuttle, or both.
There is something one should remember when looking at space equipment failures: whoever puts it up is severely limited in how thoroughly they can test it. Essentially, anything that gets put up has to be considered an alpha version. Unfortunately, there is no beta. I am astounded that much of the stuff that gets sent up works at all; that it does is a real testament to the care and skill of those who build it.
Yes, an awful lot of free software is developed at educational instutions (by both students and faculty), which can get funded because students can get loans and scolarships. However, these are investments in and contrubutions to the education of students, not the developement of software. If that were the intent, they'd lend (or give) to software companies.
I do think an argument can be made that students and faculty that develop software in an academic setting should license it for the maximum benefit to the academic community. After all, most of the scolarships and grants that help support it are meant to improve education and human knowledge.
The banks, however, are entirely irrelevant to ones moral obligation with respect to ones code. It seems a little like saying the bank that financed your car should get your pay check, because you use it to drive to work. Yes, if nobody could borrow money to buy cars, the economy would suffer, but banks hardly deserve credit for the work done by everyone who borrowed money to buy the car with which they drive to work.
-Hil
While it would be nice if Linux were easy enough anyone could run it, it isn't disasterous if it is not.
The most important strength (among several) of Linux (and open source software generally) is flexibility. Those who want (or need) the level of flexibility offered by OSS will need the technical ability to take advantage of the flexibility offered; those who need Linux most are those who are able to use it. Furthermore, it is these users who are most able to improve the software.
Although convienient, it is not essential that Linux be usable to someone with no computer experience.
I'm not claiming it wouldn't be a very good thing if it were that simple to use. However, it is not
essential, and not worth it if it makes it less useful to those that don't need quite that level of simplicity.
-Hil
I think there is an important misconception in the article, one that is pretty common in general: that education is about learning information. It is not, at least not in the way normally considered.
In science and engineering courses, the most valuable thing learned is how to solve particular sorts of problems. I am not refering the needed laws and theories directly conveyed; these can generally be looked up as needed. Instead, I refer to the skill one gets (through practice and observation of others) at solving particular sorts of problems generally. I could probably have made it through astronomy grad school knowing only the information I learned in high school. However, very few students (if any at all) just out of high school have anywhere near the problem solving skills necessary to survive graduate education in a hard science.
Can these sorts of skills be taught over a computer network? I doubt it. Where I (and those I TA'd) did most of their learning was in doing problem sets (which you can do anywhere, net or no), lab experience, and tutoring (either from other students, professors, TA's, or professional tutors). This is where the most valuable stuff gets learned in college, not in lectures. Virtual labs are no replacement for real lab experience, and net communications are still too clunky for effective tutoring.
If a student learns the skills properly, any information needed can be learned with minimal effort. (I took a few comp. sci. and EE classes at the undergrad level as a grad student in astro. I went in with less informational knowledge (and probably less talent) than most of the students, and yet did very well relative to the classes as a whole, with less effort. Why? I picked up a lot of programming and math "skills" solving physics problems.)
I suspect the situation for teaching many of the humanities over the net is nearly as bad. The most valuable aspects of these sorts of courses I took were learned in discussion with peers and professors, and in writing papers, and in careful analysis of my papers by others. (The article seems to suggest that these classes are, in general, not worth taking anyway. I disagree. They are less vocationally useful, but have helped me enjoy my life more, and I would never have known what I was missing otherwise.)
-Hil