I get a bunch of sources and I modify them to build a similar or updated application. This is clearly a derived work of the application.
I get a bunch of library sources and I modify them to do the same as above. This is clearly a derived work of the libraries.
I get the sources (header files) to a bunch of libraries, and I do not do anything to them. I just use them as components in my own application. Is including a header file into your sources a derived work of those header files ? Can this be likened to the following reasoning :
Is a clock a derived work of a pinwheel ?
Is a car a derived work of an axle ?
Etc.
Instead of treating GPL'd libraries as the basis for a derived work, I think that the GPL should turn around the reasoning and think about libraries as components. This would mean that if someone built an application using GPL libraries, he is bound to distribute the source of the libraries, but in the current case he should not be bound to make available his own sources under the GPL.
I know this is dangerous ground, but it is dangerous because the FSF has not been able to resolve this issue unambiguously, and I think that considering GPL'd libraries as components instead of as something that provides a base for a derived work will work much better.
Plant A, under evolutionary pressure, develops a mechanism with which it protects itself from moas.
Plant B, which is not under evolutionary pressure, does not develop such a system.
Evolutionary pressure disappears, but growing the defense mechanism does not constitute an evolutionary disadvantage, so it stays in place.
Under the influence of random mutations, some plants might revert back to the old style, but this is a big might, since evolution works more by accretion than by shedding things.
You forgot another thing in this evolution : the creation of two commercial entities which built Lisp computers based on the code that was available from MIT and which became closed systems.
There was never a threat to the concept that people exchanging ideas freely could do interesting things with computers. It's a very compelling concept that creates a sense of community.
But are you willing to always reinvent the wheel, because that's what you will ultimately get. In this case the GPL is also a form of pragmatism : you have the certainty that you will have access to an up to date and broad library of software with which you can do (almost) anything that you want. The almost is : if you distribute the executables, you must make sure that the corresponding sources are somehow accessible for anyone receiving the executables. Now, that is not an obstacle in the way of the above expression, isn't it ?
I find it unbelievable how things have made full circle in 12 years time.
Between 1997 and 2000 I worked as a programmer for two companies, using COBOL on WANG VS. The first one was in a small transport firm (~20 employees), and my main task was writing software for supporting the business. I also had to support the PC's, but that was strictly bare minimum. As a programmer, it was my task to know the business and be able to write software which helped the owner leverage DP to get more income or manage his resources better.
The second one was a small bank (~800 employees, programming staff included). I think there must have been around 50 programmers, but there were only three or four people doing what is now called IT : supporting PC's, administrating the WANG VS platform. They also had access to mainframes, but these where leased, so the support people there where from Honeywell.
So, in the last 12 years, people who do not know anything about programming, and thus do not know to add value to the business, have been steadily creeping into companies in order to provide all kinds of support, or rather babysitting systems, so that the people who do the real work can do their job. Decrementing this workforce is good, because it makes it possible to hire programmers and analysts to implement things not only for helping people do their job, but helping people do their job better.
I rather talk about DP than about IT, because it is DP which is able to add value to a business, due to storing, processing and retrieval of data.
The problem with Wallonia is that they used to be rich, in the times that coal and steel ruled the world. Unfortunately for them they never caught up with the times after their mines got depleted and so their working class shifted into poverty. The PS of Wallonia takes advantage of their poor for keeping their majority and nothing is done to help their poor. Some parts of Wallonia (e.g. the Borinage) are really like third-world countries. The PS uses then economic reasons to get subsidies from Europe.
There are people in Wallonia trying to make a difference, but unfortunately they are a minority.
I had enough problems in the first half of the 90's due to viruses from the sneakernet. The internet was never needed to propagate malware, it helped only to propagate it faster.
What people really want then are ISAM files. These can deliver all the performance and security of an RDBMS without the R or the SQL.
In fact, I have written applications in such a system : WANG PACE, which was a relational database system built on top of the WANG VS ISAM file structure.
Admit it, guys : deep down inside you all want to program in COBOL.
I do not know if "Structure and Interpretation of Computer Programs" is THE reference book for Computer Science curricula, but it seems to me that if one has had it in CS and solved the posed exercises, then one should already be a fairly competent programmer, seeing as the book goes into quite a level of detail in a whole lot of different problem domains.
It seems that the main background for the book is that one should be capable of reasoning about problems, and also know how to write code to implement the necessary algorithms. The book's emphasis is on theory AND practice, and if you have a look at the MIT curriculum with regard to this book, it is also emphasised that all exercises should be made.
So, how should this translate in people with a CS degree not knowing (needing) how to program ?
This is bullshit. When I started working in 1990, there where quite some viruses which ran under Mac OS. One of the vulnerabilities on the Mac was the fact that when you inserted a floppy, there was always code that was immediately executed.
Yes, but for bridges and highways there is a lower limit based upon public safety and maintenance costs. For software (per se, not in the context of another engineering branch) this is not so, and this makes it possible to sell crappier and crappier software, where the bug fixes are sold as added value and extra modules.
Emacs can be compiled without problems under Windows, but is then strictly Windows, with \ for paths and \r\n for line endings.
You can also compile Emacs under Cygwin, but then it is strictly POSIX and needs a X server to run, otherwise it runs in -nw mode.
Xemacs, however, can be compiled under Cygwin, but recognises that it is under Windows and runs all graphics natively, obviating the need for an X server. That is why I currently only use XEmacs on my job, it works POSIX with Cygwin, but runs natively under Windows.
The Actors Model is around 40 years old. Scheme was based upon it. Lisp have already shown in the eighties to be good at concurrent programming. Just NIH syndrome.
And the free market works even here. Here in Belgium you can choose between the Christelijke Mutualiteit, Bond Moyson and the neutral health care insurance.
No. I tried to update my Ubuntu 8.04 on my 64-bit AMD laptop and the ATI Radeon drivers are wrong, so I had to turn back my installation completely.
I tried to upgrade my Debian etch to lenny on my 200 Mhz PPro machine with 256 Mb memory, and even that was not completely painless because of changes in Xorg and sylpheed-claws (now claws mail), so I had to turn back that installation too.
Both experiences show that supporting legacy systems with Linux has become difficult : you need to use older distro's on older machines.
Slashdot Mirror
Perl at work, Common Lisp for hobby projects.
And I can, and do, work with vi and Emacs!
This looks suspiciously like a definition of Lisp. 50 years prior art ?
Let's take this line of thought.
I get a bunch of sources and I modify them to build a similar or updated application. This is clearly a derived work of the application.
I get a bunch of library sources and I modify them to do the same as above. This is clearly a derived work of the libraries.
I get the sources (header files) to a bunch of libraries, and I do not do anything to them. I just use them as components in my own application. Is including a header file into your sources a derived work of those header files ? Can this be likened to the following reasoning :
Instead of treating GPL'd libraries as the basis for a derived work, I think that the GPL should turn around the reasoning and think about libraries as components. This would mean that if someone built an application using GPL libraries, he is bound to distribute the source of the libraries, but in the current case he should not be bound to make available his own sources under the GPL.
I know this is dangerous ground, but it is dangerous because the FSF has not been able to resolve this issue unambiguously, and I think that considering GPL'd libraries as components instead of as something that provides a base for a derived work will work much better.
This does not prove anything.
Plant A, under evolutionary pressure, develops a mechanism with which it protects itself from moas.
Plant B, which is not under evolutionary pressure, does not develop such a system.
Evolutionary pressure disappears, but growing the defense mechanism does not constitute an evolutionary disadvantage, so it stays in place.
Under the influence of random mutations, some plants might revert back to the old style, but this is a big might, since evolution works more by accretion than by shedding things.
I really do not see anything relevant here.
You forgot another thing in this evolution : the creation of two commercial entities which built Lisp computers based on the code that was available from MIT and which became closed systems.
There was never a threat to the concept that people exchanging ideas freely could do interesting things with computers. It's a very compelling concept that creates a sense of community.
But are you willing to always reinvent the wheel, because that's what you will ultimately get. In this case the GPL is also a form of pragmatism : you have the certainty that you will have access to an up to date and broad library of software with which you can do (almost) anything that you want. The almost is : if you distribute the executables, you must make sure that the corresponding sources are somehow accessible for anyone receiving the executables. Now, that is not an obstacle in the way of the above expression, isn't it ?
I find it unbelievable how things have made full circle in 12 years time.
Between 1997 and 2000 I worked as a programmer for two companies, using COBOL on WANG VS. The first one was in a small transport firm (~20 employees), and my main task was writing software for supporting the business. I also had to support the PC's, but that was strictly bare minimum. As a programmer, it was my task to know the business and be able to write software which helped the owner leverage DP to get more income or manage his resources better.
The second one was a small bank (~800 employees, programming staff included). I think there must have been around 50 programmers, but there were only three or four people doing what is now called IT : supporting PC's, administrating the WANG VS platform. They also had access to mainframes, but these where leased, so the support people there where from Honeywell.
So, in the last 12 years, people who do not know anything about programming, and thus do not know to add value to the business, have been steadily creeping into companies in order to provide all kinds of support, or rather babysitting systems, so that the people who do the real work can do their job. Decrementing this workforce is good, because it makes it possible to hire programmers and analysts to implement things not only for helping people do their job, but helping people do their job better.
I rather talk about DP than about IT, because it is DP which is able to add value to a business, due to storing, processing and retrieval of data.
Read 'Stark'
The problem with Wallonia is that they used to be rich, in the times that coal and steel ruled the world. Unfortunately for them they never caught up with the times after their mines got depleted and so their working class shifted into poverty. The PS of Wallonia takes advantage of their poor for keeping their majority and nothing is done to help their poor. Some parts of Wallonia (e.g. the Borinage) are really like third-world countries. The PS uses then economic reasons to get subsidies from Europe.
There are people in Wallonia trying to make a difference, but unfortunately they are a minority.
Courtesy Ben Elton
Why do I always have to think about Rube Goldberg if Bill Gates proposes a plan ?
I had enough problems in the first half of the 90's due to viruses from the sneakernet. The internet was never needed to propagate malware, it helped only to propagate it faster.
Probably the biggest example of the broken windows fallacy (pun intended).
...from its metabolism ?
What people really want then are ISAM files. These can deliver all the performance and security of an RDBMS without the R or the SQL.
In fact, I have written applications in such a system : WANG PACE, which was a relational database system built on top of the WANG VS ISAM file structure.
Admit it, guys : deep down inside you all want to program in COBOL.
I do not know if "Structure and Interpretation of Computer Programs" is THE reference book for Computer Science curricula, but it seems to me that if one has had it in CS and solved the posed exercises, then one should already be a fairly competent programmer, seeing as the book goes into quite a level of detail in a whole lot of different problem domains.
It seems that the main background for the book is that one should be capable of reasoning about problems, and also know how to write code to implement the necessary algorithms. The book's emphasis is on theory AND practice, and if you have a look at the MIT curriculum with regard to this book, it is also emphasised that all exercises should be made.
So, how should this translate in people with a CS degree not knowing (needing) how to program ?
This is bullshit. When I started working in 1990, there where quite some viruses which ran under Mac OS. One of the vulnerabilities on the Mac was the fact that when you inserted a floppy, there was always code that was immediately executed.
Yes, but for bridges and highways there is a lower limit based upon public safety and maintenance costs. For software (per se, not in the context of another engineering branch) this is not so, and this makes it possible to sell crappier and crappier software, where the bug fixes are sold as added value and extra modules.
I think this is a really nice summary of most of the computer industry.
I always supposed that Ogg came from Nanny Ogg, which is also a character in the Discworld series.
Well, now you say it, that was wrong. It works with drive letters, but the paths are indeed all with forward slashes.
Emacs can be compiled without problems under Windows, but is then strictly Windows, with \ for paths and \r\n for line endings.
You can also compile Emacs under Cygwin, but then it is strictly POSIX and needs a X server to run, otherwise it runs in -nw mode.
Xemacs, however, can be compiled under Cygwin, but recognises that it is under Windows and runs all graphics natively, obviating the need for an X server. That is why I currently only use XEmacs on my job, it works POSIX with Cygwin, but runs natively under Windows.
The Actors Model is around 40 years old. Scheme was based upon it. Lisp have already shown in the eighties to be good at concurrent programming. Just NIH syndrome.
And the free market works even here. Here in Belgium you can choose between the Christelijke Mutualiteit, Bond Moyson and the neutral health care insurance.
No, but with database applications you go for IO first, CPU second.
No. I tried to update my Ubuntu 8.04 on my 64-bit AMD laptop and the ATI Radeon drivers are wrong, so I had to turn back my installation completely.
I tried to upgrade my Debian etch to lenny on my 200 Mhz PPro machine with 256 Mb memory, and even that was not completely painless because of changes in Xorg and sylpheed-claws (now claws mail), so I had to turn back that installation too.
Both experiences show that supporting legacy systems with Linux has become difficult : you need to use older distro's on older machines.