This is probably one of the crucial questions for the future development of "standards", i.e. software platforms that everyone, including commercial entrepreneurs, can use to provide better products to the end user.
Bertrand Meyer got the point when he wrote one must distinguish between free as in "gratuit" and free as in "libre".
The main questions that at this point should probably be raised are linked to the quality of the product (and responsability in case things do not work as advertised), the remuneration of those who have contributed (yes indeed), maintaining interoperability.
To take this last point, this seems usually to be done by having some sort of *dictator* deciding about what is included and what is not. This dictator can be considered benevolent (linus thorvald) of evil (bill gates...) but still there seem to be a need to have someone decide what is in and what is out.
This should of course not mean that developpers are not allowed to ass their own extensions, but it is rather about deciding what has the right to carry the name (eg Linux or Windows) and what is just proprietary. By the way this can be done by committee (as in any standardisation body) but of course you are replacing the *dictator* by a political process that should strive to be democratic. This appears to be very inefficient - it often is-, but remember "democracy is the worst of all systems except all others".
The second issue - how to be paid - must not be neglected. One of the difficulties in software is that there is often no difference between the specification of the code and the actual code (there is no metacode describing multiple codes). If this existed the solution would be simpler: the metacode is open and paid for in a kind of pool arrangement or under the leadership of a *publisher* (like any collective work - as when people were writing contributions for the Encyclopedia Britannica) and the code which implements it is sold as a proprietary product. The publisher in this cas could be private or collective.
By the way isn't it what is actually the direction taken by Linus and Linux ? Publisher Linus directs the work on Linux, distributions are sold as products. But because this duality (metacode and code) is not totally accepted, enough attention cannot be paid to interoperability.
The question of quality is also central. There was a mention of the Java Community Licence. Sun says it has set up this process to ensure quality and non-fragmentation. But many feel that is it also very self serving: the ownership of the contributions (copyrights and other IPRs) go automatically to a pool, and in the end it is only Sun who decides what and when things go into the product. No negligeable advantage, being the master of timing! This is one of the reasons Java has never been able to go through an recognised standardisation process.
To get back to initial argument, it does seem very important, for the future of software, to develop the free as in *libre* part (open access and open evolution), which is quite different from the *gratuit* part. Mixing the two is not helpful.
It seems to me that Meyer's article, by highlighting the issue and trying to steer it away from ethics, is a welcome contribution.
Slashdot Mirror
This is probably one of the crucial questions for the future development of "standards", i.e. software platforms that everyone, including commercial entrepreneurs, can use to provide better products to the end user.
Bertrand Meyer got the point when he wrote one must distinguish between free as in "gratuit" and free as in "libre".
The main questions that at this point should probably be raised are linked to the quality of the product (and responsability in case things do not work as advertised), the remuneration of those who have contributed (yes indeed), maintaining interoperability.
To take this last point, this seems usually to be done by having some sort of *dictator* deciding about what is included and what is not. This dictator can be considered benevolent (linus thorvald) of evil (bill gates...) but still there seem to be a need to have someone decide what is in and what is out.
This should of course not mean that developpers are not allowed to ass their own extensions, but it is rather about deciding what has the right to carry the name (eg Linux or Windows) and what is just proprietary. By the way this can be done by committee (as in any standardisation body) but of course you are replacing the *dictator* by a political process that should strive to be democratic. This appears to be very inefficient - it often is-, but remember "democracy is the worst of all systems except all others".
The second issue - how to be paid - must not be neglected. One of the difficulties in software is that there is often no difference between the specification of the code and the actual code (there is no metacode describing multiple codes). If this existed the solution would be simpler: the metacode is open and paid for in a kind of pool arrangement or under the leadership of a *publisher* (like any collective work - as when people were writing contributions for the Encyclopedia Britannica) and the code which implements it is sold as a proprietary product. The publisher in this cas could be private or collective.
By the way isn't it what is actually the direction taken by Linus and Linux ? Publisher Linus directs the work on Linux, distributions are sold as products. But because this duality (metacode and code) is not totally accepted, enough attention cannot be paid to interoperability.
The question of quality is also central. There was a mention of the Java Community Licence. Sun says it has set up this process to ensure quality and non-fragmentation. But many feel that is it also very self serving: the ownership of the contributions (copyrights and other IPRs) go automatically to a pool, and in the end it is only Sun who decides what and when things go into the product. No negligeable advantage, being the master of timing! This is one of the reasons Java has never been able to go through an recognised standardisation process.
To get back to initial argument, it does seem very important, for the future of software, to develop the free as in *libre* part (open access and open evolution), which is quite different from the *gratuit* part. Mixing the two is not helpful.
It seems to me that Meyer's article, by highlighting the issue and trying to steer it away from ethics, is a welcome contribution.