First lets work up the numbers for Windows, we will assume your going to upgrade every pc to XP and we will even make the assumption your hardware will run XP (honestly this is totally fool hardy trust me, ALOT of hardware and software is not supported by XP)So this is really a kind of "once upon a time" type of scenario, mind you linux will save you serious money in every scenario you can dream up. We will assume one thing in both scenarios to be the same, the consultants time will be $100 dollars a hour and they both take the same amount of time so we will ignore this factor. In reality windows takes more time to setup.

Once this is stipulated - and one may agree with it or not, but I'm exploring its ramifactions here - I feel that the basic intent of copyright is being violated with proprietary software. By not publishing the source code - the program itself, they're violating the spirit of copyright: Copyright should be sufficient protection for the program's source code - that's what it's for.

Once copyright expires, the program is supposed to enter the public domain (whether or not it happens in two centuries or tomorrow is immaterial, the issue of copyright length is an entirely different question) - but in the current state of affairs, the program will not enter the public domain because the source code of the program is not a matter of public record.

Given this, I think that only in return for publishing source - the work, strictly speaking, not the generated results of the work - can the creator of the software be deserving of the exclusive right to copy and license that software. If the creators of software release only the compiled binaries for distribution, they should not have the exclusive copyright over those binaries, because they are not the original work - only a product of the original work. Currently, that product is under copyright.

Part of the problem with modern software is that the spirit of copyright is being violated by the publishers of major and minor software works - they're trading their software into the public domain in exchange for a period of protection, but not living up to their end of the bargain; the law does not currently differentiate between compiled and source programs - in much the same way as it does not differentiate between printing plates and finished books.

While technically proprietary software makers are in compliance with copyright, in practice they are not - because the programs will effectively never be in the public domain. Only the compiled binaries of the software will be in the public domain.

Compiled software is therefore impeding progress, it is doing practical, material harm to society through the mechanism of never entering the public domain, like any other work. I think it is reasonable to agree that all works should eventually enter the public domain, and that copyright is designed for this purpose.

For this reason, I advocate that all compiled software, when distributed, should also include source code in a readily-accessible form, and that copyright should thereby be responsible for the usage of the program. There is an aphorism in some circles: if it ain't source, it ain't software, and I think that by changing the definition of "software" in legal documents relating to copyright and software to include this definition could very well fix a major part of the problem facing society and computing today.

Binaries not accompanied by source code could be in the public domain; because it does not honor the "copyright bargain", it should not have the protection of copyright.

This is the ideal situation, but it goes too far for practical purposes - it would be impossible to change the law in such a way in the existing situation. The key point, however, that binaries and source should be handled differently, is reasonable - the opposite extreme with this approach is that no terms of use whatsoever beyond copyright can apply to binaries: no redistribution, modification or *anything* can apply to binaries - and the only condition under which you are permitted to distribute binaries is with the accompanying source or a pointer to the source or a written offer to provide the source.

Modifying the idea of copyright slightly in order to differentiate between the "source" of a work (print plates, XML markup, C code, etc) and the resulting compiled document, and then stating that the compiled documents (binaries, book, etc) are not redistributable without an offer for the accompanying source might just be a practical and useful thing to do in order to ensure that the work ends, eventually (the length of a copyright term is a separate issue), in the public domain where it can be copied and used by anybody for any purpose. Thus fulfilling the purpose of copyright.

There are a few practical difficulties with this scheme - for physical objects like books, for instance, it is difficult and impractical to provide the source. Instead of making a new distinction in the law, however, adding complexity to an already complex implementation of a simple idea, I might suggest that physical access to the source documents be part of the idea - if you want to see the source to a book, or a work, you can physically go down to the building where it is stored, request to see them, and have that request fulfilled.

This allows us to use one reasonably general principle for a large variety of works and have an expectation of a minimum of special cases.

Rob Bos, This e-mail address is being protected from spam bots, you need JavaScript enabled to view it