Brave GNU World

(jpeg 7k)

Welcome to another issue of Georg's Brave GNU World. As usual I will start rather practically and then slowly advance towards theory.

Sawmill

Sawmill [5] is a window manager by John Harper that is extensible via its own scripting language in a similar way to GNU Emacs. This language was originally very close to Emacs Lisp but by now it has incorporated some Scheme elements and has become a mixture of the two.

To remain fast and lean despite all this extensibility, Sawmill does not include anything that could be done by other programs like managing the background or "application docks." Its only purpose is to manage windows in the best, most flexible and most attractive way possible.

The program itself is a primitive C-core; all higher functions are implemented in terms of the script language which allows rewriting any part according to personal preference. Personally I consider it very important that any function can be bound to any key with regard to the context (root window, inside specific window, window title and so on) which is usually determined by the mouse focus.

Of course Sawmill is also themeable [6] and comes with the imitation of some well-known Enlightenment themes. Combined with the good GTK+ and GNOME integration, all of this gives Sawmill the potential to become the "de-facto" standard window manager for GNOME.

Despite all this functionality "pure users" should not feel intimidated because pretty much everything can be customized with the help of a graphical configuration tool.

According to John Harper his future plans include KDE support and getting rid of the remaining bugs. His other objective is to prevent a "feature creep" to keep Sawmill fast, lean and efficient.

I'll continue with another project that is just about to begin.

Free Unix Benchmark Project

Stefan Carstens told me that he initiated the Free Unix Benchmark Project. Currently the available benchmarks either only measure small subunits of a system, ignore important Unix-specific tasks, or are proprietary and as such are not transparent. Another problem is that they are scattered all over the net; there is no central node.

The goal is now to create a complete benchmark suite that will allow running all tests from a single package. Including a benchmark for networking parts like the TCP/IP stack is also planned. Since such a test is not available yet, the project is guaranteed to raise interest pretty early. The choice of license will also work towards the fast distribution because it is intended to rely on the GNU General Public License and GNU Lesser General Public License. Due to the visible sourcecode this also makes the tests reproduceable and transparent which is an important point for benchmarks.

For the benefit of the users the results will be given in simple and comparable units so a "normal" user can also work with it. Different parts of the system can be profiled and compared via a kind of "horsepower." An output backend to gnuplot is already planned and I am quite sure that it will also be no problem to implement direct output into a database.

If you're now eager to grab this project and install it right now I'll have to keep you waiting for a while. But if you are interested to be part of this interesting project, I can only tell you to send mail to Stefan Carstens [7]; he was very clear about his invitation for participants.

Now these topics are covered, I would like to come to a part that is rather important for me personally.

Informational Human Rights - Part II

After my declaration of "Informational Human Rights" in issue 8 of the Brave GNU World I got a big lot of very positive feedback. One of the people contacting me was Ken Engel who sent some interesting amendments that I would like to share with you.

His first amendment to the Informational Human Rights says: "The freedom to reverse engineer shall not be abridged." His example was that it completely legal to disassemble a car and then put it back together. You may even fix it. The same is illegal with proprietary software. Doing such a thing can cost you a lot of money or even get you into jail.

The second expansion is based on experience he gathered when being forced to write an application with the help of a proprietary development environment. Due to the fact that the development environment is proprietary, his data is now locked in an opaque binary format. The only way to it is a very complicated and slow graphical interface. Hence his second amendment to the Informational Human Rights states: "Customers' and users' data shall not be locked up in the vendors product."

The last point goes against the "sweat shop" mentality and the fact that very often developers are forbidden to use the best tool for the job: "Choice of software shall not be dictated by employers who disregard their staff's expertise, thereby limiting their choice of development practices and platforms, and impeding their productivity."

I am quite sure that a lot of you will be able to identify yourself with at least one of these amendments.

Now I would like to come to the promised second part of my "Systems `99" recollection.

The Sun Community Source License

About a week before I went to the Systems, Marco Boerries, the founder of Star Division, gave an interview in a British magazine. In this interview he said that he didn't understand why the SCSL is being viewed so negatively because he considers it far superior to the GPL.

To prove this statement he explained that a customer expects a guarantee about the property rights of the code he buys and the GPL cannot deliver this. From the viewpoint of an old fashioned producing and selling business this may be a very normal statement, but in this context it shows a lack of understanding for the Free Software business models.

His first misunderstanding is that he assumes the customer buys software as his "property." Should the code be stolen the customer could then lose his money. Now launching into an explanation why the theft of proprietary sourcecode (Free Software cannot be stolen) in order to use it in Free Software makes no sense would distract from the point I was about to make. So I will allow myself to ignore it for now - should you be interested I may supply it later.

This time I would like to focus on the question why the "property model" does not make sense when talking about software. To do so I will have to delve into some basics, though.

The old-fashioned business model is to design a product, produce it and then sell it. Mass production leaves very little room for individual taste. Otherwise the costs would explode. This is why products are designed to fit as well as possible to some average person that very often doesn't even exist in reality. Afterwards it is the job of the marketing division to convince all people that this product is made for them especially and they should buy it because it'll make them happy.

This "real life" business model has been transferred without modification into the software industry. They create piles of bits and bytes that are being sold to the customer - suggesting that just the possession of this pile will solve all the problems. The actual reason for this concept, the machinated mass production, does not exist in the virtual space of software, though. Software can be custom-fitted to the customer in a way no other product can. Trapped in old patterns of behaviour they have been applied although the necessity vanished long ago.

For this reason business models for Free Software concentrate on problem solution instead of declaring possession the solution. And that is why the discussion of property rights is irrelevant. For the customer it does not matter that software is his property, only that it solves his/her problems.

In fact it is preferable to the customer if the software is not owned by anyone - neither him nor the company that he got it from. The value of software is very much determined by the amount of maintenance that goes into it - without this service software loses its use very quickly. With proprietary software a user rarely buys the software itself but the right to use a very special version. But of course this version is static and its use deteriorates fast. The customer is required a buy another version soon - even if it developed into a direction that does not help the user or his/her needs.

Additionally the customer enters a rather dangerous dependency. By choosing a certain proprietary product the success of the customers business is linked to success of the software provider - even this is not obvious in a lot of cases. Only Free Software guarantees the customer the security of his/her own success by making the success independent from the software vendor. Additionally the development of software can be influenced in several ways in order to make it fit for future needs.

And that brings me to the second point. Marco Boerries claimed the reason for disliking the SCSL was the fact that it makes Sun successful and rich. This is not true. The whole Free Software scene is very happy about the financial success of businesses like Red Hat that use the GNU General Public License on a big scale.

The SCSL's problem lies somewhere else. Even if the sourcecode is visible, it is still forbidden to change it and then pass it along. All changes have to be sent to Sun and only Sun decides whether these changes will be implemented; in the end all rights remain with Sun. What Sun managed to create this way is a kind of "proprietary Open Source license." The most important reason for the use and success of Free Software, the freedom, has not been given to the user. This is why we dislike the SCSL just like any other proprietary license.

Okay, that's it for this month. As usual I hope to receive a lot of feedback in form of ideas, comments and questions - you can send them by email to the well-known address [1].

Return to previous issue / Brave GNU World home page

Return to GNU's home page.

Please send FSF & GNU inquiries & questions to gnu@gnu.org. There are also other ways to contact the FSF.

Please send comments on the Brave GNU World column to column@gnu.org, send comments on these web pages to webmasters@www.gnu.org, and send other questions to gnu@gnu.org.

Copyright (C) 1999 Georg C. F. Greve, German version published in the Linux-Magazin

Permission is granted to make and distribute verbatim copies of this transcript as long as the copyright and this permission notice appear.

Last modified: Wed Feb 9 22:40:39 CET 2000 greve