Hacker, Mr Richard Stallman, contrives to work to help to improve computer software

Hacker, Mr Richard Stallman, contrives to work to help to improve computer software

The Economist

July 15, 1989

Mr Richard Stallman, one of America's best computer programmers, exaggerates only slightly when he calls himself the last survivor of a dead culture. He is one of the few remaining "hackers" (in the original sense of the term: a man devoted to the co-operative enterprise of computing, to the point of excluding normal social and working life). For much of the time he lives in what is little more than a cupboard in the Massachusetts Institute of Technology's artificial-intelligence laboratory, listening to gamelan music, espousing radical causes, but most of all writing programs. To put it mildly, he is no yuppie.

Such men made the computer industry possible. Then it left them behind. Now most programmers write to make their fortunes, and are employed by companies that guard their work to protect it from competitors. They work in an industry that Mr Stallman regards as wicked. Do not scoff yet. He is fighting back for the hacker ethic - and enjoying some success.

His main complaint against the software business is that competition and the desire to maximise revenue breed secrecy, which ends up being bad for consumers. To see how, consider how a program is written. A programmer writes "source code", which is translated into a form that only computers can understand before it is distributed for use. To adapt and improve a program, a user would have to alter the source code. For most commercially available programs, the user cannot do this because the source code is kept secret. If it were not, a user might recreate endless copies of the program and sell them himself or distribute them free. Other software houses could make small changes and market the new product as their own. One consequence is that users cannot iron out problems by themselves; they have to wait for the software house to issue an improved version. Neither can they customise programs to suit themselves.

In the early days of computing, the pioneering hackers who built the foundations of today's programs and systems shared their knowledge. Working mainly in universities, they built on each other's achievements, swapping tips via electronic messages. Their only concern was to improve the programs, for the sake of users. They were not saints; they just loved computers. Mr Stallman's Free Software Foundation works the same way.

He shadows and second-guesses writers of commercial software, creating programs that do the same jobs - which, even his detractors are forced to admit, he is more than clever enough to do. Then he gives them away complete with their source code. His electronic mailbox acts as a clearing house for perpetual improvements. But his programs are not placed in the public domain: if they were, anybody could copy them and then sell them without the source code, thus defeating the object of free software. They are issued under a form of copyright - known as "copyleft" - which obliges users to distribute source code for no more than the cost of reproducing it.

Mr Stallman's main project is to create a software system, known as GNU (which stands for "GNU's Not Unix"), that will be distributed freely to users. In effect it will be a version of the Unix operating system - which is beloved by programmers and scientists, and is now making its way into the business world. So far Mr Stallman has created an editing program, EMACS, based on an earlier (and widely used) version he wrote in the 1970s; a "debugger", which is used for ironing out programming problems; and a compiler, which translates a programmer's source code into instructions a computer can follow. The new NeXT computer, produced by Mr Steven Jobs, one of the founders of Apple Computer, comes with all three.

Several other computer firms in Japan, America and elsewhere are sufficiently interested in what GNU might do for their customers to support the Free Software Foundation. The foundation has just received $100,000 from Hewlett-Packard and the same amount anonymously from a British company. These and other grants have enabled Mr Stallman to hire 12 full-time programmers for the GNU project.

There may be a legal problem for him around the corner. Later this month a judge in San Francisco will hear Apple Computer's claims that Hewlett-Packard and Microsoft, a software company, have infringed its copyrights on "user interfaces" - in effect the look and feel of programs. The interface used by Apple and others, which lets users pick commands on a menu by moving pointers towards them, is such a good and easy idea that it is becoming standard on all sorts of machines. If Apple establishes a right to own the look and feel of a computer product, Mr Stallman may be in trouble. Part of the point of GNU is that its components look and feel like familiar programs.

Hewlett-Packard claims that look and feel can barely be defined, let alone protected by copyright - and are thus rather like, say the layout of pedals in a car. It also claims that the layout is not Apple's to protect. Apple's version was, after all, adapted from earlier work done at Xerox's Palo Alto Research Centre and at universities. Apple seems unlikely to win.

If the law does not stop Mr Stallman, will human nature? Nobody doubts that his co-operative approach is the most efficient way to produce good software - provided enough people join in. The question is whether enough programmers will be prepared to work for much less money than they get now. They would have to exist on grants, university jobs and by providing miscellaneous services for computer users. Mr Stallman cites the fact that there used to be enough willing hackers to build the computer business, so there could be again. If software were free, and there were no riches to be made, the hacker enthusiasts would grow their hair once more. His answer is as logical as it is idealistic.

Still, some developments in the computer industry may work to his advantage. As computers become more widespread and more powerful, their applications become more diverse. This means that programs will get longer and more labyrinthine. It may take large, co-operative groups of informed users to maintain and develop them.

There may be room for compromise. One idea suggested by Mr Richard Greenblatt, a former colleague of Mr Stallman's, is that programs should be sold, but with their source code made available. A coordinating team could then keep modifying programs as better ideas came in. A programmer could be paid according to how many of his lines of code survived the competition and for how long. The administration involved would probably be too burdensome, but the idea may be a step in a promising direction.