¿ìè¶ÌÊÓƵ

Imagine a world in which every company that made typewriters owned its
own design for the keyboard. You would have to keep buying typewriters from
a single company, or learn to type all over again when you switched to a
different model. It would be difficult to change jobs, because other companies
might use a different keyboard. Companies would spend their time devising
more clever keyboards and teaching people to use them, rather than on making
a typewriter that was well-built and reliable. In the confusion, many people
would probably avoid typewriters altogether.

Some computer programmers fear that a similar fate awaits computer software.
Companies have gone to court to prevent competitors from copying the commands
and menus that control what the computer does, and the symbols and patterns
that appear on the screen. In a series of cases in the US, judges have agreed
that if competitors copy these parts of a programme, it violates the copyright
of the author, and amounts to stealing.

This judgment ignores the fact that progress in the development of computer
software has been based on borrowing, insists Mitchell Kapor, founder of
Lotus Development. ‘Ideas slosh around. That’s the way it works,’ says Kraper,
who wrote Lotus 1-2-3, one of the most profitable computer programs in history.
He left Lotus Development in 1988 to start a new company, ON Technology.

Kapor believes that a ban on borrowing techniques and ideas from other
software is good for corporate profits, but eventually will strangle creativity
in the software industry. He is fond of reminding his listeners that the
US Constitution establishes copyrights in order ‘to promote progress in
the arts and sciences’. ‘People regularly forget that,’ he says. ‘They assume
that the prupose of the law is to help them make money.’

Computer programs can be copyrighted, just like poems, books, or sheet
music. Copyright law gives authors of literary works, such as novels and
films, complete control over their creative work. A film maker, for instance,
cannot use the plot from someone else’s novel without permission. Even copying
bits and pieces from a book, or borrowing its overall structure, can be
illegal. Similarly, you cannot simply copy the software you have bought
and distribute it among your friends. The Amrican software industry estimates
that it loses $1 billion each year in the US, and several billions more
abroad, because of software piracy.

Unlike books, however, computer programs are expected to work. There
is no social good to be accomplished by allowing an author to copy a chapter
from Brideshead Revisited when writing a new novel. But there are powerful
reasons why one computer program should adopt approaches and techniques
that are well-tested in other software. Unlike litrary works, computer software
is crucial for the safe operation of railways, hospital equipment and aircraft.
Eliminating bugs in new software is an arduous, never-ending task. The proliferation
of different approaches to solving similar problems in computer software
will result in software that is less well-tested, more bug-riudden and more
dangerous, according to many programmers.

While Kapor and other programmers agree that copying a program in its
entirety is wrong, they feel that copyright law should protect only the
program’s code – the thousands of lines of instructions that tell a computer
how to carry out each command. They say that it should not protect the program’s
effects – the patterns of words and images that the computer displays on
its screen after following the program’s instructions. This ‘user interface’,
which often consists of a set of menus, has the same function as buttons,
levers and dials of a machine: it allows the user to control the computer’s
operation, and see the results of its work. ‘Did I need copyright protection
on the menu structure of Lotus? No,’ says Kapor. ‘Did I need a copyright
on the object code? Absolutely.’

Programmers write programs in one of several languages, such as C or
Lisp. This version of the program, known as source code, is typically kept
secret by software companies. Another program, called a compiler, then translates
the source code into the long strings of binary numbers known as object
code. This version of the program, which you buy on floppy discs, is what
actually controls the computer, turning its electronic circuits on and off
at lightning speed.

It is possible to take the lines of binary code that a company sells
to its customers and ‘decompile’ it, working backwards to the original programming
language, or source code. This process, sometimes called reverse engineering,
is laborious and imperfect, but it can reveal details of how a program really
works. Some companies are trying to ban the practice. Many programmers,
however, insist that decompiling software is legitimite, because it is often
the only way to get detailed information on the state of the art programming.
They say that it should not be illegal to study new software in order to
pick up new programming ideas; they maintain that copyright should prevent
someone only from copying and using whole chunks of reverse engineered code.

Software companies use copyright law to protect the source and object
code and now want it to cover much more – screen displays, data formats
and programming techniques. In a series of cases, courts in the US have
agreed.

In June, a judge ruled that Paperback Software, a small company in Boston,
violated the law by selling a computer program that looked and worked just
like Lotus 1-2-3. This spreadsheet program contains a rectangular grid that
holds tables of numbers, such as business plan. As numbers are entered in
different parts of the grid, the program automatically calculates other
quantities, such as totals and percentages, that the person using the program
may want to know. The judge decided that Paperback software was guilty of
‘overwhelming and pervasive’ copying of one crucial part of Lotus 1-2-3:
its system of menus and commands.

Lotus itself took the the idea of a spreadsheet from an earlier program
called VisiCalc, written by Daniel Bricklin of Software Arts. Lotus developed
a system of commands that made the program much easier to use. The company
sold millions of copies and now dominates the market for spreadsheet programs.

The Lotus decision could mean that all other spreadsheet programs will
be banned from using a similar system of commands and menus to those used
in Lotus 1-2-3: two more court cases will determine whether it does or not.
The day after winning its suit against Paperback, Lotus sued two more software
companies. Borland International and Santa Cruz Operation, which use similar
commands and menus in spreadsheet programs that they sell.

The Lotus case is one of several that have been dubbed ‘look and feel’
suits, because companies are trying to prevent their competitors from selling
similar programs, or ‘software clones’. Apple Computer has sued Microsoft,
another giant among software companies, for imitating the operation of Apple’s
Macintosh computer. Microsoft is selling a program, known as Windows, that
uses menus that roll down from the top of the screen, ‘windows’ that pop
open on the screen, and small symbols such as a dustbin where files go to
when you want to delete them. All of these features were used in the Macintosh.

If Microsoft loses, computer users may no longer find Windows in their
software shops. The court could ban the program, or order Microsoft to pay
high licence fees to Apple, making it unprofitable for Microsoft to sell
Windows.

Ashton-Tate, creator of a database program called dBase, is suing two
other companies for writing software that uses the same programming language
as dBase. A database program organises information into varied lists and
formats. In the past, copyrights have never covered languages, including
programming languages.

In each of these cases, controversy has focused on what a person sees
on a computer screen when he or she uses the machine. Opponents of the idea
that a computer’s user interface can be copyrighted, point to how other
technologies, such as typewriters and cars, treat their interfaces. If it
were illegal for one car manufacturer to copy the layout of a car’s steering
wheel and brake pedals, it would be far more difficult to switch from one
kind of car to another, and the streets would be more dangerous than they
are already.

Just because two programs look alike does not mean that the programs
are identifical, any more than two cars with identical layouts of control
panels and foot pedels will run equally well. What distinguishes a program
that works quickly and reliably from one that does not, is well-written
code: two programmers may choose completely different ways to carry out
the commands that are selected from identical screen displays.

For instance, competitors of Lotus copied the screen presentation of
Lotus 1-2-3, which they thought they could do legitimately, but not the
lines of code. Instead, they tried to write their own program, which they
hoped would work as well as Lotus’s. This proved a tough task. Lotus 1-2-3
remained far more popular than Paperback’s cheaper program, because Lotus’s
programmers had spent many years making their code work as fast and as reliably
as they could. Since copying a popular user interface does not guarantee
commercial success, says Kapor, there is no reason to prevent other companies
from copying it.

Among the combatants in software’s intellectual property battles, there
is a fundamental disagreement over whether exclusive ownership of the ideas
in software is a good thing. Copyright and patent laws are based on the
belief that ownership of intellectual property promotes innovation; if your
competitors could adopt your innovations immediately, you would get no profit
from them and there would be no incentive for you to innovate. Copyright
lawyers are generally so steeped in this logic that they sometimes seem
stunned to find an entrepreneur who does not believe in it. In July, at
a meeting of lawyers and software developers organised by the congressional
Office of Technology Assessment, Kapor admitted under close questioning
by David Allen Goldberg, a copyright lawyer, that the design of Lotus 1-2-3
was an important reason for its success. ‘So how would you protect the design
of that program?’ asked Goldberg. ‘I wouldn’t,’ Koper told a disbelieving,
Goldberg.

Many programmers fear that look-and-feel suits threaten one of the most
honoured traditions in the software business: that is, borrowing good ideas
from other programmers, and using them in their own programs. Apple, which
is challenging Microsoft’s use of its screen presentation, took the idea
of presenting commands as small symbols on the screen from an earlier system
sold by Xerox, called Star. Xerox, in turn, had developed Star by using
research at the Stanford Research Institute in California. ‘I haven’t met
a single programmer who thinks that apple has any right to sue Microsoft
after they essentially borrowed the technology lock, stock and barrel from
Xerox,’ says Andries van Dam, a computer scientist from Brown University,
Rhode Island. According to an impromptu survey of 667 delegates at a conference
in the US in May last year, four out of five specialists oppose copyright
protection for user interfaces.

The legal wrangle has crossed te Atlantic to Europe, which is currently
the world’s fastest growing market for computer software. During the first
quarter of 1990, software companies based in the US, increased their sales
in Europe by 59 per cent over the previous year. Sales of American software
in Britain went up 64 per cent. In Brussels, the European Commission is
trying to draft a directive that would govern the legal status of computer
software.

The computer industry has split into two competing camps, both of which
have established offices in Brussels to push their point of view. One group,
called Software Action Group for Europe (SAGE), represents the large American
companies that dominate the European software market, such as IBM, Microsoft,
Lotus, Apple and Digital Equipment. It wants to give copyright protection
to interfaces, such as the menu structure of Lotus 1-2-3, SAGE would also
like to ban the practice of reverse engineering software – taking the object
code version of programs and working backwards to the original source code.

Without this protection, according to SAGE, EC regulations would ‘deny
software developers the fruits of their investment and innovation’. The
US government threw its support behind SAGE in a letter on 28 February from
Carla Hills, the US Special Trade Representative, to Martin Bangemann, vice-president
of the European Commission.

Opponents say that the SAGE’s proposed regulations would simply make
it harder for any European software companies to compete with the large
American firms. ‘I think it’s just an attempt by the haves to protect themselves
against the have-nots,’ says Kapor.

A competing group, the European Committee for Interoperable Systems
(ECIS), represents Olivetti, Fujitsu, Bull and a collection of smaller firms.
These companies are large enterprises but only have a small share of Europe’s
software market. They support copyright protection of only the source and
object codes of computer software. According to ECIS, copyright law should
not prevent companies from reverse engineering software and from copying
existing interfaces. Without this right, says ECIS, some companies would
face obstacles in writing progams that are compatible with existing software.
This would give other companies a monopoly of the markets, says the committee.

Different directorates within the EC have taken different points of
view, reflecting the disagreement between SAGE and ECIS. Directorate General
3 (DG-3), responsible for regulations governing the new unified European
market, favours SAGE’s point of view. DG-13, which oversees telecommunications,
wants to allow companies to re-engineer software and copy interfaces. So
does DG-4, which regulates competition and monopolies.

ECIS won a victory in July when the European Parliament passed a series
of amendments supporting more open borrowing of elements from computer software.
These recommendations are now before the European Council of Ministers,
the community’s senior legislative power, but the council is unlikely to
make a final decision on whether to adopt them or to come up with its own
recommendations before the end of the year. Either way, its decision is
scheduled to form part of the programme for completing the internal European
market by 1992 and, as such, to be in the form of a directive to member
states. A directive, as opposed to a regulation backed by European legislation,
gives member states a time limit in which to pass legislation nationally
that confirms with teh council’s wishes.

The division between large and small companies reflects a typical split
between lawyers and programmers. ‘Programmers are just getting used to this
predatory way of operating, and many are very offended by it,’ says van
Dam. But when small start-up ventures grow into huge companies selling millions
of copies of software, as Lotus did, programmers inevitably lose control
to professional managers and their legal advisers. These executives generally
want to establish strong control over the company’s intellectual property.
In the process, lawyers ‘try to fit software, which they don’t really understand,
into a system of words and precedents that is ill-suited to the technology,’
says Bricklin.

Copyright law treats software like a book, or other literary works,
even though, under US law, copyrights cannot cover a ‘procedure, process
or system’, which are all terms commonly used to describe elements of software.
Anthony Clapes, one of IBM’s senior lawyers, argues that programmers are
‘binary bards’ whose creative work is equal to that of any poet, Just as
the symbols on a page of sheet music control the performance of a symphony,
the lines of computer code direct the ‘silicon symphony’ of a computer’s
operation. Those who want to treat software differently from literature
or music betray an ‘antipathy to new technology,’ says Clapes.

But software is also like a machine, and can be covered by patents.
The patenting of computer software has become most widespread in the US,
though patents have also been awarded in Europe and Japan for computer programs
used in industrial processes. Because software has both literary and machine-like
qualities, the courts are confused over which precedents should apply.

Hundreds of patents are now being granted each year for programming
techniques, or algorithms, that the US Patent Office decides are new and
useful inventions. But programmers are in an uproar about most of them.
‘Things that we would regard as totally obvious are getting patents,’ syas
Randall Davis, director of the Artificial Intelligence Laboratory at the
Massachusetts Institute of Technology. Patents are only supposed to be awarded
for new inventions that genuinely advance the existing state of the art.
Van Dam says that most of the patents he has seen are ‘bogus’.

The basic problem seems to be that the patent system cannot cope with
the pace of developments in software. The US Patent Office currently takes
nearly two years to process an application for a patent on a computer algorithm.
Patent examiners also have great difficulty staying abreast of developments,
because most of the software is not openly published in the form of source
code. In the time that it takes the patent office to consider an application
for a patent on a particular programming technique, hundreds of programmers
around the world may have independently created something similar, thus
infringing the patent. ‘We throw away more ‘inventions’ each week than other
people develop in a year,’ says Richard Stallman, founder of the Free Software
Foundation .

According to van Dam, patents are already beginning to interfere with
software development. In one case, a company was awarded a patent on a technique
that allows the computer to display a three-dimensional object from different
viewpoints, even though the technique was developed 10 years before the
University of Utah. The patent-holder successfully threatened another company,
Raster Technologies, with a lawsuit for patent infringement. Rather than
spend time and money in court contesting the patent, Raster Technologies
paid a licence fee to the patent-holder.

One new software company recently started operations with 15 programmers
and two lawyers, notes van Dam. ‘In the old days, I don’t think people would
have started off with lawyers,’ he says. ‘It’s stopping information flow
and wasting resources.’

* * *

A crusade for free software, and a new statute of liberty

For Richard Stallman, writing software is a crusade for truth, justice
and liberty. Arrayed against him and his band of intrepid programmers are
the world’s major software corporations, a capitalist legal system and the
profit motive.

Stallman, 37, is the founder and spiritual inspiration of the Free Software
Foundation, an organisation that campaigns against all copyright protection
for programs. One of MIT’s pre-eminent ‘hackers’ during the 1970s, Stallman
looks the part, stroking his long black hair in a preoccupied gesture and
turning an intense, yet awkward gaze on his questioner. He lives in a small
room in MIT’s Artificial Intelligence Laboratory, where a visitor before
noon is likely to waken him from daytime slumber.

He has no official position at the university, but few people at MIT
begrudge him the space. He is regarded as one of the most brilliant programmers
MIT has produced, and his software comes with no strings attached. Stallman
cut his formal ties to the university in 1984 to prevent it from owning
any of his work. The same, year, with a network of other programmers, he
set up the foundation.

The foundation currently distributes a text editor called GNU Emacs,
which is widely used to edit lines of programming code or text, and a C
compiler, which converts source code written in the C language into binary
code that can run on computers. Several other programs are in the pipeline.

Free software does not necessarily mean that it costs no money: Stallman
charges enough to cover his costs. For him, the software is free in a more
important sense: its users are encouraged to copy it, change it and distribute
versions of it to others. In a snub to the computer establishment, Stallman’s
programs come with only one condition: no one may place any legal restrictions
on software derived from them.

In an effort to keep the structures of their programs secret, most software
companies sell programs in the form of object code, which is the string
of binary numbers that only a computer can readily understand. ‘That’s a
disaster that has already happened,’ says Stallman. ‘Students have no access
to large, good software.’ The Free Software Foundation lays these secrets
bare, providing the source code that its programmers write.

According to Stallman, software is a pattern, not an object; it should
be considered information, not a product, and there should be no restrictions
on its free flow. For him, free software is a symbol of free society: a
poster in his room shows the Statue of Liberty holding aloft a floppy disc.
An ‘avalanche of greed’ in the software industry, says Stallman, is choking
off the free flow of information. Companies have been systematically trying
to give themselves a monopoly over their software, first by keeping their
source code secret, then by asserting ownership of their software’s ‘look
and feel’. In protest, Stallman has organised picket lines in front of the
headquarters of Apple Computer and Lotus Development.

Stallman has supported his volunteer work for the foundation by working
a few months each year as a consultant for software companies. His skills
reportedly earn him $260 an hour. He may not need the work anymore, however.
In July 1990, the McArthur Foundation in Chicago, a charity that periodically
awards grants to ‘highly talented individuals’ to do what they like with,
gave Stallman $240,000.

According to Stallman, computer systems, like society, should be run
whenever possible on the basis of trust and openness. Around MIT, it was
common knowledge for years that Stallman’s user identity on the university’s
computer network was the initials of his name, ‘rms’, and his password was
also ‘rms’. Stallman refused to keep his password secret, allowing anyone
to log on to the computer under his identity. As a result, in January, MIT
removed him from the network. He is unrepentant: ‘I’d rather live with the
occasional vandal than the police state.’