Lapas attēli
PDF
ePub

created a very large retail market for application software--for word processing, spreadsheets, even games--as well as a lucrative market for PC operating-system software. In 1988, domestic revenues for PC software reached nearly $3 billion. 10 The widespread use of PCs also facilitated the growing use of online databases; domestic revenues for on-line business databases alone amounted to $6.5 billion in 1988.11

Rapid growth and technological innovation made markets for PCs and PC software quite volatile, even compared to the mainframe and minicomputer markets a decade earlier. New hardware and software firms would introduce products, enjoy brief success, and go out of business within the space of a Other firms built on initial successes and went on to become

few years.

industry leaders. A few years after introducing a successful product, however, they might find that a substantial fraction of their market was being taken by competitors offering similar--sometimes improved--products, often at a lower price. This phenomenon, and the volatility of PC markets, has brought new attention to questions about how best to provide intellectual-property protection for software. At the same time, they illustrate the complex relationship between intellectual-property protection and stimulation of creativity.

10 Ann Stephens, Software Publishers Association (SPA), personal
communication, Oct. 2, 1989.

11 Information Industry Association (IIA) data, 1989.

The Intellectual Property Bargain and Software

In the United States, an "intellectual property bargain" underlies

intellectual-property protection.

The bargain between creators and society

sets a balance between two social objectives: 1) encouraging the production and dissemination of new works and inventions (by providing economic incentives to creators) and 2) encouraging widespread access to and use of

12

these works and inventions. Thus, the limited monopoly granted to authors via copyright and to inventors via patent is a quid-pro-quo arrangement to serve the public interest, rather than a system established only to guarantee

income to creators.

Copyright is granted in exchange for publishing the work so that it is disseminated. But copyright, which is relatively easy to obtain and longlasting compared to a patent, is intended only to protect the expression in a work from unauthorized copying, not to protect the underlying ideas from independent creation or use.

CONTU's recommendation that full copyright protection be extended to all forms of computer software was implemented in the 1980 amendments to the Copyright Act. Even then, CONTU recognized certain inherent difficulties in applying copyright (which does not protect ideas, processes, or procedures) to software. A special concern was the impossibility of establishing a precise line between the copyrightable "expression" in a program and the

12 See U. S. Constitution, art. I, sec. 8, cl. 8. For a discussion of
this bargain and the public interest in intellectual property
protection, see United States Congress, Office of Technology Assessment,
Intellectual Property Rights in an Age of Electronics and Information
(April 1986) and Copyright and Home Copying: Technology Challenges the
Law (October 1989).

noncopyrightable processes it implements--the distinction between expression and "idea." The 1986 OTA report, Intellectual Property Rights in an Age of Electronics and Information, discussed the difficulty of applying copyright to functional works such as programs. 13 This is a major issue today, as evidenced by ongoing "look and feel" and "structure, sequence and organization" copyright suits. Moreover, market changes like the almosthundredfold increase in PC use since CONTU make the financial stakes much

higher.

14

Patent protection, which is stronger than copyright in that ideas are protected (even from independent invention), is of shorter duration. But th the requirement for patentability is stiff: the invention must be useful, novel,¡ and nonobvious compared to prior discoveries (the "prior art") that are patented, in he public domain, or otherwise widely known. Moreover, patent is granted in exchange for full disclosure of the best way of implementing or practicing the invention--the purpose of the patent is to "teach" others and thereby stimulate technological progress as they seek to build on (or invent around) the discovery. Patent protection was generally considered "not

applicable" to software until the early 1980s. Since 1981, there has been

13 See OTA-CIT-302, April 1986, pp. 78-85.

14 Some commentators believe that over 90 percent of computer programs
would not meet the patent requirement that the invention be nonobvious,
compared to the prior art; some think that only perhaps 1 percent of all
software is patentable under the current requirements. (Findings of ABA
Proprietary Rights in Software Committee (1983), cited in Cary H.
Sherman, Hamish R. Sandison, and Marc D. Guren, Computer Software
Protection Law (Washington, DC: The Bureau of National Affairs, Inc.,
1989), p. 401-8 and note 41.)

15

renewed interest in patents for software and related inventions.1 Over the past 7 years, patents have issued for software-related inventions such as a spell-checking routine, logic-ordering operations for spreadsheet programs, a brokerage cash-management system, and a bank college-savings system.

In the last year, some highly visible software-patent lawsuits and controversies concerning patents for algorithms have served as targets for more general concerns over the appropriateness of patents for software and algorithms. These arise both from fears of stifling technological progress, as well as from practical problems software and algorithms raise for patent. system administration. One of these problems is the relative paucity of "prior art" for patent examiners to use in evaluating patent applications for software and algorithmic inventions.16 Another problem is the lack of special classification for or cross-references to issued patents. As a result, it is virtually impossible to even find, let alone count or profile, all softwarerelated or algorithmic patents. This means that patent examiners and the

15 A 1981 Supreme Court decision (Diamond v. Diehr, 450 U.S. 175 (1981))
helped clear the way for patent protection for some software-related
inventions. Current examining procedures are intended to preclude
issuing patents for "bare" programs or algorithms per se. See Gerald
Goldberg, "Patentable Subject Matter: Mathematical Algorithms and
Computer Programs," Patent Protection for Computer Software: The New
Safeguard, Michael S. Keplinger and Ronald S. Laurie, Eds. (Englewood
Cliffs, NJ: Prentice Hall Law and Business, 1989), pp. 9-42.
16 Because software patenting is relatively new, because copyrighted
software deposited at the Copyright Office is not indexed or readily
searchable, and because trade secrecy has been a major form of software
protection, the prior art readily available to patent examiners and
applicants is spotty. This potentially allows patents to issue for
"inventions" that are already known in the industry or that represent
only minor improvements.

public have no effective way of searching for and studying them. Another concern is that the time lag between patent application and issuance is relatively long compared to software life cycles.

Trade secret protection, provided under individual State laws, protects against use or disclosure of the secret by others (but not against accidental disclosure or independent creation). Most foreign nations outside of Western Europe do not recognize trade secrets, either domestically or internationally. Trade secret has been the traditional favorite of the industry to protect software and its underlying ideas, logic and structure. Unlike copyright or patent, there is no limitation on its duration. When software is protected by

trade secret, information remains undisclosed for as long as trade-secret

17

status is maintained. For society, this can lead to a lack of knowledge

about the state-of-the-art.

In turn, this can adversely affect patent

examinations and lead to "re-inventing the wheel" rather than building on (or

around) prior advance.

Controversies Over Software Protection

Legal protection for computer hardware was traditionally provided by patent or trade secret; this combination served fairly well to protect major hardware advances, as well as more-incremental developments.

Because

protection for computer software did not fit so neatly within the traditional forms of intellectual property, the process by which software developers and

17 By contrast, patent rights are granted in exchange for full
disclosure of the "best method" of practicing the invention.

« iepriekšējāTurpināt »