The IT Law Wiki


Software-related inventions have enjoyed some degree of protection under the patent system since the beginning of the computer industry. In terms of distinguishing which aspects of software-related inventions could or could not be patented, the courts and the PTO have relied on a number of legal doctrines. Under one of these doctrines, computer program code per se has been held to be ineligible for patent protection because it is a writing that does not fall within one of the enumerated categories of invention. Another of these doctrines provides that processes, including those implemented in software, that are indistinguishable from the steps one would follow in applying a mathematical principle to solve a mathematical problem cannot be patented.[1] These two doctrines served to exclude protection for software-related inventions independent of machines or processes as implemented on a computer.

History of software patents[]

As early as 1964, the PTO denied the patentability of programs, characterizing them as "creations in the area of thought."

In 1966, the PTO attempted to formulate standards for patentability of software. In its first "Guidelines for Examination of Programs," the PTO made recommendations regarding process claims based solely on computer execution of mathematical formulae and process claims called. These recommendations were withdrawn after hearings in which hardware manufacturers opposed the Guidelines on the ground that software users must have access to software to promote the development of the technology. Software developers disagreed, maintaining that software was a technology entitled to protection under the patent laws.[2] These recommendations by the PTO came under attack during Senate hearings on the Patent Reform Act of 1967 for denying software manufacturers rights equal to those of hardware manufacturers.[3]

During later House hearings, former PTO Commissioner Edward J. Brenner indicated that the PTO lacked sufficient pertinent prior art since there was a lack of prior art patents and most of the literature was not in the PTO possession. The Commissioner also noted the difficulty of finding qualified examiners for computer software technology, the lack of experienced personnel, and the inadequate examination procedures in the area of software.[4]

In 1968 the PTO, disregarding the arguments and line drawing of the previous guidelines published in 1966, issued another set of guidelines in which it stated that computer programs, however claimed, were not patentable unless combined with an [[apparatus which caused the physical transformation of matter. The PTO cited examples of transformation such as knitting a pattern or shaping metal. The legal theory on which the PTO based its guidelines was the doctrine of “mental steps" — i.e., if the process could be carried out purely in one’s mind, the invention was not patentable. As discussed below, the C.C.P.A. rejected these guidelines in In re Prater, and the PTO rescinded them.

In 1968, the C.C.P.A. issued a series of decisions testing the limits of the "mental steps" doctrine and the PTO guidelines. Because a programmed computer performed calculations that, in theory could have been carried out in the human mind, it was thought that the mental steps doctrine precluded the patentability of algorithms.

Decisions in the C.C.P.A.[]

The C.C.P.A.'s first decision in this series of cases was In re Prater.[5] The court reversed the Patent Office Board of Appeals which had affirmed the rejection of appellant’s claims to a spectrographic analysis in which linear equations were solved on a programmed general purpose digital computer. In its initial decision the court expressly held that the mental steps doctrine of In re Abrams[6] and In re Yuan[7] did not preclude the patenting of software. The court held that:

[P]atent protection for a process disclosed as being a sequence or combination of steps, capable of performance without human intervention and directed to an industrial technology &mdashs; a “useful art” within the intendment of the Constitution — is not precluded by the mere fact that the process could alternatively be carried out by mental steps.[8]

The court found that in a case where the electronic or mechanical means for performing the process are disclosed in the application and where it is not reasonable to carry out the invention mentally, the invention is patentable subject matter under section 101. Because the applicants had disclosed that a computer, an electronic device, would be used for performing the process, and the process could not reasonably be performed in the mind, the court found that the invention was statutory.

On rehearing, the court sidestepped the 35 U.S.C. §101 statutory subject matter issue, and found that the mental steps doctrine did not apply because the applicants disclosed an apparatus for implementing the process without human intervention. The court, in a footnote, asserted that a general purpose digital computer programmed with a claimed process becomes a special purpose digital computer and could qualify as a patentable invention, assuming the requirements of novelty, utility, and nonobviousness are met. For this reason, the court allowed the apparatus claim. However, the court found that the applicants’ language failed to limit the process claim to its machine implementation and affirmed the patent examiner’s rejection of the specific process claim.[9]

In response to the Prater decision, Commissioner William E. Schuyler withdrew the PTO’s 1968 guidelines, stating:

We now will consider patent applications for computer programs on the basis of the merits for the specific inventions sought to be protected rather than refuse consideration for reasons such as those discarded by the court in the Prater and Wei case.

The C.C.P.A.'s conclusion in Prater was reaffirmed by the C.C.P.A. in In re Bernhart.[10] The permissible scope of subject matter was extended by the C.C.P.A. to include computer program processes in In re Musgrave[11] in which the court rejected the application of the mental steps doctrine and set forth a constitutional "technological arts" standard for determining patentability. The concurring opinion in Musgrave, concluded that "there is very little left of the 'mental steps' doctrine." This view was borne out in subsequent C.C.P.A. decisions.

Decisions in the Federal Circuit and the U.S. Supreme Court[]

The history of efforts to limit software patentability began with Gottschalk v. Benson[12] and ended with AT&T v. Excel.[13] Over the years, the U.S. Supreme Court as well as the Federal Circuit sought to draw lines between software inventions that involved physical transformations and those that represented merely mental steps or mathematical algorithms. Unfortunately, theses lines quickly eroded or had to be abandoned as unworkable.

A series of decisions rendered in 1994 by the Federal Circuit sought to clarify the boundaries of patentable subject matter for software-related inventions. In one decision, the Federal Circuit concluded that an "old" memory that was "reconfigured" through the storage thereon of a "data structure" (an ordered arrangement of information) constituted a patentable invention.[14] In other cases, both before and after this holding, the Federal Circuit concluded that a data structure, per se, and as incorporated into a process without any additional physical elements or steps in the process, did not constitute patentable subject matter.[15] The combined effect of these cases suggested that software could transform unpatentable objects into patentable ones and as such must be given weight in patentability determinations, but information per se and abstract ideas continue to be treated as non-statutory subject matter. The trend — as far as can be ascertained — is to provide a broader eligibility for software aspects of inventions than was available previously.


  1. The courts developed a test — the Freeman-Walter-Abele test — to distinguish claims covering "mathematical algorithms" from those on products and processes that use or rely upon mathematical principles. See In re Freeman, 573 F.2d 1237 (C.C.P.A. 1978), as modified by In re Walter, 618 F.2d 758, 766-68 (C.C.P.A. 1980); In re Abele, 684 F.2d 902, 907 (C.C.P.A. 1982). See also In re Meyer, 688 F.2d 789, 796 (C.C.P.A. 1982). The Patent and Trademark Office promulgated guidelines for interpreting and applying the two-part test for statutory subject matter for inventions involving mathematical algorithms. See 1106 Off. Gaz. Pat. Office 5 (Sept. 5, 1989); 1122 Off. Gaz. Pat. Office 189 (Jan. 1, 1991).
  2. Report of the Hearings on the Patent Office’s Guidelines to Examination of Programs (1966).
  3. Hearings on S. 1042 Before the Subcomm. on Patents, Trademarks and Copyrights of the Senate Comm. on the Judiciary, 90th Cong., 1st Sess. (1967).
  4. Hearings on H.R. 5924, H.R. 13951, and Related Bills Before Subcommittee No. 3 of the House Comm. on the Judiciary, 90th Cong., 1st Sess., Serial No. 11, pt. 1 at 37 (1967).
  5. 415 F.2d 1378 (C.C.P.A. 1968), aff'd on rehearing, 415 F.2d 1393 (C.C.P.A. 1969).
  6. 188 F.2d 165 (C.C.P.A. 1951).
  7. 188 F.2d 377 (C.C.P.A. 1951).
  8. Id. at 1389.
  9. 415 F.2d 1393 (C.C.P.A. 1969).
  10. 417 F.2d 1395 (C.C.P.A. 1969).
  11. 431 F.2d 882 (C.C.P.A. 1970).
  12. 409 U.S. 63 (1972).
  13. 172 F.3d 1352 (Fed. Cir. 1999).
  14. In re Lowry, 32 F.3d 1579 (Fed. Cir. 1994).
  15. In re Trovato, 42 F.3d 1376 (Fed. Cir. 1994); In re Warmerdam, 33 F.3d 1354 (Fed. Cir. 1994).