Software-Related Patents:
A Global Survey
copyright 1997, 1998 Donald M. Cameron , Aird & Berlis
Contents |
|
Introduction
|
Hardware/Software
|
Software-Related Patents:
A Global Survey copyright 1997, 1998 Donald M. Cameron , Aird & Berlis
Contents |
|
Introduction
|
Hardware/Software
|
There is a global market for software; however, the marketplace is sometimes foreign in more ways than one. The laws relating to the protection of software-related inventions are not identical around the world.
There are several major distinctions between the U.S. patent system and those of the rest of the world, which although commonly known are worth repeating in the context of electronic systems and software patents.
The United States is one of the last countries in the world to retain the "first-to-invent" system of priority of invention. All other countries have the "first-to-file" system where the first person to file a patent application is the person entitled to priority over all subsequent applications for the same invention.
Other countries have an "absolute novelty" requirement: the invention must be secret until a patent application is filed, or else no valid patent will issue. Some countries (such as Canada) have a grace period allowing commercialization of the invention before a patent application need be filed.
These countries also "lay open" for public inspection or publish pending application 18 months after the application's filing date which can detrimentally affect the secrecy of an American application which has not yet issued.
Unlike laws of nature, the laws of humans are not uniform throughout the world. Different countries differ on the applicability of certain principles sometimes used to analyse the patentability of a computer software related invention. Those principles are:
1. You can't patent math or science.
2. Computer programs per se are not patentable; computer processes are patentable.
3. If the invention is more than just math or science,
(a) Is it patentable "as a whole"?
(b) Is the only thing new the math or science? [point of novelty]
(c) Ignoring the math or science, is it otherwise patentable? [Otherwise patentable]
(d) and if it's a process, is there a "technical result", that is not just data?
The applicability of these principles, according to the legislation and jurisprudence are set out below.
1. Patentability of Math or Science Patents do not protect science. Patents protect applied science.
In Japan , the application of a scientific principle is a prerequisite to patentability. Statutory subject matter includes as inventions "any highly advanced creation of technical ideas by which a physical law of nature is utilized". n1 If this "cause and effect" relationship is linked by a law of nature, then the subject matter is statutory. If the relationship is a mathematical formula only, then the subject matter is not statutory. Thus, inventions based on human mental activities are non-statutory.
Subsection 27(3) of the Canadian Patent Act provides that "No patent shall issue for ... any mere scientific principle or abstract theorem."
2. Patentability of Computer Programs Per Se A logical extension of the prohibition against algorithms or abstract principles is a prohibition against the patentability of computer programs per se.
3. Patentability of Computer Processes Gerald Goldberg, Group Director of Electrical Computer Group 2300 of the USPTO, the section primarily responsible for examining computer program-related inventions phrased it succinctly when he said:
"Do not claim a computer program...Do claim a computer implemented process." n2
Because Patent Offices are comfortable in allowing patents on processes, then the claims should be drafted to recite traditional subject matter and not prohibited subject matter.
In Diamond v. Diehr (209 USPQ 1, 1981) the U.S. Supreme Court held:
The wholistic approach asks whether the invention is patentable "as a whole". Using this analytical technique, the algorithm or computer program becomes "just another element" in the claim and neither adds to nor detracts from the patentability of the invention.
"...claims must be considered as a whole. It is inappropriate to dissect the claims into old and new elements and then to ignore the presence of the old elements in the analysis."
The European Patent Office Guidelines for Examination provide in section 2, "Inventions" Part C:
"...the Examiner should disregard the form or kind of claim and concentrate on its content in order to identify the real contribution which the subject-matter claimed, considered as a whole, adds to the known art."
2. The Point of Novelty Approach
The point of novelty approach asks whether the only thing new is the math or science. If the algorithm is the only new element in an otherwise old system, then it is the "point of novelty" for the system. The "point of novelty" approach dissects the invention into its old and new parts. This is contrary to examining the invention "as a whole" and is currently viewed with disfavour. Nevertheless, a corollary to this technique is the "otherwise patentable" approach that dissects the patent in a slightly different manner.
3. The "Otherwise patentable" Approach
Perhaps a derivative of the "wholistic" approach, the "otherwise patentable" approach asks whether, ignoring the math or science, the invention is otherwise patentable. This approach segregates the algorithm from the rest of the claimed invention, to determine whether the rest of the invention falls within a traditional category of patentable subject matter.
The "otherwise patentable" test arises from the second part of the Freeman-Walter-Abele n3 . test:
1. Is the algorithm mathematical or non-mathematical? [It is mathematical if it essentially recites, directly or indirectly, a method of computing one or more numbers from a different set of numbers by performing a series of mathematical computations.] Ex parte Logan (20 U.S.P.Q. 1465 at 1468 (U.S. Patent Office Board of Patent Appeals and Interferences)).
2. Is the claimed invention as a whole, no more than the algorithm itself or is it applied to or limited by physical elements or process steps of an otherwise statutory process or apparatus claim? In other words, does the claim wholly preempt the algorithm? Does the algorithm operate on a claimed physical step? In Re Grams (888 F.2d 836).
The product of the claim must not be a pure number. Would the claim otherwise be statutory without the algorithm?
In other words, since the algorithm cannot be "statutory subject matter", the patentable, statutory subject matter must be found elsewhere in the claim.
4. Requirement of a Technical Result
The fundamental requirement for the patentability of computer-related inventions in the European Patent Office is that the invention have a technical character or must provide a technical contribution to the art and not simply a mathematical one. An Application by IBM Corporation under No. 80 107 625.8 (IBM application 80 107 625.8) decision of the European Patent Office dated October 5, 1988 p.7.
Provisions are found in member states' statutes to include this requirement. For example, Article 2585 of the Italian Civil Code provides patent protection to "...the technical application of a scientific principle, provided the latter yields direct industrial results." The exclusions to patentability in the European Patent Convention have in common that they refer to activities which do not aim at any direct technical result but are rather of an abstract and purely intellectual nature. "Decisive is what technical contribution the invention makes to the prior art". 8 Vicom Systems Inc. decision of the Technical Board of Appeal, European Patent Office, Journal of the Europe Patent Office, No. 1, 1987, p.14.
Under Canadian law, the modification of data has been held to be non-patentable . (1981) 56 C.P.R. (2d) 204.
In the EPO the Vicom Systems Inc. decision of the Technical Board of Appeal, European Patent Office, (Journal of the Europe Patent Office, No. 1, 1987, p. 14.) case held that an algorithm that produced numerical results was not patentable. Financial systems and document abstraction inventions are generally disallowed in the United Kingdom .
Hardware elements in a computer system provide the necessary physical and electrical structure for implementing software which has been designed to run on the system. Alpha-numeric keyboards, CRT displays, printers, disc drives, CPU's, and I/0 ports are all examples of well known computer hardware elements which co-operate to bring about a desired result in accordance with executed software.
On the other hand, software forms the instruction set for controlling operation of the hardware elements. These programs can be divided into two classifications: operating systems and applications programs. Operating systems are designed to control and synchronize operation of the various computer hardware elements, while applications programs typically perform predesignated tasks through execution of a series of algorithmic steps which are implemented by the hardware.
Not all software can be said to execute algorithmic steps. Artificial intelligence (Al) software packages are currently available which have the ability to perform tasks in a nondeterministic manner by operating on a database of collected information in accordance with a set of predefined rules.
The distinction between hardware and software becomes somewhat less clear when considering "firmware" elements of a computer. The term firmware is used to designate hardware elements in which software has been stored, (e.g. programmed floppy discs, ROMs, PLAs and gate arrays).
"Knowledge-based" inventions require their own peculiar patent drafting techniques due to their nature.
In Canada, an invention can be claimed in one patent in more than one statutory class: ie. as an apparatus or a process or both. For instance, an algorithm or computer program may be defined by a process claim, or as an apparatus which executes the algorithm or program in "means-plus-function" format, as will be discussed in greater detail below.
Although the Canadian Patent Office has not expressed any preference for one claim format over another, combination-type claims are often best suited to software-related inventions. A combination-type claim is one where the invention is described as a combination of elements or steps which achieve a desired result or interaction.
In some rare instances, the invention may be better defined by either an improvement claim (a claim where only the improvement to an existing apparatus or process is claimed), or European-style claim (one where the preamble describes the known prior art).
A claim for an apparatus should not contain a process step as an element; similarly a claim for a method should not contain an article as an element. Such claims are referred to as ones with "mixed elements". An example of the first type is:
A digital signal processor comprising:
a) a first register for storing signal data,
b) transferring said signal data to a second register...
It is permissible however to incorporate a process limitation in a claim for an apparatus. For example:
A computer comprising:
"means for storing signal data..." or
"a first register for storing data..."
The clause beginning with the word "for" describes the function or process carried out by the apparatus.
Likewise, a claim can recite a process step that contains a hardware or machine limitation. For example:
A method for processing seismic data, comprising the steps of:
collecting the time-varying seismic detector output signals from a plurality of seismic sensors positioned in a cable ... [emphasis added]
In the latter example, the signals are collected from a recited structure. The structure is the hardware or machine limitation within the process step.
Combination-type claims traditionally begin with a preamble which is designed to identify the class of invention and recite any prior art elements with which the invention co-operates.
The invention should not be described in terms of prohibited subject matter but rather in terms with which the Patent Office is comfortable. To avoid a Patent Office subject matter rejection the preamble should not contain such words as "algorithm", "computer program", "software", etc. (8) For instance, where the invention might be characterized as an algorithm or computer program, the preamble should preferably define the class of invention as a "process". Hardware and firmware related inventions can normally be referred to in the preamble by their commonly accepted technical terminology, without fear of provoking a subject matter rejection.
In a combination-type claim, the invention is described as a combination of parts which interact to achieve a certain result or function. In formulating a combination-type claim, functional limitations may be expressed in "means-plus-function" format. The court will interpret the recited "means" as including the preferred means disclosed in the specification as well as functional equivalents, subject to any express limitation in the disclosure.
In order for there to be patentable subject matter, the Canadian Patent Office has held that the end result must fall within a field of activity that would normally be considered patentable if a computer was not involved (i.e. the computer operation must further be integrated into a practical process) (9) . This requirement is similar to the "technical result" requirement of the European Patent Office. Thus, within the body of the claim, there should be a recitation of how the claimed combination affects matter, energy or information, or how the hardware elements interact to provide input and output signals and data.
When the invention relates to data processing, the invention should be described in such a manner that the processed data or results of calculations are not the end result in and of themselves but are instead parameters used within an overall system which satisfies the criteria of statutory subject matter. (10)
After drafting independent claims in one or more statutory classes, the patent agent usually then drafts a number of dependent claims for defining subordinate features to those recited in the independent claims or to further define features and/or structure of the essential claim elements. In the context of a software related invention, the dependent claims may be used to expand sub-routines of a main program.
Under the provisions of the amended Rules for the Canadian Patent Act (11) , a dependent claim may refer to any number of preceding claims, and each immediately preceding claim must refer to no more than one preceding claim.
the description (the disclosure)
It is trite law that the claims of a patent define the subject matter of the invention. Paradoxically however, the Canadian Patent Office refers to the disclosure in determining "what" has been invented, within the two-part Schlumberger test. Thus, careful drafting of the disclosure is important for avoiding a subject matter rejection by the Canadian Patent Office.
According to s. 34(1) of the Canadian Patent Act, the disclosure (or description) must describe the invention and its use as contemplated by the inventor. More particularly, the nature of the invention must be defined, and the manner in which the invention is to be carried out must be described. In the case of a machine (i.e. hardware based invention) the best mode of operation must be described. In the case of a process (e.g. algorithm) the necessary sequence of steps must be explained for distinguishing the invention from the prior art. Moreover, the disclosure must be sufficient to support the claims by describing how to make or use the invention defined in the claims.
The disclosure must also refer to any drawings forming part of the application. Many patent agents begin sketching informal or conceptual precursors to the formal drawings while drafting the claims and then finalize the drawings while drafting the disclosure.
Flow-charts and state diagrams are useful for illustrating software based inventions. Realtime software inventions may require specific timing of events. Thus, timing diagrams may be useful for bridging the gap between the software and surrounding hardware elements.
Some patent agents prefer to include a source code listing of the software as an appendix to the disclosure, for ensuring complete support for the claims. Other practitioners feel that the inclusion of a source code listing should be avoided as redundant when the invention can be adequately described with reference to flow charts and state diagrams.
Where source code listings often contain proprietary information or trade secrets which an applicant may wish to keep confidential, particular consideration must be given to whether it should be included in the disclosure. If certain information is necessary in order to make or use the invention, then it must be included in the disclosure or else the patent will be declared invalid for failing to describe the best mode of operation of the invention or its necessary sequence of steps. (12)
1. Japanese Patent Act, Article 2.
2. AIPLA Bulletin, January-February, 1994 p. 209.
3. In Re Freeman 573 F.2d 1237 (1978); In Re Walter 618 F.2d 758 (1980); In Re Abele 684 F.2d 902 (1982)
8. Manual of Patent Office Practice 12.03.01(g).
9. Re Application for Patent of General Electric Co. (now Patent No. 1,188,775) (1985) 6 C.P.R. (3d) 191, 194.
10. Re Application for Patent for General Electric Co. - supra
Re Application for Patent of Bartley et al (now patent No. 1,167,549) (1985), 3 C.P.R. (3d) 396.
11. Rule 26, amended Patent Rules, C.R.C. 1978, c. 1250, as amended.
12. Ductmate Industries Inc. v. Exanno Products Ltd. (1985) 2 C.P.R. (3d) 289 (F.C.T.D. per Reed J.).
Return to:
Cameron's IT Law : Home Page ; Index
Cameron's Canadian Patent & Trade Secrets Law : Home Page ; Index