OTD protects academic principles and independence while advancing discoveries toward commercial development.

A Guide to Patent Law

Identifying Patentable Inventions

For an idea to be patentable, it must consist of statutory subject matter. United States patent law permits the granting of a patent on the following statutory subject matter:
  • A process, such as a method of applying a vapor barrier for silicon materials.
  • A machine, such as a new instrument to deposit uniform layers of metallic compounds.
  • An article of manufacture, such as an assay kit for an infectious disease or class of diseases.
  • A composition of matter, such as a new molecule (characterized by amino acid sequence or base-pairs) or a new chemical compound.
  • New and useful improvements of the above.
  • Any distinct and new variety of plant which is asexually reproduced.
  • Any new, original, and ornamental design for an article of manufacture.

If an idea fits into one of the previous categories, United States patent law also requires that the invention be: 1. New or Novel: The invention must be demonstrably different from publicly available ideas, inventions, or products (so-called "prior art"). This does not mean that every aspect of an invention must be novel. For example, new uses of known processes, machines, compositions of matter and materials are patentable. Incremental improvements on known processes may also be patentable. 2. Useful: The invention must have some application or utility or be an improvement over existing products and/or techniques. 3. Non-Obvious: The invention cannot be obvious to a person of "ordinary skill" in the field; non-obviousness usually is demonstrated by showing that practicing the invention yields surprising, unexpected results.

Each of these three criteria is open to interpretation. Establishing novelty and/or usefulness often relies on arbitrary value judgments. For example, it may not be clear to the United States Patent and Trademark Office (USPTO) that a new gene-sequencing instrument is "demonstrably different" from other existing models, or that the use of a new hybridoma for narrow research purposes is sufficiently significant to be called a "useful" improvement over existing techniques.

Despite the possible variety of definitions of novelty and usefulness, the concept of "non-obviousness" is the most complex of these three patentability requirements. It is subject to broad and often inexact interpretation. For example, it might be argued that a new method of controlling protein production in bacteria is obvious in the face of prior art because it relies on a collection of well-known, existing and proven concepts. Conversely, one could argue the same method is not obvious because certain specific elements of the method yield surprising, unexpected results. Judging what is obvious to one of "ordinary skill" in an art is rarely straightforward, especially in technologically complex and rapidly changing fields.

The three criteria of novelty, usefulness and non-obviousness generally are the focal points of the USPTO's evaluation of patent applications. Obviousness is the reason most frequently cited by patent examiners as to why an invention is not patentable.

Finally, patent law states that inventions may be patented if they have been reduced to practice, even if a physical embodiment of the invention has not been realized ("constructive" reduction to practice, as described in the preceding section). Therefore, some extrapolation about an inventive concept can be included in a patent application.

The extent of permissible extrapolation, however, can be the subject of dispute with the USPTO. For example, an invention which claims a broad class of chemical compounds is rarely granted a patent, unless evidence is presented that several different representative members of the class have actually been made. A claim even to a single new compound usually must include a detailed description of its actual synthesis in order to receive patent protection. In the case of machines or mechanical devices, the requirements of the USPTO are much less strict.

Conception of Invention

For patent purposes, inventions are made in two stages:
  1. "Conception," which is the intellectual act of creating a complete inventive concept, including ideas for making the invention and methods of using it.
  2. "Reduction to practice" of the inventive concept. This is defined either as "constructive" reduction to practice (patent application), even though physical embodiment of the invention may not have been realized, or "actual" reduction to practice which involves making a physical embodiment of an invention (for example, a prototype), and using it successfully for its intended purpose.

For example, a scientist might claim to have invented a novel surgical instrument or a new environmental monitor and a patent application could be filed on the invention even though the device had never actually been built. In this example, the conception would be the intellectual concept of the instrument and its use for medical or environmental applications. The reduction to practice would be a constructive reduction to practice achieved by the filing of a patent application describing the instrument and its intended use. (If the invention had been built and used, actual reduction to practice would have been achieved.)

For chemical and bio-pharmaceutical products, however, it is usually not possible to obtain patent protection unless at least one example of the product has actually been made. If the invention claimed is broad enough to cover a class of products, it will be necessary to have made a representative number of the products of that class. In general, it has not been necessary to have proof of the product's usefulness if the claim of usefulness is plausible.

In order to obtain a patent, all inventors, through their patent attorneys, ultimately have to convince the U.S. Patent and Trademark Office (USPTO), and its foreign counterparts, that their invention meets the requirements of patent law summarized in the next section.

Patenting of Computer Software

For software-related invention(s), if the invention is integral to a particular task or process, it may be considered as patentable by the USPTO. The USPTO, however, will not grant a patent on a pure scientific discovery, or a method of doing business, or a process which does not require hardware, or one that simply duplicates the "mental steps" of a human being (such as a purely mathematical algorithm). The distinction between "statutory" (patentable) and "non-statutory" (unpatentable) subject matter continues to be a topic of debate among software developers, academics, lawyers, and USPTO examiners and is currently determined by case law rather than by hard and fastrules.

The courts, however, have been ambiguous in defining what exactly constitutes "patentable subject matter," the first hurdle to be cleared in obtaining a patent. In response to this confusion, the USPTO recently (in March 1996) issued new Examination Guidelines for Computer Related Inventions. The new guidelines had two general objectives: to assist USPTO examiners in determining which categories of software are patentable, and to help inventors submit more successfully crafted patent applications.

The guidelines set up three permissible categories for patentable subject matter involving software inventions. First, a computer or other programmable apparatus whose actions are directed by a computer program or other form of software may be considered a "machine." Second, a computer-readable memory that can be used to direct a computer to function in a particular manner when used by the computer is presumed to be an "article of manufacture." Third, a series of specific operational steps to be performed on, or with the aid, of a computer is deemed a "process."

In addition to satisfying the criterion for statutory subject matter, a software-implemented invention must also meet the patentability criteria of novelty, utility and non-obviousness. One advantage of computer software inventions is that they may be protected in a number of ways which are not necessarily mutually exclusive. For example, when an aspect of a software-related invention which is protectable by patenting, is reduced to code, the code itself may be further protected by copyright.

You may also wish to refer to the Computer Software page on this site.

Publication and Patent Rights

The University is committed to an academic policy ensuring that faculty members can disseminate the results of their research as they wish. In keeping with this commitment, the University has adopted the intellectual property policy and procedures, that encourage the faculty's customary manner of disseminating research results, while providing the means to preserve rights to intellectual property.

In terms of patent law, publication or public disclosure of an invention means the non-confidential transfer of knowledge--orally or in writing, by exhibits, demonstrations, public use, or sale of the invention. Of course, journal articles and catalogued theses are considered publications. Abstracts, typewritten papers, slides and projected material, which are distributed or discussed at non-confidential meetings, conferences, seminars, or forums, are also examples of what may constitute publication or public disclosure. Finally, grant applications once awarded are put into the public domain and may also constitute a public disclosure. Electronic transmission of abstracts, articles or research reports is also a form of publication or public disclosure. Scientists should be aware that many journals and scientific societies often place material on the World Wide Web prior to written publication, creating an increased potential for loss of patent rights.

The key factors which determine whether disclosure of scientific data represents a public disclosure which can be construed as prior art include the nature of the disclosure (for example, a lab talk or formal presentation), and whether what was disclosed "taught one of normal skill in the art" to carry out the invention. In cases of uncertainty, it is usually best to seek the advice of patent counsel through Harvard's technology transfer offices.

In the United States, patent applications may be filed up to one year after an invention's first public disclosure. In contrast, most foreign countries require that patent applications be filed before any public disclosure of an invention. Despite these strict rules, if a U.S. patent application is filed before any public disclosure, most foreign countries permit filing of corresponding applications for up to one year after the U.S. filing date, even if a disclosure of the invention was made after the U.S. filing. Therefore, a U.S. patent application filing prior to any publication or oral disclosure preserves both U.S. and foreign patent rights.

Patent Prosecution Process

Whether or not an invention has been constructively or actually reduced to practice, the process through which the USPTO ultimately awards patents is rarely simple or straightforward. It is a negotiating process between the inventor (represented by an expert patent attorney) and the USPTO, usually centering on whether or not the invention is "non-obvious."

While every case is different, the process has two phases. First, the patent application is submitted on behalf of the inventor and the University by an outside patent counsel. It consists of:

  • an abstract,
  • the "specification," a text section which explains the invention's history, describes it in broad terms and teaches the public how to make and use it, and
  • the "claims," which are precise, numbered statements describing exactly the products or processes that would be protected by the patent. The claims of a patent are the essence of an invention. They are a written boundary describing exactly what part of the practical arts the holder of the patent has the right to forbid others from practicing and, as such, inventors should pay special attention to the crafting of their patent's claims to ensure that they obtain the broadest protection possible for their invention.

Second, in approximately twelve to eighteen months, the USPTO issues its first "office action," or reply to the initial patent application. This first office action usually rejects most or all of the inventor's claims of invention on the grounds of lack of novelty or obviousness; several references (either issued patents or scientific papers) are cited to support this contention. The inventor and patent attorney then must argue that the references do not render the invention obvious. This basic give-and-take continues until a patent is either granted or denied; the entire process can take anywhere from a year and a half to ten years. The total cost of obtaining a U.S. patent through these steps varies from $10,000 to $25,000. An application for a patent in other countries follows a similar procedure; obtaining foreign patents costs approximately $10,000 to $20,000 per country.