by Douglas J. Malewicki

I first contacted Patent Attorney Patrick D. Kelly after reading this superb article that he authored for the Fall 2002 issue of Tau Beta Pi's THE BENT magazine (shown to the right).  Tau Beta Pi is a National Engineering Honor Society.  I am a  life member going back to my days at the University of Illinois. (MMM - must remember to add that accomplishment to my resume!)

This article and Pat's other article about Trademarks are MUST READS for all aspiring inventors.  You will find that studying these articles thoroughly will require time and focus.  Do the work - the valuable information you gain will save you time and money. 


by Patrick D. Kelly, Patent Attorney, St. Louis, 314-828-8558, FAX: 314-828-8998     America Online users:
Copyright © 2002, by Patrick Kelly. All rights reserved.


Patrick D. Kelly, received an engineering sciences degree from the University of Texas at Austin in 1974 and a juris doctor from Harvard Law School in 1980. He has practiced patent law since 1980 and specializes in biochemistry and medical technology.

Mr. Kelly worked for the EXXON Company in 1974-77 and the Monsanto Company in 1982-85. He has written a guide to trademarks and trademark law that was published in the Spring and Summer 1998 issues of THE BENT. In 1994, he was a nominee for the U.S. Congress. He can be reached at

      This article is for inventors who want to be actively involved in drafting patent applications on their inventions. Most inventors want to make sure they're still a full partner on the team when a patent attorney takes over and starts writing the application. Some inventors go beyond that and want to write either or both of the following: (i) a "provisional" patent application, as discussed below; and/or, (ii) the first draft of a "utility" patent application.

      Provisional patent applications are discussed in a separate paper by this same author; that paper, and an example of a provisional application, is also reprinted in these conference proceedings. In general, a well-written provisional application should contain the same organization and content as a utility application, as discussed below. The two most important differences between provisional and utility applications are (i) a provisional application has much more leeway to include “informal” drawings, which don’t need to use “callout numbers” at all; and (ii) a provisional application is not required to contain any claims. The question of whether a provisional application should contain claims  is discussed below, after some basic rules of claim drafting are explained.

      Regardless of whether an inventor writes a complete provisional application (or a complete first draft of a utility application), or hands that task over to a patent attorney, any inventor can benefit greatly if he/she knows what to expect from a patent attorney, how the information should be organized, what the specific goals are, and how to tell good work from not-so-good work. The goal of this article is to share that information with inventors, so they can work smoothly, quickly, and effectively with a patent attorney, regardless of how the roles of inventor and attorney are adjusted and balanced, once they begin working together. 

General  Comment: Know Your Audience(s)

      Over the course of my life, I’ve learned that when someone tells me something which appears to be about me, I should generally assume, “Half of what he’s saying is about me. But the other half is about him, not me. Part of my job is to figure out which is which.”

      In the same way, whenever you read a patent, you should first ask who owns the patent, and who wrote it, because that will say a lot about how it was written.

      If a patent was written by patent attorneys working for a big company facing ferocious competition, the most likely goal of that patent is to cover and protect investments that have already been made. There’s nothing wrong with that, but keep in mind, that’s not the same problem that has to be faced by independent inventor who is trying to persuade a company or team of investors to invest in the idea).

      For better or worse, part of the goal of nearly any patent obtained by a billion-dollar company facing tough competition will be to make it hard for competitors to figure out what actually happened in the company’s research labs; so, a patent like that will not readily disclose or highlight where the really valuable and crucial “nuggets” are, in that mass of verbiage. Various tactics are used by patent attorneys in those situations. One tactic involves using peculiar, awkward, self-selected terminology, either to raise questions about what was really done or why, or to avoid admitting that some component is merely commonplace and obvious (my favorite was a patent that created and used the term “liquid motors”, to avoid admitting that they were just ordinary “pumps”). Another common tactic (heavily used by pharmaceutical companies) involves mentioning the chemical name of the best chemical compound that has been identified and tested to date, but burying the name of the “lead compound” in a long list of “candidate” compounds that are no longer of interest (I would challenge anyone, including any patent attorney, to take the name or chemical structure of any billion-dollar drug, and then go find the patent which claims and protects that drug). Yet another example involves deliberately making it impossible to tell which chemical compounds, in a long list of compounds, were actually created and tested, and which ones were only thought up, written on paper, and then named based on their drawn structures without ever being actually synthesized.

      As a result, quite a few patents bring to mind a comment by a computer expert about “Microsoft Word”, the word processing program: “It’s a curious and difficult combination of a word processor, and an encryption program.” The simple fact that there is no “Reveal Codes” option in MS Word (as provided by most other word processing programs) makes it clear that Microsoft doesn’t want users to know what is really going on inside their system. If that makes it harder for someone to figure out why some hidden formatting code is messing up a document, or how to fix the problem … well, that’s too bad. Microsoft has figured out how to run its business, it knows what the competition is doing, and it makes a lot of money.

      In the same way, there is no “Reveal Codes” option that can be used to help newcomers decipher patents that were written for big companies. So, if you’re getting started in this type of work, and if you haven’t yet had a lot of exposure to patents, the best way to start learning how patents should be organized and written is by analyzing patents that were written for independent inventors, or small companies, rather than big companies.

      If you’re trying to write a patent application that will be used to obtain or justify new investments that have not yet been made (as distinct from providing cover for investments that have already been made), then your patent application should try to explain, as clearly as possible, what is actually happening, and how and why your new invention works.

      That explanation should be written for five very different classes of readers: (1) scientists, engineers, and other technical specialists; (2) patent examiners, who have degrees in science or engineering but who may not have any expertise in that specific field of technology; (3) patent attorneys working for companies that might be interested in licensing or buying the patent; (4) business managers, who typically have accounting, finance, MBA, or similar degrees, and who will make the final decisions on whether to spend company money on a new project; and, (5) judges, who may be required to analyze and rule on a patent if it is ever challenged in court.

      Obviously, those five different groups cover a lot of different areas, and it should be recognized that the last two groups usually have had no training at all in science or technology, except for a couple of freshman courses in science, way back in college, 20 or 30 years ago.

      If you have an idea that deserves protection and investment, the best way (the only way) to reach and persuade all five groups is to be clear, methodical, and straightforward. Avoid legalistic phrases such as, `the hereinbefore mentioned component.' Plain, clear English is best, and before you even try describe the invention in written words, give special care and attention to the prior art, the drawings, and the Background section.


      Prior art is a legal term that refers to information that is publicly available (usually but not always in printed form) on the date you file your application. It includes patents, articles from journals and magazines, chapters from textbooks or encyclopedias, etc. It also includes catalogs, sales brochures and specifications, ads, technical and field literature such as instruction manuals issued by a manufacturer, etc, so long as the publications are reasonably available to interested people.

      The relevant prior art includes anything that (1) resembles your invention; (2) bears directly on your invention in any way; (3) describes any major component of your invention, even if it does not include other crucial components; or (4) describes how other people tried to solve the same problem you're trying to solve. Bring anything which might be relevant to the attention of the patent attorney. He (or she) can help you decide what should be brought to the attention of a patent examiner, and what can be omitted.

      No matter what an inventor's level of involvement will be in drafting a patent application, an inventor should be the expert on what he has done, how the technology works in that field, and what other people have done (and tried to do) in that line of technology. Don’t rely on a patent attorney to do that kind of technical work; his job is to take what you tell him, and convert it into legalese. It’s your job, more than his, to know how the invention actually works, and what makes it new and different.

      With the explosion of information that is freely available on the Internet, any inventor can and should do an initial patent search, download the relevant patents in printable form, and study those patents. Then, when it's time to talk with a patent attorney, the initial search will already have been completed, at the lowest possible cost, and the patent attorney can start building on a foundation, rather than having to start at ground zero with an excited inventor who is probably talking too quickly to follow completely.

      Get three clean copies of the most relevant prior art you can find. One copy is for your own "patent file". Keep it separate from your other files, so you can find articles quickly, without having to hunt for them. Typical billing rates for a good patent attorney are over $200/hour, so don't keep one waiting on the phone for fifteen minutes while you try to find a copy buried in some pile somewhere. The second copy is for the attorney's files. The third copy will be submitted with the application.

      Patent searching is a complex task, but an inventor who makes a diligent effort at it can often come up with one or more items that will offer a good start. A great place to start is the free U.S. patent database run by the Patent Office, available at To do free searching for patent applications filed in other nations, go to  (you'll set up an account, but it's free), and to

      Once you get into any of those sites, you may need to click various options that will lead you to a "Boolean" search. This is simply a weird word that means, in essence, that you can search by fields (such as by inventor's names, by assignee names in case a certain patent is owned by a company you're interested in, and by key words anywhere in title, abstract, or text). A Boolean search also lets you use "connecting" words such as AND or OR to fine-tune the search. For example, if you want to search for, say, any patents on heated ice skates, a Boolean search would let you search for "ice" AND "skat" (which would find both "skate" and "skating") AND ("heat" OR "warm").

  Once you know the patent numbers, patents from nearly any country can be downloaded from a site called  (you'll need to create an account name and number, but that’s free). The easiest way to work with these files is by downloading them as “pdf” files; this refers to “portable document format”. These files will print out a patent in exactly the same format as issued in print by the Patent Office, with columns, drawings, etc. Pdf files can be read and printed by software called the Adobe Acrobat Reader, available free from US patents can be downloaded for $3 each from Delphion. Since they’re public domain documents, they can be printed or distributed without limitation.

      If desired, text versions of US patents (which can be loaded into a word processor, and then searched or otherwise manipulated using the standard features of the word processing software) can be downloaded for free from the official PTO database, at

      If you obtain a relevant patent, it can help you find more prior art. For example, the front page lists all items of prior art (patents as well as other publications) that were considered while that patent was being examined. Some databases also list any subsequent patents that cited a certain patent as prior art.

      If you have Internet access, it’s a good idea to do your own search for articles in the scientific literature, for two reasons. First, a professional patent searcher might miss a relevant item during a manual search. They work with limited budgets and limited times, so you can never be sure they've found everything. Second, many professional patent searchers do not search any databases except patents. Many of them don't use computerized databases, and they don't consult sources such as Chemical Abstracts. If you're paying for a prior art search, ask the searcher to list any and all databases that were searched.

      Prior art is not limited to publications printed in America, or in English. For example, if a footnote or some other reference mentions a foreign article that might be relevant, try to obtain an abstract that has been translated into English. Various databases available on the Internet contain English abstracts of foreign articles, and it’s often possible to obtain those, either free, or through a subscription database that a company or patent law firm already has.

      If an inventor is seriously interested in doing as much prior art searching as possible before getting a patent attorney involved, he/she should also find out what kinds of databases or libraries are available in that field of science or technology. As just one example, anyone who works professionally with biochemistry or medicine should know about the National Library of Medicine (NLM), located in Bethesda, Maryland, which is paid for by government funds. The NLM makes its online database available to the public for free, at Start with a simple search, by typing in (for example) an author name and a known or suspected title word, in the small empty rectangle near the top of the screen. If that doesn't work, look for more advanced searching options. You can also go hunting for a Boolean search screen, and for medical subject terms that have been assigned to articles by a team of technician-specialists at the NLM.

      That’s just one example of an extraordinarily powerful and useful free resource, in one field of research. Various others are also available, and the easiest way to find them is by either (i) calling a  government agency that works in that field, and asking for the phone number of someone who works with computer databases that are run or used by that agency. In addition, call a couple of universities that are large enough to have science and/or engineering libraries, and ask for a reference specialist at that library. Such librarians are excellent sources of guidance on where to look for information, and they often will do a for-pay literature search, for a local inventor, at rates which are less than the costs of a typical patent search.


      The headings listed below are standard, and are generally required by laws and rules adopted by the Patent Office. Subheadings may be added at will. In the early stages, it's convenient to start each section listed below on a new sheet of paper, in case you need to add more information to it later.

      The title should be short, usually about 10 words or less. It's a finding aid only, and does not restrict the coverage of the patent. If you’re a big company, you can use an obscure phrase that doesn’t disclose anything useful (such as “Chemical Composition”). But, if you’re an independent inventor, this is your first chance to make a good first impression, so choose a good, helpful, descriptive title that can help you explain and sell your idea. 

Government  Support

      Omit this heading unless government funding was used. Otherwise, indicate which agency provided the funds. The grant number is optional.

Field  of  the  Invention

      This section is intended to help clerks in the Patent Office figure out which examining group ("art unit") should receive and examine the application.

      In the first sentence, list a broad field of technology (such as recycled materials, chemistry, pharmacology, automotive, construction equipment, computers, electronics, surgery, agriculture, etc.) Then add a second sentence which cites a more specific field or goal, such as "More specifically, this invention relates to drugs for treating XYZ syndrome" or "This invention relates to an improved machine and method for doing XYZ."

 Background  of  the  Invention

      This section should contain four different types or “clusters” of information.

      First: describe the overall context of the invention. Although many patents provide very little background information, this is a good chance to create a good first impression. Draft a short essay describing that particular field of science or technology, two to five pages long (double-spaced), so that an intelligent person who has never studied that area of technology will be thinking, "That's interesting; I never knew that," when he or she finishes reading that section.

      Second: explain each important term, at any length necessary to convey a clear understanding of that term. Point out distinctions, subtleties, and substitute terms, especially if a certain word isn't used the same way by everyone. Avoid dictionary definitions; those were written by grammarians who had to be brief, rather than scientists or engineers who appreciate the complexities of science and technology. Explain key terms not just as definitions, but in terms of how each component relates to other components. And, in case you leave anything out, give citations to a couple of standard textbooks or review articles used by scientists or engineers who work in that specialty.

      This set of paragraphs, establishing context and terms, can also help set the stage, by describing broad problems, such as industrial, technical, or social problems the invention can help solve.

      The third set of information, after the context and terms have been set, will describe each of several pieces of prior art. When citing patents, the best format is, “US patent 5,123,456 (Smith et al, 1988)” (the name and date can protect you, just in case there is a typographical error in the number). A full citation to an article published in a journal can be inserted into the text, or if numerous references are cited, abbreviated citations (e.g., "Smith 1980") can be used, if you also provide a list of complete citations at the end of the text. Titles of articles are not mandatory, but usually are a good idea, since most readers will wonder what they’re about if the title is kept hidden.

      For each piece of prior art, provide three different pieces of information. The easiest and clearest way is to draft three paragraphs. First, summarize the item of prior art. How does it work, what does it do, and how is it used? Point out directly relevant passages by page number (or by column and line numbers, in patents).

      Second, describe the limitations of that item of prior art, in factual language. Don't say it's obsolete, clumsy, or absurd; instead, point out its limitations in clear, factual terms as a preface to describing why your invention is better. Comparative terms such as `less expensive,' `more convenient,' and `relatively high' are fine, if it's clear what they're being compared against. If the prior art could not accomplish a desirable goal, describe that goal and say why it could not be accomplished until now. If the prior art has shortcomings (for example, prior machines had numerous moving parts, which made them expensive to manufacture, difficult to operate, and prone to breaking or jamming), describe those shortcomings. This sets the stage for the usefulness of the invention.

      And third, for each item of prior art, point out why your invention is better.

      These three paragraphs might be combined or reorganized in the final application, but setting forth each paragraph separately will help an inventor organize the information, and convey it clearly and quickly to a patent attorney.

      This advice, about creating 3 different paragraphs on each item of prior art, is a very good and very cost-saving piece of advice. The problem is, most inventors think they can do it, without actually doing it. They can--and almost always do--easily tell the patent attorney everything they know, about each item of prior art. And to the inventor, that approach works fine. The problem is, the patent attorney probably won't actually start drafting the patent application until several days or weeks later, and by then, several pieces of prior art are all schlumped together in a pile or file, and when that day finally arrives, the patent attorney will have to sort through them, try to remember what the inventor said about each one, and either dictate or write a narrative analysis that sorts everything out.

      So, here is probably the single best piece of advice in this entire article, for inventors who want to save time, money, and aggravation when working with a patent attorney. For each and every important piece of prior art, write down, using no more than 1 or 2 sheets of paper, an overview, using each of the three paragraphs listed above (the brief summary; limitations of that piece of prior art; and why the new invention is better). Then staple or paperclip that page, with your 3 paragraphs, onto that piece of prior art. Then give it to the patent attorney.

      If you'll do that, and actually do it (instead of just thinking about what you'd say, if you did it,), you'll be amazed at what a good patent attorney can do from that starting point, and how quickly he or she can do it. It will help the patent attorney draft an application better and faster, and then, a year or two later, after a Patent Office Action arrives with an "obviousness" rejection, it will help the patent attorney remember and figure out what happened, and why, a lot faster (which will save you even more money). So don't just think about it--do it. If you want to save money, write up all 3 paragraphs for each of the closest items of prior art, and give those analyses to the patent attorney.

      Finally, the fourth “cluster” of information that needs to have its own separate group of paragraphs, in a patent application, will be a list a set of goals and objectives for the invention. For example, "One object of this invention is to create a simple, inexpensive, and durable hyper-widget with fewer moving parts, and which can bounce higher than any previous widget or hyper-widget. Another object of this invention is to create a hyper-widget which can generate a more concentrated vacuum, in a smaller space. Another object is to create a hyper-widget which can be manufactured with fewer radioactive elements or toxic chemicals, and which still tastes pleasant and has few if any noxious odors, and which generates less waste during the manufacturing process."

 Summary  of  the  Invention

      This should be a brief (one page or less) overview for someone who doesn't know anything about the invention. This section (often repeated verbatim as the Abstract) will be published by a number of data services after the patent issues, and it will be read by people who will decide whether to order a copy of the entire patent. After describing the invention itself, add one or two sentences indicating how it is used and/or why it is useful.

      A lot of people will read the paragraph above, and nod their head and say, "Yeah, yeah, I can do that," when they really can't do it capably, because they've never actually tried it or practiced that skill.

      So, if you can, please take this approach and attitude: “It is a noble challenge to try to summarize a complex and important idea, in a single page of writing. And, as in so many other things, the proof is in the performance. The test of my invention is not whether I like it or feel proud of it, but whether I can persuade other people to invest their money in it. And the test of my ability to describe it is whether I can write a good summary, on a single sheet of paper.”

      Any inventor is urged to accept that challenge, and actually try to do it. One of my standard practices, when I begin working with an inventor, is to tell him or her, "Give me a three-to-five minute overview, with a summary of the whole invention. Don't start me out by walking me down a street and pointing out one building after another. Instead, show me a map of the place. Tell me where we're starting out, and tell me where we'll be going. And then, once we start walking down those streets, and seeing those buildings, I'll have some sense of how everything fits together, and where it's heading."

      To me, that's an entirely reasonable, obvious, and logical request. And yet, at least half of all inventors will respond to that request by waving their hand in the direction of a map, and then ignoring it, and start me walking down some street I’ve never seen before, while I have no idea where that street is going, or why I'm going in that direction. I'll usually listen patiently for about 15 minutes, then I'll gently interrupt the person and say, "Remember how I asked for a 3-minute summary, about 15 minutes ago? Well, I never got that overview, and I still need it. This tour you're giving me is very interesting, but I don't have any idea where it's leading, or how these pieces all fit together. So, let's start over. Can you please give me about a 3-minute summary, and overview, of what this invention is about, and what it does, and why it's useful?"

 Drawings  (also  called  Figures)

      Anyone who is interested in creating a patent, but who has not seen any patent drawings, should simply download a few mechanical patents, as pdf files, from Delphion, and see what they look like. It can be difficult to get them done just right, and a patent draftsman usually has to help put them in final “also called “formal” form; however, it’s pretty easy to figure out how they’re approached, and what the general goals are.

      The two main goals of patents drawings can be explained easily. First, they must help explain the invention. Any pictures or graphics you can create, which will help a patent attorney or business manager figure out what is happening and how this invention works, are good, and should be shown to the patent attorney or included in a provisional application.

      The second goal applies only to drawings that are submitted as part of a utility application. Any patent drawing that will be printed in a utility patent must be cheap and easy to print, and it must remain clear and legible even after it has been copied into fourth or fifth generation photocopies. As a result of that second goal, more than 99% of all patent drawings in utility applications are limited to black-and-white line drawings, with large lettering and callout numbers. This includes line drawings with callout numbers, but it nearly always excludes photographs unless absolutely essential. Drawings can be graphs, flowcharts that show the steps carried out in a multi-step process, and just about anything else that can be converted into a plain black-and-white graphical format that is easy to print and will remain legible in  A 4th generation photocopy.

      Small strip-charts and other machine-generated printouts can be submitted as drawings, if they are properly inked to convert them into clearly reproducible black-and-white drawings.

      “Stippling” (tiny black dots which create a shade of gray) is prohibited, since it will not print or copy well, but nearly any kind of cross-hatching is okay, so long as the cross-hatching lines are bold and are not confused with the drawing lines. Any letters or numbers must be at least 1/8 of an inch tall, so they will print well.

      Tables containing data, and chemical formulas and structures, are rarely submitted as drawings; instead, those types of material are preferred as inserts in the text , even if they require the 2-column format of a patent to be interrupted. This is because a patent examiner can handle anything in the text, including tables or chemical structures, while a separate review and approval, by the PTO’s Office of the Draftsperson, for all drawings.

      The PTO is very reluctant to print photographs in patents, because of the printing expense. However, any good photographs should be shown to the attorney, and they can be submitted to an examiner, as exhibits which accompany a declaration or affidavit that will not be printed as part of the patent.

      Carefully consider the best way to show your invention visually. Many patent attorneys and examiners look at the drawings before the text, so drawings can create the first impression. For that reason, they're worth extra effort. Study the drawings of several patents in the same field as your invention. That will give you a sense of how patent drawings are arranged and what their goal is.

      When you're ready, draw up your own rough drafts. Even if they aren't high quality, they'll help the draftsman. Your only goals are to make them large, and clear. Try to make each drawing (even drawings of sub-components) fill up at least two thirds of a sheet of paper. Paper is cheap, and few things are more annoying than staring at a tiny little squiggle, while listening to an inventor point out a dozen parts that are invisible to anyone but him. Computer-generated drawings are often helpful, and they can sometimes be used as the formal drawings.

      Drawings in full-scale utility applications patents normally use "call-out numbers" which are explained in the text. Inventors should avoid those; don't even try to assign any numbers. Instead, give the patent attorney two copies of your hand-drawn versions. On one copy (the "plain vanilla" copy), don't put any words or numbers at all; just draw the device, graph, or whatever, and leave the drawing as open and clean as possible, so the patent attorney will be able to use it in any way he chooses. On the second copy, add labels (words or phrases) to identify each component, or to point out specific things that deserve extra attention in the drawing. Put any labels out around the edges of the paper, and use arrows to point out each labeled component, line on a graph, etc. You will end up using those two drawings to help the patent attorney figure out and assign any names that need to be assigned to components that need to be named. Later, the patent attorney will assign call-out numbers, after he develops a better sense of how the final drawings should appear.

      Drawings of prior art are rarely printed in patents. However, such drawings can help people figure out how your invention works, so if there are good prior art drawings, make copies and include them with your first draft. Clearly label them as prior art (for example, "This is the device described in patent 4,567,890 by Smith"). 

Drawings  in  Provisional  Applications

      As noted above, the requirements listed above apply to drawings that will be printed by the Patent Office. And as noted earlier, provisional patent applications are never even examined, let alone printed, by the Patent Office.

      This leads to a conclusion: there is unlimited flexibility and leeway, in drawings, photographs, and other graphical materials that are submitted with provisional applications. Have you got some good color photographs, showing what you’ve done? Or some color printouts from a Powerpoint presentation? Fine; you can include them with your provisional application, but make sure they look impressive and professional, and are of a caliber that can help sell your proposal to a business manager.

      And, you don’t need callout numbers in a provisional application Instead, if you want to make your drawings as easy as possible to decipher, feel free to put written-out names around the margins, with arrows pointing to each named component. And feel free to add a block-lettered legend at the top or bottom of the sheet, explaining what that drawing shows.

      It should also be recognized that, even in utility applications, final (“formal”) drawings are not required to be submitted with the initial application. One way to speed things up and save money is by filing a utility application with “informal” drawings. These can be drawn by an inventor or patent attorney, using simple tools (straight-edges, curved templates, etc.), or they can be drawn by computer graphics. If adequate versions can be generated, the expense of having formal drawings done by a professional draftsman can be delayed until after the application has been examined, and fully allowed. If a Notice of Allowance is received, formal drawings can be created; and, if the application isn't allowed, that expense can be avoided entirely. And, if an inventor is in a hurry to get an application filed, informal drawings can be used to eliminate a waiting period while a professional draftsman finishes his work. 

Brief  Description  of  the  Drawings

      This is a very important section, since nearly any patent attorney or patent examiner will quickly go to the drawings, and try to figure them out, before reading anything other than the abstract or summary. Instead of using these “brief descriptions” (which will be no more than 2 or 3 sentences, for each  drawing, in sequence) to master the entire invention, the attorney or examiner will look briefly at a drawing, try to figure it out, and then read its brief description, to see if he was right about what that drawing is about.

      Typical brief descriptions might read:

      Figure 1 is a plan view showing the hyperwidget of this invention.

      Figure 2 is a side elevational cutaway view, along line 2-2 in Figure 1.

      Figure 3 is a circuit diagram of a control unit for the hyperwidget of this invention.

      Figure 4 is an exploded view showing the components of the hyperwidget.

      Figure 5 is a chromatograph showing the purification of the ABC substance by the hyperwidget.

      Figures 6A and 6B are before-and-after depictions, showing the untreated surface prior to treatment (FIG. 6A) and the same surface after the treatment (FIG. 6B).

Detailed  Description

      This section (which used to be called, “Description of the Preferred Embodiments”) contains a narrative explanation of the invention. This section must contain a complete and detailed narrative which describes the figures in detail, listing and explaining each and every callout number used on any drawing. This is also the proper place to describe any potential enhancements, such as additional or alternative components, methods, etc. that might be incorporated into the invention at some time in the future, even if they are not present currently.

      If an invention involves chemistry or biology, the examiner will want proof that the process or compound exists and functions in the way the inventor claims. This usually requires a separate section called "Examples" which will specify detailed recipes (such as exactly what weights of what compounds were mixed together, how long they were stirred or incubated, etc.) and experimental results, as described below. The "Detailed Description" section should be a narrative overview of those examples; without repeating the exact recipes, this narrative will explain why those steps were taken, and what was happening during each step of the process as a certain compound was being made or tested.

      An inventor must reveal the "best mode of carrying out the invention" that is known on the day the patent application is filed. The inventor cannot hide the best version and keep it as a trade secret while using an inferior version to obtain a patent. This is a rigid requirement which is strictly enforced. If there is any hint that an inventor was playing hide-and-seek games with the best mode requirement, most courts will simply strike down the patent. So, be up-front and entirely forthcoming with the best version you know about on the day the application is filed.

      However, there are two major rules that clarify and adapt the best mode rule, and both should be known to any inventor. First: improvements that are discovered after an application is filed can be kept as trade secrets. Therefore, the way to win at the best mode requirement is by filing an application promptly, then following up the initial discovery with more research after the application is filed.

      Second: even though the best mode must be disclosed (which means, it must be mentioned somewhere in the application), it does not need to be identified, highlighted, or emphasized as the best mode. As an example, it is standard practice for drug companies to place their “lead compounds” (the compounds that performed best in initial tests, and which will be studied in more detail in subsequent tests) somewhere the middle of a much longer list of compounds which did not perform well and are no longer of any interest in any further tests or trials. Their attitude is, “If my adversary and enemy that wants to steal the fruits of my work and wants to know which compounds worked best, fine. Let him do the work, and he’ll find out. I’ve already told him where to look, and that’s all he’s entitled to get from me.”

      There is another rule that also deserves careful attention. Inventors and their attorneys must act with complete candor and good faith in bringing all relevant information to the attention of a patent examiner. Inventors must be very careful about how they treat unsuccessful experiments, early prototypes that didn't succeed, etc. If they try to conceal negative data, the patent can be invalidated after it is issued. Bring all the data to the attention of the patent attorney, and clearly point out the best mode, to the attorney. The attorney will help decide on the best way to present the information.

      In addition to pointing out the best mode, the inventor should try to anticipate possible modifications to his invention. Typical language: "In an alternate preferred embodiment of this invention, the positions of the cardioblot and framistat can be reversed. This would cause . . ."

      It is also possible to use functional language in describing and claiming an invention. For example, suppose that in a machine, it is necessary to securely mount part A on part B, but it doesn't matter how you do it. The two parts could be screwed together, bolted, welded, nailed, glued, clamped, etc. Instead of specifying ways of attaching those parts, you can simply refer to "means for attaching part A to part B," and give one or two suitable examples, such as bolts or welding.

      Similarly, it is often possible to use functional language in the chemical or biological field, such as "any pharmaceutically acceptable salt of this compound," or "at a dosage sufficient to generate an immune response in an injected animal, using a suitable carrier substance. Such carrier substances are well known to those skilled in the art." 


      Most chemical and biological patents must include a section called “Examples”. These must contain any hard data, to assure a patent examiner that the molecules, mixture, cells, etc., have been created and perform as claimed. Otherwise, people could get speculative patents merely by synthesizing a compound, without having to prove its utility. Such patents would discourage research rather than promote it.

      If examples are necessary, they should contain enough detail and data to allow "someone with ordinary skill in the art" to repeat the experiment and get successful results. "Ordinary skill" refers to someone who can carefully follow written instructions, rather than a creative whiz who can find some way to overcome any obstacle. The instructions can be very complex and demanding, and the level of skill required to follow it can be high; for example, in the field of biotechnology, "ordinary skill" refers to people with PhD degrees, since nearly any biotechnology company will have one or more PhD's. The key point is that "ordinary skill" must not require a high degree of imagination, or creativity in overcoming unforeseen obstacles. Instead, the invention must be capable of being carried out without “undue experimentation.”

         Whenever a reagent, piece of equipment, or other item is mentioned, the supplier should be listed along with the city where the supplier is located. Model numbers (for equipment or hardware), product codes or trademarks, and purity classifications or other specifications (for chemicals) should be specified the first time that item is mentioned. If you use a trademark, describe what that item is in technical terms, or by referring to some publication that describes it (e.g., "This product is a granular polyacrylamide with an average molecular weight of X daltons, an average particle diameter of Y microns, and a charge density of Z coulombs per gram," or "This chip is described in the Microprocessor Specification Handbook, published by the Magnarola Company, of Piston, Michigan"). If you used a special cell or plasmid, indicate not just where it is described, but where it is available to the public (if it's not available, it might be necessary to deposit a culture with a public depository). If you refer to a process that is described elsewhere, give a citation to the article or book where it is described, and point out any deviations from the published version. Even if the only deviation was increasing the quantities used, describe what actually happened. 


      If citations were given in abbreviated form (such as "Smith 1980") in the text, the full citations (including titles of articles or chapters) should be listed here, in alphabetical order. Example: Smith, W.C., et al, "A New Treatment for Payne Syndrome," Science 123: 83-87 (1980). 


      The claims of a patent define the property of the inventor. Everything else is merely explanation, which is necessary to support the claims. Anything which is described but not properly and carefully claimed becomes freely available for public use.

     Claim drafting is a highly specialized skill, and inventors emphatically should not try to draft their own patent claims. The tricks and traps of claim drafting are so tortuous and complex that if an invention does turn out to be worth millions of dollars (which in most cases will not be known for several years), and if the patent is challenged by competitors who are absolutely determined to move in and grab as much of that cash as they can get without being sent to jail, pitting an unskilled novice who tried to draft his own patent claims, against a skilled and seasoned patent attorney, would be like sending a naïve, unarmed kid to fight a strongarm thug with a bad attitude, a long arrest record, and a knife. It would not be pretty, and it would not be respectful or polite.

     Nevertheless, inventors should understand some basic facts about patent claims, so they can tell whether their attorney is working hard and fighting for their best interests, or taking an easy way out and settling for claims that are too narrow to offer any real value or protection against infringers.

      There are two crucial rules in interpreting, applying, and enforcing patent claims:

      1. Each and every "limitation" in a claim must be infringed, or else that claim is not infringed.

      2. If any claim in a patent is infringed, then that entire patent is infringed. The patent owner does not need to show infringement of more than one claim.

      The first rule acts is a major warning, and threat, for inventors. In the broadest claims, do not include any limitations that are not absolutely essential. Every limitation in any independent claim should have a clear reason for being in that claim, and the only two reasons for including a limitation in an independent claim are (1) to define and describe the invention, or (2) to avoid prior art.

      For example, suppose you invent a machine with four essential parts (A through D), and an optional part E which helps it work better. Your broadest claim (which will be an independent claim that stands alone and does not refer to any other claim) should specify only parts A through D. It will also need to specify the relationship between those parts, and it must describe how the machine is constructed and/or how it functions; you cannot claim a machine merely by listing its parts.

      After you have listed essential parts A through D in an independent claim, part E should be specified in a “dependent” claim, which will be written in a form such as, “The machine of Claim 1, which also contains part E.”

      Why is it done this way? Well, if your broadest claim lists all five components (A through D), then competitors can legally avoid that claim (they can make, use, and sell your invention, without paying royalties) by merely getting rid of part E, or replacing it with some other component.

      After the patent attorney has drafted the claims, study every word carefully. See if any words can be taken out while still distinguishing your invention from the prior art. And, because of rule 2, try to draft several independent claims from several different angles, using different phrases. One of those claims might cover it in a way that other claims missed.

      While you're working with a patent attorney, don't be afraid to question, challenge, and probe. It's your child, and you're paying the bills. A good attorney won't get upset, any more than you should get upset at him if he challenges you to back up your assertions. In a good relationship, each side works hard while challenging the other side to do the same. If an attorney cannot or will not clearly explain why he did something, or why he used a certain phrase in a claim, that's a danger sign. 

Should a Provisional Patent Application Contain Claims?

      The comments above, about claims and claim-drafting, lead back to a major question: should a provisional patent application contain claims? This question arises because the patent law says that provisional applications do not need to have claims.

      I’d like to address that question, because the answer leads back to an important piece of advice about attitude, salesmanship, and persuading investors to actually invest hard-earned cash in idea and a patent application.

      Some patent attorneys take the position that any provisional application should have claims at the end, to provide additional evidence of what the inventor actually regarded as his/her invention. However, the author of this article leans strongly in the opposite approach, and takes the position that -- unless a particular reason points strongly toward including claims, and unless a patent attorney is involved in making that decision and drafting those claims -- then the safest approach is to simply omit any claims, especially if an application is being prepared by an independent inventor or small company.

      Because of their very nature, provisional applications should be written and filed quickly, without months of delays for extensive, exhaustive efforts to make sure the prior art has been thoroughly evaluated, the claims have been drafted carefully, etc. A provisional application by an independent inventor or small company says, in effect, “I was just getting started, and I knew I needed to hurry up and get something filed, and I wasn’t sure, at that time, which direction or directions this idea would end up taking. So, I was just trying to get the patent process started. The patent law says that claims are not required, so instead of spending time trying to draft claims at that very infant stage, I decided to spend my time working on improving the idea.”

      Just as importantly, putting a set of claims in a provisional application creates a major risk of turning those claims into a “sitting duck”, where opponents (i.e., lawyers who are getting paid by the hour) can argue at great length (and expense) that the provisional claims were invalid, overbroad, or whatever. And, your opponents can argue strenuously that, if you didn’t adequately claim some particular aspect of your invention, then you obviously failed to realize or anticipate an important aspect or enhancement of your invention, which your opponent is now trying to take away and claim as his own patented improvement.

      Rephrased in the vernacular: don't try to climb up on a horse, unless you're ready to ride it. Even if you think you know what will probably happen, you can’t be sure what that horse will do, after you're on it. If a car horn suddenly honks nearby, or a dog or rabbit suddenly darts out of a nearby bush, that horse might bolts, and take off running, and it won’t stop just because you ask it to.

      In the same vein, unless you’re ready to start fighting for whatever claims you put into a provisional application, you’ll most likely be safer if you don’t have any claims in it at all, for opponents and competitors to attack and criticize. Instead of proclaiming that you're a skilled rider, before you're even up on that horse, approach it with respect, caution, and humility. If it's a real horse, it'll be bigger, heavier, and stronger than you.

      In the same way, everything about the tone, content, and attitude of a provisional patent application should suggest, “I think this might be a good idea. I've been working on it a while, and from what I've seen so far, no one else seems to be doing it.” No one will try to rip your head off for making a statement like that. Indeed, if you're “the little guy,” trying to do something worthwhile and help other people, they'll probably start rooting for you.

      By contrast, if you walk into the office of a patent attorney (or an experienced business manager) and start saying, “I've thought of a super-colossal idea, and it's going to make millions, and it all belongs to me,” I can promise you – that manager or patent attorney has heard that song before, too many times, from people who never ended up doing what they said they could do. Instead of seeing respect and admiration, you’ll see eyes glazing over, and a face that just can’t or won’t understand what you’re trying to tell it.

      So, this author’s advice is, don't start out by making claims. Don't take the position that you think you've got the right to decide, and claim, what you think belongs to you. Instead, roll up your sleeves, start working on the problem, and try to convince any onlookers or passers-by that you're more concerned with making a useful contribution, than in how much money you're going to make.


      Patents aren't really scientific documents; they're legal and business documents, and the essence of both law and business is conflict and competition. Patent examiners get paid to challenge your application and find any flaws. And if your patent becomes valuable, people who want to infringe or ignore it will try even harder to find any flaws in it. It's best to find those flaws before an application is filed, even if it means a bit of a struggle.

      In the words of Frederick Douglass, "If there is no struggle, there is no progress." And in the words of William Shakespeare, "Life has given nothing to mankind without great struggle." The goal is not to avoid conflict; the goal is to learn how to work with conflict and competition, and learning how to turn them to your advantage.

      The struggle to create something important and valuable is the finest, most rewarding, and potentially most enjoyable part of any career. But it can be difficult. The era is long dead when an explorer could claim a huge tract of land merely by erecting a flag somewhere. Today, if you want to be recognized as the creator of something, you'll have to work for it.

      Most inventors enjoy challenges, and trying to get a good patent can be a real challenge, in every sense of the word. I hope this article can help, and I hope you'll enjoy being an inventor, as much as I enjoy working with inventors. They’re some of the finest, most interesting and creative people who have ever lived, and it’s a genuine privilege and a pleasure to be able to live the kind of life where I get to work with inventors for a living.


            Pat Kelly received an Engineering Sciences degree from the University of Texas at Austin (1974), and a Juris Doctor from Harvard Law School (1980). He has practiced patent law since 1980, and specializes in biochemistry and medical technology. He is also an inventor in his own right.