Introduction to Patents and Patent Searching

C. Understanding Keyword Searching

When you use a computer program to search for patents, you often must search for them by entering words into a "query" box and asking the search program to match your words with words stored in its database.

As you might expect, performing keyword searches is a skill with a learning curve. Sure, anyone can put one or two words into a box and pull up all the patents with those words. No skill there. But the overall number of patents you pull up is likely to be huge and the number of the patents that are relevant to your search are likely to be low. To pull up a manageable number of patents and to assure that most of them will have some relevance to your own invention, you will need to know at least some of the basic techniques for choosing your search terms and combining them into meaningful search queries.

1. The Role of Wildcards in Keyword Searching

One powerful tool that is often used during keyword searching is called the wildcard. A wildcard is a special character inserted into your keyword. This character tells the computer search program to do something special with the keyword within which it's used. The two most often used wildcard symbols are the asterisk (*), and the question mark (?).

The asterisk wildcard is used at the end of a word root to take the place of any number of additional letters that may come after that root. For example, assume you have invented a new type of dance shoe. The shoe can be used for ballroom, ballet and tap dancing. In addition to the keywords "ballroom," "ballet" and "tap," you will certainly want to search for the word "dance." But there are several variations of the word "dance," "dancing," "dancer," "danced," and even "danceable." By using "danc*" as your keyword, the asterisk replaces any other possible characters that would follow the four letters, "danc."

Figure 6, below, contains the search results from searching the titles of U.S. patents issued in the years 1997-1998 for the word "dance." The patent titles that have the word "dance" in them are listed and numbered. As you can see, there are four patents that have the word "dance" in the title. The first title relates to a dance practice slipper, the second title concerns the sole of a dance shoe, the third title relates to a type of dance and the fourth title relates to a portable dance floor.

Figure 7, below, contains the search results from searching the titles of U.S. patents issued in the years 1997-1998 for the word "danc*." As you can see, we now have eight patents listed. The first two titles are the same ones that we obtained before. However, title numbers 3 and 4 contain the word "dancing" as opposed to "dance." The first four letters (danc) are the same as in the word "dance," but the wildcard (*) was used for the letters "ing." Similarly, title numbers 5 and 8 contain the word "dancer," as opposed to "dance." Here, the wildcard (*) was used for the letters "er."

The question mark (?) wildcard can be used to replace any single character in a word. Continuing with our dancing example, the words "foot" or "feet" could be searched by using "f??t" as our keyword. Obviously, you would not want to use the keyword "f*," as this would return every word that started with the letter "f." By using "f??t," every four-letter word that starts with "f" and ends with "t" would be searched for by the computer. For example, along with the words "feet" and "foot," the words "flat" and "fast" would also be reported to you in the search results.

In Figure 8 below, we show a portion of the search results obtained from the IBM electronic patent database when searching for the word "f??t" in patent titles for the years 1995-1998. The first patent (Patent Number 5,697,106) contains the word "foot" in the title, whereas the next patent (Patent Number 5,696,529) contains the word "Flat" in the title. The third and fourth patents listed have the word "Fast" in their titles. Finally, the fifth and sixth patents listed (Patent Numbers 5,695,527 and 5,695,526) contain the word "foot" in the title.

There are other wildcards used on the various computer search systems available to the public. However, the asterisk (*) wildcard is, by far, the most popular.

2. The Role of Boolean Logic in Keyword Patent Searching

A search technique known as Boolean logic can be used to combine individual keywords into powerful searches. Boolean logic uses a total of four words (called "logical operators") to define the search: AND, OR, XOR and ANDNOT. The AND operator is by far the most useful. A graphical representation known as a Venn diagram will help you to understand how these operators work.

In Figure 9 above, we have a circle that has been shaded. The area inside the circle represents all of the patents that contain the keyword represented by the letter A. The area outside the circle represents all the other patents that do not contain the keyword represented by A. In other words, if we were to search a database of patents for all the occurrences of the keyword A, our search results would be contained in the shaded circle above.

In Figure 10, we have two keywords represented by the circles A and B. Searching for individual occurrences of the keywords A or B would result in a lot of search results. It would take a long time to review these results and most of them would be irrelevant.

For example, let's suppose we have an invention idea for a new type of telephone cable. A search for the keyword telephone would return numerous references to different types of telephones. Similarly, a keyword search for the word cable would return patents related to cable television, bridge support cables, cable cars and so on. What we need is a way to search for both the keywords telephone and cable within the same patent. This is where Boolean operators come into the picture.

3. The AND Boolean Operator

In Figure 11 above, we have used the Boolean operator AND to combine the keywords A and B. The shaded area where the circles overlap represents the search results that contain both keywords A and B. As you can see, the AND operator is a great way to narrow the scope of the search.

When a match is found between a keyword (or a combination of keywords) and a patent, the result is called a "hit." When patent searches are conducted, the number of hits, or occurrences, of a keyword match is usually reported to the user. By using the AND operator, the user reduces the quantity of hits that need to be reviewed.

For example, let's suppose that you have invented a new type of steam engine. A steam engine is a machine for converting the heat energy in steam into mechanical energy by means of a piston moving in a cylinder.

The search results using the keyword Engine are shown in Figure 12 above. The shaded circle represents all of the patents that contain the word Engine. This could be quite an extensive list. For example, all the various types of internal combustion engines would be included in this list. A steam-powered vehicle is an external combustion device; the steam is usually obtained from an external boiler. However, if we only searched for the word Engine, we would have to review search results that contained references to gasoline-powered engines for cars, trucks, trains and all other engine-powered devices.

Figure 13 above shows the search result obtained when using the Boolean AND operator to combine the keywords Steam and Engine. The resulting number of hits is represented by the small shaded area in the diagram, where the two circles overlap. We can see at once why AND is the most often used Boolean operator. It allows the searcher to narrow the scope of the search and obtain more meaningful results.

In Figure 14 above, we see the results of a search using the keywords Steam, Engine and Car. In this case the overlapping area is even more precisely defined. It would be necessary for a patent to contain all three keywords before being reported as a match.

Throughout this book, we will identify critical concepts used for effective patent searching. We call these concepts Searcher's Secrets. The use of the AND operator brings us to Searcher's Secret #1.

4. The OR Boolean Operator

In Figure 15 above, we have used the Boolean operator OR to combine the keywords represented by the letters A and B. The shaded area within the circle labeled A represents all of the patents that contain the keyword represented by the letter A. Similarly, the shaded area within the circle labeled B represents all of the patents that contain the keyword represented by the letter B. When you use the Boolean OR operator, you can't tell from your search results whether a particular reference contains just one of your key words or both. Using our Venn diagrams to represent one possible set of search results, we see that in Figure 15, above, there were no hits that contained both of the keywords represented by the letters A and B. If the search results did have some patents that contained both keywords, the resulting Venn diagram would look like Figure 16 below.

In Figure 16 above, we have the two circles, A and B, with a small overlapping area. The lightly shaded areas of A and B that do not overlap represent patents that contain only one of our keywords. The heavily shaded, overlapping area represents patents that contain both keywords.

Returning to our steam engine example, Figure 17 represents the number of hits returned when we use "Steam OR Engine" to search the patent database. What this means is that any patent that contained the word Steam or the word Engine would be returned as a match. The lightly shaded areas of the circles represent patents that contain the keyword Steam or the keyword Engine, but not both. The heavily shaded area, where the two circles overlap, represent patents that contain both keywords. Remember, however, that you couldn't tell this from your research results; the Venn diagrams are only being used to explain what happens in fact. Use of the OR operator brings us to our next Searcher's Secret.

5. The XOR Boolean Operator

The exclusive OR operator is symbolized by the XOR letters. This operator is very similar to the OR operator, but with one important difference. The overlapping area is not included in the search results. So, if we used "Steam XOR Engine" to search our database, we would obtain a list of patents that contained the word Steam, or the word Engine, but not both. This brings us to our third Searcher's Secret.

6. The ANDNOT Boolean Operator

The final Boolean operator we will be reviewing is the ANDNOT operator. The ANDNOT operator is actually a combination of the AND and NOT operators. The NOT operator, by itself, simply finds all the patents that do not contain the keyword used. The reason the NOT operator is combined with the AND operator can be seen in Figure 19. If you were to use the NOT operator, by itself, with just the keyword Engine, your search results would include all the patents that do not contain the word Engine -- a very large search result indeed.

An example of the correct use of the ANDNOT operator is shown below. If you wanted to search for steam engines used in all devices except trains, you could compose a query like:

(Steam AND Engine ) ANDNOT Train

This would return patents concerning steam engines in cars, boats, etc. However, any patent containing the word Train would be excluded. This is true even if the words Steam and Engine were contained within the train-related patent. Figure 20 below shows the search result obtained when using the above query. The resulting number of hits is again represented by the small shaded area in the diagram where the circles representing the keywords Steam and Engine overlap. However, a small section of that overlapping area has been excluded. This excluded area represents the patents that contain the keyword Train.

Use of the ANDNOT operator brings us to our next Searcher's Secret.

7. Use of Parentheses

Also, notice that we have made use of left and right parentheses -- ( ) -- around the words Steam and Engine. This means that the words within the parentheses are evaluated first, then the ANDNOT condition is applied.

We will cover the use of parentheses for advanced keyword searches later in this book (Chapter 6).

You can also combine wildcards with Boolean operators. To return to our dance shoe example, we can combine the keyword "danc*" with the keyword "Shoe," and exclude the keyword "Tap" with the following query.

(Danc* AND Shoe) ANDNOT Tap

The resulting patents would have the words "shoe" and one or more words like "dancing," "dancer" or "dance," but not the word "tap."

Summary

What Is a Patent, and What Does It Do for Me?

• A patent is a right of exclusion, granted by the government, for a term of years. It is a document as well as an abstract right.
• A utility patent covers the functional aspects of the invention. A design patent only covers the appearance of an invention.

How Keyword Searches Work

• A computer program matches the words you type in (keywords) with words stored in its database.

How Wildcards Work

• The asterisk (*) wildcard can take the place of any number of letters following its location in the word.
• The question mark (?) wildcard can be used to replace any single character in a word.

Boolean Logic

• Use Boolean logic to combine keywords into powerful searches.
• The more keywords used with the AND operator, the smaller the number of matches obtained and the more meaningful each match is to the searcher.
• The OR operator is used to widen the scope of the search results.
• One, and only one, of the keywords combined with the XOR operator will appear in each of the patents in the search results.
• The ANDNOT operator is used to exclude keywords from the search results.
• Left and right parentheses -- ( ) -- are used in complex Boolean logic to determine which terms are evaluated first.
• You can also combine wildcards with Boolean operators.

Back to Top