Suppose you’re an inventor hoping to get a piece of the electric vehicle market share, and you’re in the process of preparing a patent application describing your brilliant idea: an improved electric unicycle.  Once you’ve worked out the fundamental mechanics, the next step to ensure that you aren’t the “third wheel” is to run a prior art search.  This may seem straightforward enough.  Type in a quick set of keywords into Google Patents, say electric unicycle, and voila: you’re supplied with over 3000 results.

However, as with database searching in general, naïve queries may return more (or less) information than desired.  It is up to you to determine which terms will trim the excess and which terms are already proving too limiting.  In the former case, this is not only a matter of making your search specific, but of removing irrelevant prior art.  There are several tricks that you can use to facilitate this ordeal.

One trick is to remove terms that you don’t imagine will show up in any publication of interest by using a minus sign.  For instance, the search electric unicycle -toy* will omit all patents which are directed to toys instead of some means of personal transportation (note the use of an asterisk here – this is a wild card, which will remove not only “toy,” but any term that begins with “toy” as well, such as “toys”).

Another trick is to lop off older prior art which may have little relevance.  If you’re at the forefront of your field, perhaps you’re only in danger of reproducing the newest innovations.  Most patent search engines will allow you to search within a specific range of publication dates; in Google Patents, you may do so by appending a date in YYYYMMDD format as, for example: after:publication:20000101 (conversely, you may omit everything published before this date with before:publication:19991231).

Of course, the process of exhaustively narrowing down everything that isn’t relevant to you may prove a herculean task.  You shouldn’t spend an inordinate amount of time at this stage.  Simply appending quotations to your search term can immediately hand you publications truly reciting electric unicycles.  Indeed, “electric unicycle” only returns ~600 results.  An important caveat, though, is that while you may be certain that these results discuss electric unicycles, you may be omitting cases which discuss electric unicycles in a non-obvious or patent-like manner (e.g., “electric monocycle,” “unicycle with an electric motor,” etc.)

To that end, you may add flexibility to your search by taking advantage of Boolean search operators.  In Google Patents, AND operators can be represented implicitly by placing two keywords in separate sets of parentheses.  Thus, the search query (electric) (unicycle) is equivalent to electric AND unicycle (in Google Patents, this is also equivalent to the first example query electric unicycle).  Perhaps you’d like to see results on electric monocycles as well.  Simply adding an OR operator will account for this: (electric) (unicycle OR monocycle).

Be advised that generic combinations of these directed search techniques will only be of limited utility in the preparation of any given application, and must therefore be judiciously tuned to the specific invention at hand.  In the next few posts, we will explore how to further narrow search queries to serve a more particular purpose.

Comment