I spent four years developing a patent-able idea, but there already is a U.S. patent that claims to produce exactly the same result but has a critical flaw that renders it inoperable.

Any person skilled in the art following the the outlined claims and interpreting the terms in said claims with an authoritative english dictionary, or even constructing it according to the specification, will quickly realise that the entire construction can amount to nothing but a very complex paperweight because something that is mentioned explicitly makes it non-working.

Obviously either the inventor meant something and had the right idea, but the patent clerk wrote something else down, or the inventor didn't fully understand what they were talking about but patented a gut feeling.

What I mean to say with this is, if I were to produce and sell a working product that fulfils 99% of the claims, and were to be sued, I am certain I would win.

However, my question is, how likely can I patent my version of the widget?

99% of the claims would be verbatim, except for one word, which is similar, but critically different. The specification and drawings however would differ substantially.

would they call it an obvious correction and not really a novelty? can a patent that cannot work be considered prior art?

will I get a patent or will the whole thing become effective public domain?

2 Answers 2


In a patent specification, you are free to define a word so that it means just what you define it to be. You could also give examples to say that the word does not mean certain other things. This definition would then govern the interpretation of your patent and not the general meaning of the word. For example in ordinary English there can not be a diagonal to a closed curve as the curve has no vertex. But in your patent you could define a diagonal as a line that cuts the curve into two or more parts. This way you could be sure that your claims are read properly.

  • This is partly at what I am looking for, thank you. What about the prior Art part? Is there a precedent?
    – guest
    Sep 29, 2016 at 15:18
  • This is guess work and I have no knowledge if there are other patents with such language. Oct 1, 2016 at 1:31
  • You could write equivalent of two paragraphs in your prior art description. The first paragraph would describe prior at as it was before the 'non-working' patent. The second paragraph would describe the non-working patent where you would demonstrate why it does not offer an improvement over prior art. Oct 1, 2016 at 1:36
  • From Wiki pedia article on Clod Fusion Oct 4, 2016 at 15:05

The previous patent can be cited against your application to show that your claims are not novel or are obvious. It is assumed that granted patents are for things that work and are enabled in the patent.

If, in fact, it does not work then it can be dismissed as prior art. You will need to show that it does not work (without your added twist) by affidavits from experts.

  • I provided this late answer since the accepted answer does not address this key point.
    – George White
    Dec 28, 2021 at 23:53

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .