Human to human language is widely unrestricted, even nonsyntactic, in patent documents. Yet those documents are carefully written by the inventor, carefully critiqued through at least two office actions by the PTO examiner, and the claim language is debated, reviewed, checked against prior art, and finally allowed to issue for some small fraction of patent applications.

My software treats them as unrestricted text, and I think that is by far the most feasible and effective way to treat them. It is fairly easy to break up the text into sentences, with periods, question marks and exclamations able to detect at least 80% of the sentence endings. False endings (especially with periods and acronyms) can be effective endings in most of the other 20%, with a few mismarks remaining in nearly every patent.

Claim language, though unrestricted, has some statutory words and punctuations which can be detected and used to effectively break long claim statements (well over 100 words) into elements.

The elements also contain rarer words that designate the claimed materials, ideas, systems and methods. Those rarer words are useful in tracing each claim element back to sentences in the specification which form the disclosure of the invention.

So my software uses mostly unrestricted forms of English, except for the statutory lexicon needed to structure the interpretation out of the nearly unrestricted specifications, claims, abstract, and more highly structured database-like columns.

-Rich

Hi John,