The prior art nobody checked

What actually happens

Most inventors search the way an engineer searches: for products on the market, in their language, in their industry. Examiners search patent databases across jurisdictions, including applications from twenty and thirty years ago that never became products, in Japanese, German, Chinese and Korean.

The gap between those two searches is where most novelty objections come from. The cited document is rarely a competitor. It is usually an abandoned application from a company that no longer exists, describing something nobody ever built.

Why the search changes the drafting, not just the decision

The point of searching is not only to decide whether to file. It is to find out where the closest existing disclosure sits, so that the claims can be written to step around it deliberately and the specification can be written to explain why the difference matters technically.

An application drafted with the closest prior art in front of the drafter looks completely different from one drafted without it. The first anticipates the objection; the second discovers it eighteen months later, by which time amendment options are constrained by what was disclosed on filing.

What a search can and cannot tell you

No search is exhaustive. Applications are not published for eighteen months, so there is always a window of pending filings nobody can see. Non-patent literature - theses, standards documents, forum posts, product manuals - is poorly indexed. A clean search is evidence, not a guarantee.

What it does give you is a realistic view of the field and, often, an early answer to a more useful question: is this invention worth the next five years of fees and attention?