Novelty is assessed against a single instant: the effective filing date. Everything made available to the public before that instant sits in the prior art, and it makes no difference to the analysis who put it there. An inventor's own conference talk is as damaging to their own claims as a competitor's patent would be — more so, in fact, because it describes exactly the thing being claimed.
This is the least intuitive rule in the field, and the one that produces the most irreversible losses. It cuts against every instinct a person has when they have built something that works. The impulse is to show it: to a manufacturer, to a buyer, to a forum, to anyone who will confirm the thing is as good as it feels. The rule says the showing must come second.
§ 01 Definition
What the Law Counts as Telling Someone
The statutory categories are wide and deliberately open-ended: patented, described in a printed publication, in public use, on sale, or otherwise available to the public. Courts have read that final catch-all generously. A disclosure need not have been noticed, read or understood by anyone; it needs only to have been accessible to a person interested in the subject who exercised ordinary diligence.
In practice that captures a longer list than most inventors expect. A demonstration at a trade fair. A crowdfunding page with drawings. A dissertation shelved in a university library and indexed by subject. A product listing taken down after a week. A poster session. A pitch deck circulated to twenty potential partners without a confidentiality agreement in place. A video posted for an audience of eleven. None of these requires publication in any formal sense; each can put the invention beyond reach in most of the world. This is one of the five recurring failures set out in the wider account of how applications are lost before examination, and it is the one with the shortest path from mistake to consequence.
§ 02 Geography
Twelve Months in One Place, None in Most Others
Some jurisdictions provide an exception: a disclosure made by the inventor, or by someone who obtained the subject matter from the inventor, does not count against the application if the filing follows within twelve months. That exception is the public disclosure grace period, and it is a genuine safety net — narrow, but real. It also does something less widely understood: a qualifying disclosure by the inventor can, within that window, block a later independent disclosure by somebody else from being used against them.
Many other territories have no equivalent. It applies absolute novelty, and the only exceptions are an evident abuse against the applicant and display at a small class of officially recognised international exhibitions, both subject to a six-month window and formal requirements. A few territories offer limited exceptions that require a declaration filed alongside the application and supporting documents within a fixed period; they are recovery mechanisms with paperwork, not a general licence to publish. Several other major systems offer nothing.
The grace period is not a period of protection. It is a period during which one jurisdiction agrees not to hold your own words against you, while every other jurisdiction already has.
On what the exception actually buys
The arithmetic that follows is unpleasant. An inventor who demonstrates a product and files nine months later is inside the domestic window and outside every foreign one. They may still obtain a domestic patent. They have also removed the international rights that make a licence worth signing, because a manufacturer weighing a category rarely wants a right that stops at a single border. Practitioners who have spent decades working alongside independent inventors report the same conversation repeatedly: the invention is sound, the domestic filing is intact, and the valuable half of the portfolio was given away at a trade stand.
§ 03 Commerce
The Sale That Does Not Have to Be Public
The on-sale bar operates independently of whether anybody learned how the invention works. A commercial offer for sale — a quotation, a supply agreement, a purchase order for units not yet built — can place the invention on sale even where the counterparty was bound to secrecy and the technical details never left the room. What matters is that the invention was the subject of a commercial transaction and was ready for patenting, meaning either reduced to practice or described well enough for someone skilled in the field to build it.
This traps a specific and sympathetic category of inventor: the one who behaves carefully. They sign a confidentiality agreement, they avoid publicity, they negotiate quietly with a contract manufacturer, and they treat the resulting silence as protection. Confidentiality prevents the disclosure limb from being triggered. It does not prevent the on-sale limb. A licence negotiation is generally safer than a supply agreement here, because a licence transfers rights rather than goods, but the distinction is fine enough that the sequence matters more than the label. File first; negotiate afterwards.
§ 04 Exception
Testing in the Open Without Losing the Right
There is a long-standing exception for genuine experimentation. The canonical case involved a wooden pavement laid in a public road and observed for six years; the use was held experimental rather than public because the inventor was testing durability under real conditions and retained control over the installation throughout. The doctrine survives, and it is narrower than its reputation.
What supports it is evidence of an actual testing purpose: records of the parameters under observation, control over who had access and what they could do, an obligation of confidentiality where practicable, no charge for the article, and a programme that ended when the technical question was answered. What defeats it is anything resembling commerce or promotion — taking payment, seeking orders, gathering marketing feedback, or continuing long after the data was in. Crucially, the doctrine is applied by a court years later, on the strength of documents created at the time. An inventor relying on it without a contemporaneous record is relying on nothing.
§ 05 Practice
An Order of Operations That Survives Contact
The workable sequence is short. Write the description first, in full technical detail, because it is needed for the filing in any event and because writing it usually exposes the gaps that a demonstration would have exposed publicly. File before any external conversation that is not covered by a signed confidentiality agreement. Keep a dated log of every disclosure made after filing — to whom, when, what was shown — because the twelve-month window is counted from the earliest one, not the most recent, and reconstructing that date from memory two years later is guesswork.
Two habits are worth adopting alongside. First, treat the public disclosure grace period as an emergency measure rather than a plan; the number of inventors who used it deliberately and well is small, and the number who discovered it after the fact is very large. Second, note that the rules bite hardest in fields where the work is naturally public. Environmental and green technology projects, which attract grant reporting and conference presentation, and design-led work such as inventor-led architectural innovation, where the output is meant to be seen, both run this risk structurally. In those fields the filing calendar has to be built around the disclosure calendar, not the other way round.
None of this asks an inventor to work in secret indefinitely. It asks for one reordering: the description exists, and is filed, before the demonstration happens. Everything that makes the grace period matter follows from getting those two events in the wrong order.
End of sheet 05a