Protecting an Idea in the Window Before Filing

There is a period, usually a few months long, in which an invention exists and no application covers it. Nothing in the file has a date at the granting office. Nothing is pending. And during that period the inventor generally needs to talk to people — a machinist about tolerances, a supplier about materials, a potential buyer about whether anyone would want the thing at all. Each of those conversations is either harmless or catastrophic, and the difference is procedural rather than a matter of good faith.

The mechanism is novelty. A patent can only be granted for something not already available to the public, and the public in this sense includes anyone under no obligation to keep quiet. Disclosure by the inventor counts. This is the single most counter-intuitive rule in the system: an inventor can single-handedly destroy their own patentability by describing their own work, without any competitor doing anything at all.

The Window

Why the Gap Exists at All

The gap is not laziness. Filing too early is a genuine error. An application is assessed on what it discloses on the day it is lodged, and a description written before the mechanism is understood produces a narrow, brittle document that cannot later be broadened without losing its date. Material added afterwards is new matter with a new date, and in a contested case the new date is often the difference between a valid claim and an invalid one.

So there is a real trade-off. Wait, and the design becomes clear enough to claim broadly — while the risk of an intervening disclosure, an independent filing by somebody else, or an accidental publication accumulates every week. File immediately, and the date is secured around something half-formed. Protecting an idea through this stretch means managing both sides at once rather than pretending only one exists.

Novelty

What Counts as a Disclosure

Disclosure is broader than publication. It does not require an audience, a sale, a transcript or anyone actually paying attention. The test in most systems is whether the information was made available to a person free to pass it on. On that test, all of the following have destroyed patent rights: a trade-stand demonstration; a crowdfunding page with an exploded diagram; a conference talk; a video showing the mechanism operating; a public thesis on a library shelf; an offer for sale of a working unit; a supplier drawing sent without terms attached.

Two nuances matter. First, grace periods. Some jurisdictions allow a limited period — commonly twelve months before filing — during which the inventor's own disclosure does not count against them. Others apply absolute novelty, where the first public description is prior art from the moment it appears, with no forgiveness at all. Relying on a grace period therefore forecloses filing in a large part of the world, and that consequence is usually discovered only when an applicant tries to extend protection abroad and is told the door shut a year ago.

Second, the experimental-use question. Testing a prototype in circumstances that necessarily expose it — a component tested in a working vehicle, a device trialled by users — may in some systems be excusable where the testing was genuinely necessary to complete the invention and was conducted under restraint. The exception is narrow, fact-specific and a poor thing to bet on. Where a trial is unavoidable, it should be run under written terms and documented as a trial, with its purpose recorded at the time.

Instruments

Agreements That Work, and Ones That Do Not

Confidentiality is the main instrument available before filing, and it works on one condition: it is in place before the information moves. An agreement signed after a meeting is a contract about the future conduct of somebody who already knows everything.

• Signed in writing, dated, before any technical detail is exchanged. A verbal understanding is enforceable in principle and unprovable in practice.
• Defining the confidential subject matter specifically enough to be enforceable, without reciting the invention in full in a document that will circulate widely.
• Naming a term long enough to matter — five years is common, and anything shorter than the expected time to filing is decorative.
• Covering the recipient's employees, contractors and advisers, since those are the people who will actually read it.
• Logged in the disclosure ledger with date, counterparty and scope, so the position can be reconstructed years later.

Some counterparties will not sign. Larger companies routinely refuse unsolicited confidential submissions, precisely to avoid claims that they took something. Where that happens, the options are to file first and disclose afterwards, or to disclose only at a level of generality that teaches nothing — the problem solved and the benefit delivered, never the mechanism that delivers it. Advisers to a professional practitioner are a different case: communications with a patent attorney are privileged in most systems, which is why the first substantive technical conversation should generally be that one.

Two instruments deserve dismissing. A posted, unopened envelope addressed to yourself proves a postmark and nothing else — not authorship, not completeness, not a filing date. And a copyright notice on a drawing protects the drawing as an image; it leaves the mechanism it depicts entirely free for anyone to build.

Filing

Closing the Window

The window closes with a filing, and the choice of filing determines what comes next. A provisional or similar priority application is never examined, never published and never becomes a patent; what it does is fix a date and permit honest use of the phrase patent pending for the twelve months that follow. Official fees for an individual filer typically run in the low hundreds, which is why it is treated as the cheap option — and why it is so often written as though cheapness were the point.

It is not. A later full application can claim the benefit of a priority filing only to the extent the earlier document actually described what is now being claimed. A thin provisional supports a thin sliver of the eventual claims, and the shortfall surfaces at the worst possible moment: after twelve months of disclosure made in reliance on it. Write it as though it were the full specification. The distinctions between the available routes are worth reading properly, and the granting offices set out the categories of application and proceeding in more detail than most first-time filers expect.

Three dates then govern everything. Twelve months from the priority filing, to convert to a substantive application and to decide on filings in other territories. Eighteen months from priority, when the application publishes and confidentiality ends by operation of law rather than by choice. And thirty or thirty-one months, where an international application is used to defer the decision about which national procedures to enter. Diarise each at nine months as well, because preparation, not filing, is the part that takes the time.

Sequence

The Order That Keeps It Open

Almost every avoidable loss in this stage is an ordering failure rather than a knowledge failure. Run properly, the sequence is: write the description fully, search the prior art, take privileged advice, file the priority application, and only then begin the commercial conversations that need detail. Run in the order most people actually use — talk to a manufacturer, build a prototype, put up a page to test demand, then think about filing — and the right is often gone before anyone consults anyone.

Documentation is what makes the sequence auditable. Every disclosure logged with a date and a counterparty; every agreement filed with the version that was signed; every decision to proceed without an agreement recorded with the reason. Kept that way, the pre-filing period leaves a clean record instead of a set of unanswerable questions, and the case for a durable archive is set out in full in the inventor's evidence file.

There is a broader point about pace. Ideas that reach a market almost never do so on the first geometry — accounts of how invention ideas actually become products describe long stretches of revision, and devices built around measured physical performance, such as vibration-based pain relief wearables, spend much of that time producing the test data the eventual claims rest on. That is exactly the material the window is for. Protecting an idea is not a matter of secrecy for its own sake; it is a matter of keeping the option to file intact for as long as the technical work needs, and then filing on something worth defending.

Describe it fully, search it, file it, then talk. The order is the whole discipline.