Section 112 Compliance Checklist: Enablement, Written Description, Definiteness, and Means-Plus-Function

By Casey Scott McKay · 2024-11-23

Section 112 is the quietest way a valuable patent dies, and almost all of it is decided before filing. This checklist runs the analysis in twelve phases: map claim scope to disclosure, audit functional limitations for disclosed algorithms, verify genus support, check definiteness term by term, sweep for admissions and absolutes, verify priority support, then respond to a Section 112 rejection, and finally read an issued patent for exposure whether you own it or are defending against it. Each box states why it exists, the authority, and the trap. A worked matter follows one application from drafting through an office action. The companion article explains the doctrine and the companion guide explains how to draft against it.

IP and Technology > Patent Counseling Transactions | Checklist | Published 23 November 2024 - Updated 30 September 2025 | Casey Scott McKay - marksy.us

Summary. Section 112 is the quietest way a valuable patent dies, and almost all of it is decided before filing. This checklist runs the analysis in twelve phases: map claim scope to disclosure, audit functional limitations for disclosed algorithms, verify genus support, check definiteness term by term, sweep for admissions and absolutes, verify priority support, then respond to a Section 112 rejection, and finally read an issued patent for exposure whether you own it or are defending against it. Each box states why it exists, the authority, and the trap. A worked matter follows one application from drafting through an office action. The companion article explains the doctrine and the companion guide explains how to draft against it.

Keywords: Section 112 checklist, enablement, written description, definiteness, means plus function, algorithm disclosure, support map, claim scope, genus claims, terms of degree, antecedent basis, prophetic examples, priority support, continuation claims, prosecution response, invalidity analysis, patent drafting, specification review, nonce words, claim ladder

How to use this checklist

| Phase | What it covers | When | |---|---|---| | 1 | The support map | Before filing | | 2 | The algorithm audit | Before filing | | 3 | Genus and scope support | Before filing | | 4 | Definiteness, term by term | Before filing | | 5 | The admissions and absolutes sweep | Before filing | | 6 | Examples and prophetic disclosure | Before filing | | 7 | Priority support | Before filing and at every continuation | | 8 | Responding to a Section 112 rejection | On office action | | 9 | Amending safely | On office action | | 10 | Continuation claims | Whenever new claims are drafted | | 11 | Reading an issued patent for exposure | Diligence and defense | | 12 | Portfolio review | Annually |

Phases 1 through 7 run before filing and are the whole game. Everything after is either recovery or diagnosis. Boxes marked [Gate] should clear before filing.

The matter. Ostrander Systems filed on a streaming-analytics improvement. The independent claim recited "an evaluation module configured to determine whether a record satisfies a threshold condition." The specification described the module in three sentences and said the determination used "techniques known in the art." Every problem in this checklist appears in that one claim.

Phase 1. The support map

[ ] [Gate] List every claim limitation in one column and the supporting paragraph in the next.
- Why. It is the last moment at which an unsupported limitation can be fixed. After filing, new matter cannot be added.
- Authority. 35 U.S.C. § 112(a); 35 U.S.C. § 132(a).
- Trap. Assuming support exists because the drafter knows what was meant.
[ ] Resolve every limitation with no paragraph.
- Why. Add the description or remove the limitation; both are possible now and neither is possible later.
- Trap. Filing with a known gap and planning to argue it.
[ ] Map each limitation to a figure where one exists.
- Why. Drawings support written description independently, and a structure shown but not described in words can still evidence possession.
- Authority. 37 C.F.R. § 1.83.
- Trap. Figures showing elements that appear in no claim and claims reciting elements in no figure.
[ ] Confirm every claim term appears in the description with the same meaning.
- Why. A claim term absent from the specification is a written-description problem and a construction opportunity for the other side.
- Trap. A term introduced during claim drafting and never back-filled into the description.
[ ] Retain the map with the file.
- Why. The prosecutor arguing support for an amendment three years from now needs it.
- Trap. A map built once and discarded.

Phase 2. The algorithm audit

[ ] [Gate] List every claim limitation phrased functionally.
- Why. Each is a candidate for construction under 35 U.S.C. § 112(f), which limits it to the disclosed structure and equivalents — or renders it indefinite if no structure was disclosed.
- Trap. Assuming the provision applies only where the word "means" appears. It does not.
[ ] Flag every nonce word.
- Why. Module, unit, element, mechanism, component, device, logic, and engine supply no structure and are treated as placeholders.
- Trap. "Engine" and "logic," which feel technical and are not structural.
[ ] Match each functional limitation to a disclosed algorithm.
- Why. For a computer-implemented function, the corresponding structure is the algorithm, not the processor.
- Authority. 35 U.S.C. § 112(b); 35 U.S.C. § 112(f).
- Trap. "Performed using known techniques," which discloses nothing.
[ ] Confirm each algorithm is expressed in an acceptable form.
- Why. A flowchart, pseudocode, step-by-step prose, or a formula with defined inputs and outputs all suffice. Source code is not required.
- Trap. A block diagram with labeled boxes and no internal description.
[ ] Cross-reference each algorithm to the claim term explicitly.
- Why. "The segmentation engine 240 performs the process of FIG. 4" makes the correspondence unmistakable.
- Trap. An algorithm present somewhere in the specification with no link to the claimed element.

Ostrander, Phase 2. The audit found one functional limitation, one nonce word, and zero algorithms. The fix before filing was a flowchart, a paragraph describing the comparison and the threshold derivation, and a claim amendment reciting the operations rather than a module configured to perform them.

Phase 3. Genus and scope support

[ ] Ask what taught the full scope, not the embodiment.
- Why. Enablement requires the disclosure to teach the entire breadth claimed without undue experimentation.
- Authority. 35 U.S.C. § 112(a).
- Trap. A single embodiment supporting a claim covering a category.
[ ] [Gate] For any genus claim, confirm the genus is defined structurally rather than functionally.
- Why. A genus defined by what members do, in an unpredictable art, requires a research program to practice — which is undue experimentation.
- Trap. The beautiful functional claim, which is the most reliable way to lose a valuable patent.
[ ] Confirm representative species span the claimed variation.
- Why. Representative means distributed across the range, not merely numerous.
- Trap. Twelve species clustered at one end of a claimed range.
[ ] Assess predictability of the art honestly.
- Why. Mechanical and electrical arts support broad extrapolation from one embodiment; chemistry and biology do not.
- Trap. Applying software drafting instincts to a life sciences application.
[ ] Check ranges for intermediate disclosure.
- Why. A claim to a sub-range carved from a broad disclosure, with nothing described in between, is vulnerable on written description.
- Trap. Reciting only the broadest range and the single working value.

Phase 4. Definiteness, term by term

[ ] List every term of degree and locate its baseline.
- **Why. **"Substantially," "about," "approximately," "sufficient," and "optimal" require something in the specification supplying an objective measure.
- Authority. 35 U.S.C. § 112(b).
- Trap. A term of degree the drafter considered obviously bounded.
[ ] Flag every subjective term.
- Why. Scope that depends on taste has no reasonably certain boundary.
- Trap. "User-friendly," "efficient," "aesthetically pleasing."
[ ] Check for mixed statutory classes within a claim.
- Why. An apparatus claim reciting a method of use leaves infringement undefined.
- Trap. "A system comprising ... wherein a user performs ..."
[ ] Check antecedent basis throughout.
- Trap. "The controller" with no prior "a controller."
[ ] Check negative limitations for support.
- Why. Claiming the absence of a feature requires a basis in the disclosure for excluding it.
- Trap. A negative limitation added in prosecution to avoid art.
[ ] Confirm any special definition is stated clearly.
- Why. A patentee may be its own lexicographer, and a clear definition controls over ordinary meaning.
- Trap. A definition that binds you more than it helps. See What the Claim Means.

Phase 5. The admissions and absolutes sweep

[ ] Search for "conventional," "well known," "standard," "routine," "typical," "commonly."
- Why. Each is an admission that a challenger will quote at Alice step two and an examiner will cite for conventionality.
- Authority. See What Can Actually Be Patented.
- Trap. Boilerplate describing the computing environment as conventional, which is the most common self-inflicted eligibility wound.
[ ] Search for "must," "always," "requires," "essential," "critical," "necessary."
- Why. Each is a construction limitation waiting to be imposed.
- Trap. "The invention requires," which defines the invention as requiring it.
[ ] Search for "the present invention is."
- Why. It characterizes the whole invention and is quoted to narrow every claim.
- Trap. A summary written in that form as a stylistic habit.
[ ] Check the background for prior-art admissions.
- Trap. Identifying a specific reference as prior art unnecessarily.
[ ] Check the abstract for narrowing characterizations.
- Trap. An abstract written last, describing only the preferred embodiment.

Phase 6. Examples and prophetic disclosure

[ ] Include working examples where the art is unpredictable.
- Why. They are the strongest evidence of both enablement and possession.
- Trap. A chemical application with no examples.
[ ] Include comparative data against the closest known alternative.
- Why. It supports unexpected results for nonobviousness later, and the comparison must be against the closest prior art.
- Authority. 35 U.S.C. § 103; declarations under 37 C.F.R. § 1.132.
- Trap. Comparing against a convenient baseline.
[ ] [Gate] Write prophetic examples in present or future tense.
- Why. Describing experiments not performed as though they were conducted has been treated as a candor problem.
- Trap. A template that renders all examples in past tense.
[ ] Label examples clearly as actual or prophetic where the distinction matters.
- Trap. An examiner or a court discovering the distinction from a deposition.

Phase 7. Priority support

[ ] [Gate] Read the provisional against the claims that rely on it.
- Why. A provisional supports only what it describes, judged under 35 U.S.C. § 112(a). A claim not supported does not get the earlier date.
- Authority. 35 U.S.C. § 119(e).
- Trap. A slide deck filed as a provisional, relied on for claims drafted a year later. See The Priority Chain.
[ ] Verify the cross-reference to related applications is present and correct.
- Authority. 35 U.S.C. § 120.
- Trap. An omitted or defective priority reference, which silently breaks the chain.
[ ] Re-run the support map against the earliest application relied on.
- Trap. Support that exists in the parent and not in the grandparent, where the claim needs the earlier date.
[ ] Record which claims depend on which priority date.
- Why. It determines which art applies to which claim.
- Trap. A family where nobody can say what date any given claim gets.

Phase 8. Responding to a Section 112 rejection

[ ] Identify which of the four requirements the examiner invoked.
- Why. Enablement, written description, definiteness, and means-plus-function indefiniteness are different failures with different cures, and a response treating them as one addresses none.
- Trap. A generic traverse.
[ ] For enablement, address the factors the analysis uses.
- Why. Quantity of experimentation, guidance provided, working examples, nature of the invention, state of the art, skill level, predictability, and claim breadth.
- Trap. Arguing that the embodiment works, which is not the question.
[ ] For written description, point to possession evidence.
- Why. Structure, formula, distinguishing characteristics, or representative species.
- Trap. Arguing enablement in response to a written-description rejection.
[ ] For definiteness, supply the specification's baseline or amend.
- Trap. Asserting that the term is clear to a skilled artisan without citing anything.
[ ] For a means-plus-function indefiniteness rejection, identify the algorithm by figure and paragraph.
- Why. If it is there, the rejection resolves. If it is not, the claim must be amended to recite the operations.
- Trap. Arguing the provision does not apply, when amending to recite structure resolves the issue outright.
[ ] Respond to every ground and every claim.
- Authority. 37 C.F.R. § 1.111.

Phase 9. Amending safely

[ ] Confirm support for every amendment before filing it.
- Authority. 35 U.S.C. § 132(a).
- Trap. A new-matter objection that costs a round.
[ ] Prefer elevating a dependent limitation to drafting new language.
- Why. Support is established and estoppel is more contained.
- Trap. Novel claim language invented at the response stage.
[ ] Recite operations rather than modules configured to perform them.
- Why. It avoids the construction and the eligibility problem at once.
[ ] Cite the supporting paragraph in the remarks.
- Trap. Leaving the examiner to find support, which invites an objection.
[ ] Record what the amendment surrenders.
- Why. It limits equivalents later and will be quoted at claim construction.
- Authority. See Proving Patent Infringement.

Phase 10. Continuation claims

[ ] [Gate] Run the support map afresh for every new claim.
- Why. Claims drafted years later against a competitor are the ones most likely to lack support, because they were written against a product rather than against the disclosure.
- Authority. 35 U.S.C. § 120.
- Trap. Assuming the parent's allowance validates the family.
[ ] Confirm co-pendency before the parent issues.
- Trap. Paying the issue fee first, which closes the family permanently.
[ ] Check whether new claims trigger a different priority date.
- Trap. A continuation claim supported only by a later-filed parent, exposed to intervening art.
[ ] Watch for double patenting and the terminal disclaimer consequence.
- Why. A terminal disclaimer ties term to the earlier patent and conditions enforceability on common ownership.
- Authority. 37 C.F.R. § 1.321.
- Trap. Filing a disclaimer reflexively in a family whose members may be sold separately.

Phase 11. Reading an issued patent for exposure

[ ] Count the functional limitations in each independent claim.
[ ] Measure claim breadth against described embodiments.
[ ] Check for working examples where the art is unpredictable.
[ ] Search the specification for each claim term.
[ ] Check where the claims came from — original, added in prosecution, or first in a continuation.
[ ] Read the provisional if the patent relies on it.
[ ] Check terms of degree against the specification.
[ ] Record the result as a risk-adjusted view, not a verdict.
- Why. The presumption of validity is real and the challenger's burden is heightened.
- Authority. 35 U.S.C. § 282.
- Trap. Telling a client a patent is invalid when what you mean is that it has exposure. See IP Due Diligence Toolkit.

Phase 12. Portfolio review

[ ] Flag patents whose claims are materially broader than their examples.
[ ] Flag patents with functional limitations and no disclosed algorithm.
[ ] Confirm a continuation remains pending in each significant family.
- Authority. 35 U.S.C. § 120.
[ ] Consider reissue for a correctable defect, within the limits.
- Why. Broadening reissue is available only within two years of grant; narrowing is available at any time; new matter can never be added, and the recapture rule bars regaining surrendered subject matter.
- Authority. 35 U.S.C. § 251.
[ ] Distinguish repairable defects from terminal ones.
- Why. Antecedent basis and terms of degree are fixable; a missing algorithm or an undescribed genus is not.
- Authority. 35 U.S.C. § 254; 35 U.S.C. § 255.
[ ] Feed the result into the maintenance-fee decision.
- Authority. 35 U.S.C. § 41(b). See Patent Portfolio Management Toolkit.

Ostrander, outcome. The pre-filing audit caught all of it. The claim was rewritten to recite the comparison operations, a flowchart and a supporting paragraph were added, the "known in the art" sentence was deleted, and a dependent claim reciting the threshold derivation was added as the allowance path. The application received a §102 rejection and no §112 rejection. Had the original claim been filed, the algorithm gap would have surfaced in litigation eight years later, where it could not have been fixed.

Phase 13. Technology-specific Section 112 screens

The general phases apply everywhere; the pressure points differ by field.

Software and computer-implemented inventions

[ ] Treat the algorithm audit at Phase 2 as mandatory rather than advisory.
- Why. It is the dominant failure mode in this field by a wide margin.
- Trap. A specification that describes what the system accomplishes and never how.
[ ] Describe data structures explicitly — fields, layout, invariants.
- Why. A data structure is structure in a way that a process step is not, and it supports both definiteness and eligibility.
- Trap. "A database" as the only description of the storage layer.
[ ] Describe the system at more than one layer.
- Why. A continuation may need to claim the client, the server, or the protocol, and each must be described to be claimable.
- Trap. A single-perspective disclosure that forecloses two thirds of the family.
[ ] Describe what is externally observable.
- Why. A claim to an unobservable internal process is difficult to prove infringed.
- Trap. Claiming only server-side steps nobody outside the defendant can see.

Chemistry, formulations, and materials

[ ] Confirm the genus is structural, with enumerated substituent groups.
- Authority. 35 U.S.C. § 112(a).
- Trap. A genus defined by activity.
[ ] Confirm working examples span the claimed variation at each variable position.
- Trap. Examples clustered at one corner of the claimed space.
[ ] Confirm ranges are disclosed at broad, intermediate, and specific levels.
- Trap. A sub-range claimed in prosecution with no intermediate support.
[ ] Confirm comparative data uses the closest prior art.
- Authority. 37 C.F.R. § 1.132.

Biotechnology

[ ] Confirm sequence listing compliance in the required format.
- Why. A defect here is a filing-date risk rather than a formality.
- Trap. Treating it as a clerical step delegated without review.
[ ] Confirm structural characterization of any claimed genus.
- Trap. An antibody genus defined by binding and function alone.
[ ] Confirm deposit requirements where biological material cannot be described adequately in writing.
- Trap. A deposit made after filing, or not made at all.

Mechanical and electromechanical

[ ] Confirm figures show every claimed element with a reference numeral.
- Authority. 37 C.F.R. § 1.84.
- Trap. Claim elements visible in no drawing.
[ ] Confirm tolerances, materials, and assembly relationships are described.
- Trap. Relying on figures for dimensional relationships that the claims recite numerically.
[ ] Focus the review on definiteness rather than enablement.
- Why. Predictable arts support broad extrapolation from one embodiment; terms of degree remain the live risk.

Medical devices and diagnostics

[ ] Separate device claims from method-of-use claims into distinct families.
- Why. They face different eligibility exposure and different infringement realities, including divided infringement where a physician performs steps.
- Authority. See Proving Patent Infringement.
- Trap. One application carrying both, where a problem with either infects the family.

Phase 14. Cost and sequencing

| Step | Relative cost | When | |---|---|---| | Support map | Low — two hours | Before filing | | Algorithm audit | Low — one hour | Before filing | | Admissions and absolutes sweep | Very low — automated search | Before filing | | Second-drafter cold read | Low — one hour | Before filing | | Adding described alternatives | Moderate | Before filing | | Working examples and comparative data | High, and generated by the science | During development | | Responding to a §112 rejection | Moderate | Prosecution | | Reissue to narrow a defective claim | Moderate to high | Post-grant, within limits | | A §112 invalidity fight in litigation | Very high | Avoidable at drafting | | An unsupported genus claim held invalid | Total loss of the claim | Not repairable |

The ratio worth stating. The four pre-filing checks — support map, algorithm audit, admissions sweep, cold read — total about five hours and address the overwhelming majority of Section 112 exposure. The alternative is discovering the same problems in litigation, on a developed record, against an adversary with experts, at a point when the specification has been unchangeable for a decade.

And the sequencing rule. Claims, then specification against the claims, then the four checks, then file. Applications drafted in that order rarely see a Section 112 rejection. Applications drafted specification-first routinely do, and the ones that do not are simply the ones where nobody has looked hard yet.

Phase 15. Using Section 112 offensively

The checklist so far assumes you own the patent. Defending against one runs the same analysis with the opposite objective, and it is frequently the cheapest available defense.

[ ] Identify every functional limitation in the asserted claims.
- Why. Each is a candidate for a means-plus-function construction limiting the claim to the disclosed algorithm and its equivalents — which usually excludes the accused product.
- Authority. 35 U.S.C. § 112(f).
- Trap. Assuming the argument is unavailable because the claims avoid the word "means."
[ ] Search the specification for structure corresponding to each one.
- Why. No disclosed algorithm means indefiniteness, which invalidates rather than narrows.
- Authority. 35 U.S.C. § 112(b).
- Trap. Accepting a block diagram as structure.
[ ] Compare claim breadth against the described embodiments.
- Why. Claims materially broader than the disclosure are the enablement and written-description target.
- Trap. Attacking breadth in a predictable art, where extrapolation is accepted.
[ ] Trace when each asserted claim was drafted.
- Why. Claims added late in prosecution or first appearing in a continuation are the most likely to lack support, because they were written against a product rather than against a disclosure.
- Authority. 35 U.S.C. § 120.
- Trap. Reading only the issued claims and not the file history.
[ ] Read every application in the priority chain.
- Why. If the claim needs the provisional's date to avoid intervening art, and the provisional does not describe it, the claim loses the date and the art comes in.
- Authority. 35 U.S.C. § 119(e).
- Trap. Assuming the priority claim is good because it was accepted.
[ ] Choose the forum deliberately.
- Why. Section 112 cannot be raised in an inter partes review, which is confined to novelty and obviousness on patents and printed publications. It can be raised in a post-grant review within nine months of issuance, and always in court.
- Authority. 35 U.S.C. § 311(b); 35 U.S.C. § 321(c); 35 U.S.C. § 282.
- Trap. Planning a single-forum defense when the strongest grounds split across two. See PTAB Practice Toolkit; Patent Assertion Defense Toolkit.
[ ] Raise the construction argument at the Markman stage, not later.
- Why. A claim narrowed to a disclosed flowchart is frequently a claim the defendant does not infringe, and that outcome is available at claim construction rather than at trial.
- Authority. See Litigating Claim Construction; Claim Construction Checklist.
- Trap. Treating Section 112 as a trial issue when most of its value is realized in construction.
[ ] Be candid with the client about the burden.
- Why. Invalidity must be established against a presumption, and expert testimony cuts both ways.
- Authority. 35 U.S.C. § 282.
- Trap. Promising a defense that is a strong argument rather than a certainty.

Phase 16. Building the checks into the practice

The individual boxes matter less than whether they run every time, which is a workflow question rather than a legal one.

[ ] Make the support map a docketing item, not a good intention.
- Why. Checks that depend on the drafter remembering are performed on the applications with the most time and skipped on the ones under deadline pressure — which are the applications that need them.
- Trap. A firm standard that exists in a memo and not in the file-opening template.
[ ] Automate the admissions and absolutes sweep.
- Why. It is a text search for a fixed word list, it takes under a minute, and it catches the most common self-inflicted eligibility and construction wounds.
- Trap. Doing it by reading, which misses instances buried in boilerplate.
[ ] Assign the cold read to someone who did not draft.
- Why. The drafter cannot see the gap between what they meant and what they wrote; a colleague reading the claims against the specification finds it in an hour.
- Trap. A self-review labeled as a second review.
[ ] Keep the support map with the file permanently.
- Why. It is needed at every amendment, at every continuation, and in any later validity analysis.
- Trap. A working document deleted after filing.
[ ] Re-run Phases 1, 2, and 7 at every continuation.
- Authority. 35 U.S.C. § 120.
- Trap. Treating the parent's issuance as validating the family's support.
[ ] Record which checks were run and by whom.
- Why. It is the contemporaneous evidence of diligence, and it is the institutional memory when a different lawyer picks up the family in six years.
- Trap. No record, and a family nobody can reconstruct.

A closing note on who this checklist is for. Phases 1 through 7 belong to the drafter and run before filing. Phases 8 through 10 belong to the prosecutor. Phase 11 belongs to whoever does diligence or defends an assertion. Phase 12 belongs to the portfolio owner who signs maintenance-fee decisions. The single most common organizational failure is that Phases 1 and 2 belong to nobody in particular — they sit between drafting and filing, they take a few hours, and under deadline they are the first thing cut. Putting them in the docket rather than in a practitioner's judgment is what makes them happen.

Key Authorities at a Glance

| Authority | Proposition | Phase | |---|---|---| | 35 U.S.C. § 112(a) | Written description, enablement, best mode | 1, 3, 7 | | 35 U.S.C. § 112(b) | Definiteness | 2, 4 | | 35 U.S.C. § 112(f) | Means-plus-function construction | 2 | | 35 U.S.C. § 101 | Eligibility; the admissions sweep serves it too | 5 | | 35 U.S.C. § 103 | Nonobviousness; comparative data | 6 | | 35 U.S.C. § 119(e) | Benefit of a provisional | 7 | | 35 U.S.C. § 120 | Continuations; benefit only for what was described | 7, 10, 12 | | 35 U.S.C. § 132(a) | No new matter | 1, 9 | | 35 U.S.C. § 251 | Reissue and its limits | 12 | | 35 U.S.C. § 254 | Certificate of correction, Office mistake | 12 | | 35 U.S.C. § 255 | Certificate of correction, applicant mistake | 12 | | 35 U.S.C. § 282 | Presumption of validity; invalidity as a defense | 11 | | 35 U.S.C. § 41(b) | Maintenance fees | 12 | | 37 C.F.R. § 1.75 | Claim form | 4 | | 37 C.F.R. § 1.83 | Content of drawings | 1 | | 37 C.F.R. § 1.111 | Reply responsive to every ground | 8 | | 37 C.F.R. § 1.132 | Declarations | 6 | | 37 C.F.R. § 1.321 | Terminal disclaimers | 10 |

The five things people get wrong

One: they draft the specification before the claims. A specification written first describes what the inventor built. One written against the claims describes what the claims require, and only the second reliably supports them.

Two: they leave functional limitations without algorithms. For a computer-implemented function, the structure is the algorithm. This single omission produces more software invalidity holdings than any other drafting error, and it takes an hour to prevent.

Three: they claim a functional genus in an unpredictable art. The claim covering every compound with a property, supported by a dozen species, is attractive, sometimes issues, and does not survive. Claim the structural genus instead and accept the narrower scope, which is worth more because it holds.

Four: they never run the support map. Two hours before filing, listing each limitation against its supporting paragraph, is the last moment anything can be fixed. Almost nobody does it, and it is the highest-return step in the entire process.

Five: they assume issuance means support. Examiners have limited time and no adversary; litigation defendants have budget, experts, and motive. An issued claim broader than its disclosure has deferred the fight, not won it. See Drafting a Patent Specification That Survives.