Treat every patent question as a sequence of gates. Gate 1, novelty: one prior-art document against every claim feature, no mixing sources. Gate 2, inventive step: the difference plus a motivation argument. Gate 3, construction: list and interpret each claim feature. Gate 4, comparison: map features onto the accused product element by element. For amendments, trace every added feature back to the specification text before filing the response. Readiness checks: state both prior-art gates without notes, complete a five-element claim map with every entry labelled, draft an amendment with zero new features, and name the duty at stake in ethics vignettes. Administrative details for this credential, such as dates, formats, and fees, are set by the Professional Standards Board for Patent and Trade Marks Attorneys (Australia) and should be confirmed directly with the issuer.
Why novelty and inventive step are two separate gates
Novelty asks whether one prior-art disclosure shows every claim feature. Inventive step asks whether the difference over that disclosure is obvious. The gates use different evidence and different tests, so analysing them together produces attacks that collapse under their own logic.
A novelty attack lives or dies on a single document. You must point to one prior-art disclosure that shows each feature of the claim, either expressly or as something that disclosure inevitably makes available. Pulling a missing feature from a second document to patch the gap is the recognised error of mosaicing, and it converts a novelty argument into an obviousness argument without saying so. Track the features one by one against one document before reaching any conclusion at all.
Scenario: a claim recites features A, B, and C. Document 1 shows A and B; Document 2 shows C. The tempting move is to declare the claim not novel because every feature appears somewhere in the art. The better decision is to concede nothing on novelty against Document 1, since C is missing, and argue inventive step instead: would a skilled person, starting from Document 1 and facing the problem that C solves, adopt C from Document 2? The distinction matters because an obviousness attack needs a reasoned motivation, and the claim survives if that link cannot be drawn.
| Gate | Core question | Evidence allowed | If the attack fails |
|---|---|---|---|
| Novelty | Does one document show every claim feature? | A single prior-art disclosure, including what it inherently makes available | Move to the inventive-step gate; do not re-argue novelty |
| Inventive step | Is the difference over the closest art obvious to a skilled person? | Multiple documents plus common general knowledge, linked by motivation | The claim stands as inventive; revisit construction before attacking again |
| Mosaicing check | Am I combining documents to fill novelty gaps? | Not permitted at the novelty gate | Reserve all combinations for the obviousness analysis |
Parsing claim structure before you draft or attack it
A claim is one sentence whose punctuation carries meaning. Before evaluating validity or infringement, identify the preamble, the transitional phrase, each feature, and the antecedent basis linking them.
Three structural habits pay for themselves. First, mark the transitional phrase: open-ended language such as comprising and closed language such as consisting of set different boundaries, and their precise effect varies between jurisdictions, so verify the local rule in the issuer's materials. Second, check antecedent basis: every the should trace back to a preceding a or an, and a missing antecedent creates ambiguity you can name precisely. Third, map the dependency chain so you know exactly which limitations each claim actually carries.
Then rewrite the claim as a numbered feature list before any substantive conclusion. Features hidden inside clauses — such as a wherein clause defining a material or a relationship — are the ones that vanish from a rushed analysis, and dependency mistakes make a narrow claim look broad or the reverse. The feature list also becomes your working document for the novelty gate, the support check, and the infringement map, so building it early means every later step reads from the same accurate base rather than from memory.
When a broad claim outruns its specification
Claims may not extend beyond what the specification teaches. Enablement asks whether a skilled person could work the invention across the claimed scope; a disclosure gap can sink a claim that is perfectly novel.
The specification and the claims form a bargain: the disclosure teaches, and in exchange the claims define exclusive rights. A genus claim — say, a polymer defined by a composition rule covering many conditions — needs teaching that lets a skilled person achieve the result across that scope. A mini-scenario shows the trap: the description exemplifies one temperature range only, and a drafter claims all temperature ranges because the composition language permits it. The better decision is either to add teaching for the wider conditions or to narrow the claim to the range the disclosure actually supports, because a claim unsupported at its edges invites a sufficiency attack that novelty arguments cannot answer.
In drafting tasks, run a support check as a separate pass. Trace every feature of the independent claim to a specific place in the description and ask two questions: is the feature described, and is it described with enough teaching to justify the breadth used? A feature mentioned in a single passing list is described but weakly supported, which is a different problem from being absent altogether. Naming which problem you face determines the fix: more teaching, narrower claim language, or moving the feature to a dependent claim.
Choosing amendment versus argument in an office action response
Every response decision starts with the same rule: amend from the description, not from memory. Narrowing with a disclosed feature and arguing over the rest usually beats inventing new limitations under time pressure.
Scenario: an examiner rejects a claim as obvious over Document 1 combined with common general knowledge. The drafted response adds a limitation — a drainage channel in the base — that appears nowhere in the description as filed. That is the mistake: the amendment introduces new matter, which draws its own objection and can jeopardise the claim to the original filing date. The better decision is to narrow using a feature already disclosed, typically one carried by a dependent claim or a described embodiment, and add an argument explaining why the combination still fails, for instance because the prior art gives no reason to adopt the remaining distinguishing features.
Know the amendment types and their different risk profiles. Adding a limitation narrows scope but must trace to the disclosure. Deleting claims is low-risk and sometimes the whole answer. A disclaimer narrows by excluding subject matter and needs its own basis. Correction amendments fix errors but must not change what was disclosed. Before finalising any response, run the trace: find the exact text supporting each new word, and if the text is missing, convert that amendment into an argument instead.
Mapping claim elements onto an accused product
Infringement is an element-by-element comparison between a properly construed claim and the accused product. If even one essential element is absent, the usual conclusion is no infringement, whatever the overall resemblance.
Keep two steps separate. Construction first: decide what each feature means, especially functional language and relational terms, before looking at the product. Comparison second: place each claim feature against the product and label it present, absent, or unclear. The all-elements principle means absence of one essential feature usually ends the analysis. Also quarantine validity from infringement: a claim can be infringed yet invalid, and a valid claim can be uninfringed. An opinion that answers only one question should say so explicitly rather than blending the two into a single verdict.
Exercise: take any granted claim with at least five features and write a short product description embodying only four of them. Map every element as Present, Absent, or Unclear. Rubric: no element skipped; every Unclear entry names the construction question it turns on; you reach an actual conclusion instead of hedging. Expected observation: Unclear entries cluster on relational and functional features — how parts are coupled, what configured to covers — which shows exactly where your construction effort belongs in both exam answers and real opinions.
Paris Convention and PCT routes in scenario answers
Both routes preserve a first filing date, but they work differently: the Paris route means filing in each country during the priority period, while a PCT application defers country-by-country decisions and gives a search before those commitments.
The Paris Convention priority right lets a later filing in another country claim the date of the first filing, provided it happens within the priority period set by the treaty. The PCT route starts from a single international application that produces an international search and, if requested, a preliminary examination, with national or regional phase entry in each chosen country later. The practical trade-off is timing and information: the Paris route commits to each country early, while the PCT route lets the applicant read the search report before deciding where to spend national-phase money.
In scenario questions, build a date timeline first: first filing, any later filings, and whether each later filing claims priority. Then check which route the facts describe, because the two routes are easy to conflate — both preserve the first filing date, yet only the PCT route defers country choices and provides a search before those commitments. The classic error is treating a PCT application as if it were a granted worldwide patent — it is not; national phase and grant remain separate steps in each jurisdiction. Exact periods and national requirements vary by treaty rule and national law, so for any specific deadline in an answer, reason from the timeline structure and flag where the local rule must be checked.
Ethics fact patterns: name the duty before the facts
Every ethics question turns on identifying which professional duty the facts engage — candour to the IP office, client confidentiality, conflict of interest, or competence — then applying it to the specific conduct described.
Practise the duty-first pattern. A fact pattern showing a practitioner aware of a materially adverse prior-art document while making submissions to the office engages candour and honest dealing. A pattern showing a practitioner who moves firms and continues advising on an invention learned from a former client engages confidentiality and conflicts. A pattern showing advice outside the practitioner's technical competence engages competence and the duty to decline or refer. Naming the duty first disciplines the answer: the facts then serve to test whether the duty was breached, rather than inviting a vague essay about good conduct.
Preparation sequence, adaptable to your calendar: weeks one and two, prior-art gates plus daily claim-parsing drills; week three, drafting and the support-trace habit; week four, office action responses practised under time limits; week five, infringement mapping and the Paris-versus-PCT timelines; week six, ethics vignettes and one full timed practice paper. Closing readiness checks: state both prior-art gates without notes; complete a five-element claim map with every entry labelled Present, Absent, or Unclear; produce an amendment with zero unsupported features; name the duty at stake in ten ethics vignettes. These are learning milestones, not predictions, and administrative exam details should be confirmed with the issuer at psb.org.au.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.