Unit 8 — Seminar Notes, Referee Reports, and Responses to Criticism

1 Learning outcomes

This unit has the stable identifier O017-U08. After completing it, you will be able to:

1. turn a mathematical talk into a claim record with times, locations, types, grounds, verification statuses, and confidence levels;
2. mark inaudible or unrecorded passages without inventing replacement content;
3. distinguish summaries, verifications, questions, and objections, and separate a problem’s severity from confidence in its diagnosis;
4. write referee comments that are specific, bounded by the evidence, respectful of the author, and request feasible action;
5. assess recommendations to accept, request minor revisions, request major revisions, or reject on the basis of the manuscript’s state, not the author’s reputation or assumptions about the author;
6. respond to criticism by stating agreement or disagreement, action, evidence, the location of changes, status, and unresolved matters;
7. correct theorems and proofs without concealing the old wording or claiming a broader repair than has been proved; and
8. produce a seminar–review–response package that can be checked without contacting anyone outside the exercise or publishing anything.

2 Reviewing is an evidence contract, not a performance of certainty

Seminar notes, referee reports, and author responses serve different functions.

• Seminar notes record what can be observed at a particular time and location, together with the note-taker’s checks or questions.
• Referee reports assess a frozen manuscript within a stated scope, present evidence, and formulate revision requests.
• Author responses pair each comment with a decision, a change or a reason, evidence, a new location, a status, and remaining issues.

These three must not be conflated. The note “the speaker stated Theorem 2 on slide 6” establishes that the claim was heard or seen; it does not yet prove the theorem. A report saying “the third step does not follow from the hypotheses” must identify the step and give a reason; an assertive tone is not evidence. The response “corrected” is not auditable without the new wording and its location.

3 The recording contract: write what was observed, mark what is missing

Auditable seminar notes distinguish three layers.

1. Observation record: what is written on slides, the board, or accompanying materials, and what is heard during a particular segment.
2. Note-taker’s actions: summaries, checks, questions, or objections produced by the note-taker.
3. Post-seminar status: answered, verified, refuted, or still open after particular materials have been checked.

If a recording stops between 06:10 and 06:42, write “segment not recorded, 06:10–06:42”. Do not fill the gap with a sentence that seems to fit the next slide. If a symbol is inaudible, write “symbol inaudible; the slide only shows a letter that looks like $q$”. Do not turn a guess into a transcription.

4 Template for seminar notes with locators

Record the seminar’s identity before making inferences.

Then use one row per claim or note-taker action.

The location must be sufficient to find the content again. “In the middle of the seminar” is not enough if slide 5 or the right-hand board can be named. A timestamp alone is insufficient for material that has been re-edited; pair the time with a conceptual location.

5 Synthetic seminar case: one claim, one gap, one objection

The following package constitutes all the available seminar material.

On slide 5, the speaker defines $g(x)=f(x)-x$ and shows $g(0)=f(0)\geq0$ and $g(1)=f(1)-1\leq0$. The recording then enters the gap. When the recording resumes at 06:42, the right-hand board contains

$$0<f(y)-f(x)<y-x \qquad (x<y),$$

and slide 6 states “thus $g$ is strictly decreasing, so its zero is unique”. There is no Lipschitz hypothesis, contraction hypothesis, or other increment inequality on slides 1–6. During questions, a participant mentions $f(x)=x$; the speaker replies that the example needs to be checked after the seminar.

6 From notes to referee comments

An actionable referee comment has five components.

1. Locator: page, section, theorem, equation, or anchor.
2. Exact claim: the wording being assessed, without a strengthening invented by the referee.
3. Diagnosis and evidence: the failing step, counterexample, or missing dependency.
4. Bounded consequence: which parts are affected and which remain valid.
5. Request: an action that can be checked, not a demand to “fix everything”.

Compare the following two statements.

This proof is careless, and the author seems not to understand monotone functions.

That statement judges a person, gives no locator, and offers no test.

In Theorem 1, the third proof paragraph uses $f(y)-f(x)<y-x$ for $x<y$. Strict increase gives only $f(y)-f(x)>0$; the identity map satisfies every hypothesis but has many fixed points. Please retain the existence part, then revise or remove the uniqueness claim and audit the results that use it.

The second statement is firm, but its objects are the manuscript and its inferences. It makes no assumptions about the author’s ability, intentions, or motives.

7 A synthetic manuscript and the referee’s diagnosis

We now move from the seminar recording to a complete manuscript excerpt that freezes the claim and its proof.

P1 and P2 are valid. In P3, the first inequality $0<f(y)-f(x)$ follows from strict increase, but the second inequality $f(y)-f(x)<y-x$ does not. Being increasing does not bound the size of an increment relative to the interval’s length.

Take the identity map $f(x)=x$. This function is continuous, strictly increasing, and maps $[0,1]$ to $[0,1]$. However, $f(c)=c$ for every $c\in[0,1]$. The uniqueness claim in V1 is therefore false, while its existence claim remains true.

8 A nondefensive author response

A good response is not measured by the number of expressions of thanks. It is measured by whether readers can link comments to decisions and new results. For each comment:

1. state whether you agree, partly agree, or disagree;
2. paraphrase the diagnosis accurately;
3. identify the action taken or the evidence-based reason for not taking it;
4. show the evidence and the location of the change;
5. give an honest status; and
6. retain unresolved issues as explicit entries.

The response “we apologize for the confusion” is inadequate when the theorem is false. The response “this is only a notational issue” is also inadequate when there is a counterexample. Conversely, authors need not accept a suggestion stronger than necessary; declining it with a mathematical reason can be entirely professional.

9 A minimally revised theorem and proof

“Minimal” here means minimal relative to V1’s proof contract: the valid existence part is retained, while the uniqueness hypothesis is replaced by the exact property used to prove that $g$ has at most one zero. This is not a claim that the following wording is the weakest characterization among all possible formulations.

The new proof does not use strict increase of $f$. Removing an unused hypothesis is part of the revision, not a deficiency. If the author wants to give a more recognizable condition, a separate corollary may be added—for example, a particular inequality ensuring that $g$ decreases—but each corollary needs its own statement and proof.

10 The response matrix as a closure tool

A matrix does not replace the response letter; it prevents comments from getting lost between paragraphs. Suggested statuses are complete, partly complete, withheld, not adopted, with reasons, or open.

Each row must point to evidence or a limitation, not merely a promise. “Will be checked” is an open status; do not mark it complete. If a comment contains two requests that can be closed separately, split it into two rows.

11 Writing a referee report

A useful report enables an editor or author to make decisions without guessing the referee’s grounds. The following structure suffices for many mathematical manuscripts.

12 Four recommendations and their evidence gates

A recommendation describes the manuscript’s state relative to a stated scope and standard. It is not a judgment of the author.

V1 receives major revisions, not automatic “rejection”: the central uniqueness claim is false and dependent results need auditing, but the existence part is valid and V2 offers a checkable correction path. The number of comments does not determine the category. One counterexample to a central theorem may matter more than twenty typographical corrections.

Comment severity and the overall recommendation are not identical either. A major finding can be closed by a clear revision; a recurring set of local problems may indicate that the overall audit is not yet mature. Explain how the findings are combined.

13 Responding to disagreement

Authors may disagree with a comment. An auditable response has the following form:

1. state the part of the comment you understand;
2. state precisely which part you disagree with;
3. give local evidence or clarification;
4. explain whether the manuscript is nevertheless changed to prevent misreading;
5. give the location; and
6. allow the editor or referee to assess that evidence.

The following example responds to a comment about the finite geometric-series identity $\sum_{k=0}^{n}x^k=(1-x^{n+1})/(1-x)$ for $x\ne1$.

We agree that the old wording did not specify the domain of $n$. We disagree that the identity fails at $n=0$: the left-hand side is $x^0=1$, while the right-hand side is $(1-x)/(1-x)=1$ because $x\ne1$. In V2 we specify $n\in\mathbb N_0$ immediately before the identity and add the $n=0$ check in the following paragraph.

This response does not call the comment “nonsensical” or appeal to the author’s status. It answers with a domain, a calculation, and a location.

14 The complete seminar–report–response protocol

Use the following sequence.

1. Freeze the context. Record the seminar or manuscript identity, version, copy, time, location, and unavailable parts.
2. Separate observation from evaluation. Create summary entries before adding verifications, questions, or objections.
3. Mark gaps. Do not reconstruct unrecorded speech, symbols, or intentions.
4. Normalize claims. Use the quantifiers, domains, hypotheses, and conclusions actually written.
5. Check evidence. Use results from Unit 3, 6, or 7 where relevant; preserve grounds and boundaries.
6. Assess comments. Separate severity from confidence and state the reasons for each.
7. Write comments under the contract. Locator, claim, evidence, consequence, and request.
8. Formulate the recommendation. Combine comments by centrality, impact, and repair path; do not count comments.
9. Build the response matrix. One row per independently closable request.
10. Revise additively. Retain V1, create V2, and identify new evidence; do not change the history.
11. Audit remaining issues. Keep statuses open or withheld until evidence is available.
12. Cross-check. Ensure that the report, matrix, new manuscript, and unresolved-issue list describe the same state.

The protocol ends with a local package. Sending a report, contacting an author, opening an issue, or publishing a revision requires separate authorization and does not occur in this unit.

15 Checklist before closing a round

1. Does every seminar note have a time and location?
2. Is unrecorded content marked rather than completed by guessing?
3. Are summaries, verifications, questions, and objections distinct?
4. Is confidence in transcription kept separate from the truth of a claim?
5. Does every objection state the exact claim and evidence?
6. Do severity and confidence have separate reasons?
7. Do comments assess the manuscript rather than the author’s character or intentions?
8. Are the parts that remain valid acknowledged?
9. Does every response have an action, evidence, location, and status?
10. Can the old version still be found?
11. Do unresolved issues remain open?
12. Does the recommendation follow the evidence gates and repair path?
13. Is the report or response still local and not treated as already sent?

16 Exercises

1. O017-U08-E01 — Four entry types. Classify each of the following sentences as a summary, verification, question, or objection:

   1. “slide 3 states $a_n\to0$”; (b) “is compactness used in selecting the subsequence?”; (c) “direct calculation gives $T(0)=1$, so the claim $T(0)=0$ fails”; and (d) “from monotonicity and boundedness, convergence of the sequence does follow”. For each, state its grounds and verification status. See O017-U08-H01.

2. O017-U08-E02 — A recording gap. A seminar recording is missing 14:20–15:05. Before the gap, the speaker states a lemma; afterwards, the board reads “by the lemma, $A\subseteq B$”. Write three valid note entries: an observation before the gap, a gap record, and a question after the gap. Explain one sentence that must not be written. See O017-U08-H02.

3. O017-U08-E03 — Fixed-point diagnosis. Check P1–P3 in manuscript V1. State exactly which parts are valid, the first unsupported inference, every hypothesis satisfied by $f(x)=x$, and the conclusion that fails. Explain why this example does not refute the existence part. See O017-U08-H03.

4. O017-U08-E04 — Actionable criticism. Turn the comment “this section is confusing and the proof is poor” into one referee comment about P3 with five components: location, claim, evidence, consequence, and request. Add severity and confidence with separate reasons. See O017-U08-H04.

5. O017-U08-E05 — Response matrix. A referee requests

   1. separation of existence and uniqueness, (ii) use of a contraction condition, and (iii) an audit of a corollary whose copy is unavailable. Create three comment–action–evidence–location–status rows. You may decline the second request, but you must give a mathematical reason. See O017-U08-H05.

6. O017-U08-E06 — Evidence-based recommendations. Give recommendations for three situations: (a) the central theorem is valid and only two locators are wrong; (b) one central theorem is false, but complete replacement wording and proof are available, and dependent results can be audited; (c) the central result depends on an appendix that has not been supplied. For each situation, state the category, grounds, and one condition that could change the decision. See O017-U08-H06.

17 Hints and answer guidance

1. O017-U08-H01. (a) Summary; confidence may be high that the slide contains the claim, but the truth of the limit has not been checked. (b) Question; its grounds are a dependency not yet evident.

   3. Objection if the value $T(0)=1$ and the domain have been checked exactly; the claim is refuted. (d) Verification if the monotonicity and boundedness hypotheses are genuinely available and the theorem used is among the prerequisites. The type label does not replace the status column. Return to O017-U08-E01.

2. O017-U08-H02. Record the lemma’s wording with a time and location, create a “14:20–15:05 not recorded” entry, then ask which step connects the lemma to the inclusion on the board. Do not write “the speaker proved $A\subseteq B$ during the gap”; the recording does not support that sentence. Return to O017-U08-E02.

3. O017-U08-H03. P1 is valid because the difference of continuous functions is continuous. P2 is valid because the range supplies the endpoint signs and the Intermediate Value Theorem gives a zero. P3 first fails at $f(y)-f(x)<y-x$. The identity is continuous, strictly increasing, and has range exactly $[0,1]$, but every point is fixed; uniqueness therefore fails. Existence is satisfied, not refuted. Return to E03.

4. O017-U08-H04. An adequate comment points to P3, writes the inequality that does not follow, gives the identity as a counterexample, retains P1–P2, and requests revision of the uniqueness claim and an audit of dependent results. Major severity comes from the claim’s centrality; high confidence comes from the exact example, not the comment’s tone. Return to O017-U08-E04.

5. O017-U08-H05. The first row may be marked complete with the V2 Theorem and Proof. The second may be marked “not adopted, with reasons”: contraction is sufficient but stronger than the strict decrease of $f(x)-x$ used here. The third row must remain withheld because the corollary’s copy is unavailable; do not promise an audit that has not been performed. Return to O017-U08-E05.

6. O017-U08-H06. (a) Generally minor revisions if traceability can still be restored locally; this changes to acceptance once the locators are corrected. (b) Major revisions because a central result changes, although there is a repair path; the recommendation may change once the proof and all dependent results have been audited.

   3. Withhold the recommendation rather than rejecting an unseen appendix; the decision changes when the exact appendix is available and has been checked. Return to O017-U08-E06.

18 Unit completion task

Assemble a complete package of seminar notes, a referee report, an author response, and a revision for the following synthetic materials. The package constitutes the entire available context. Do not invent additional names, DOIs, institutions, pages, recordings, or comments.

At 08:40–09:12, slide 5 displays the choice of $N$. The segment 09:12–09:50 was not recorded. At 09:50, the left-hand board contains one value of $N$ used for $t$ near $x$. At 15:30, a participant proposes $f_n(x)=x^n$. The speaker only replies, “I will check the quantifiers.”

Use the following local facts, and check their application:

• each $f_n(x)=x^n$ is continuous on $[0,1]$;
• the pointwise limit is $f(x)=0$ for $0\leq x<1$ and $f(1)=1$, so $f$ is not continuous at $1$;
• if $f_n\to f$ uniformly on $[0,1]$ and each $f_n$ is continuous, the triangle-inequality proof can choose one $N$ valid for all points and establish continuity of $f$.

Your product must contain:

1. the identity and boundaries of the recording, including one explicit entry for the gap;
2. at least eight timed and located entries using all four types S, V, Q, and O without inventing the gap’s content;
3. separation of confidence in transcription, mathematical verification status, severity, and confidence in diagnosis;
4. a referee report with a neutral summary, scope, at least two valid parts, one major comment, one genuinely local minor comment, open issues, a recommendation, and report boundaries;
5. a quantifier diagnosis showing that pointwise convergence gives only an $N$ that may depend on the point, whereas P01 uses one $N$ over a neighborhood without justification;
6. a complete check of the example $f_n(x)=x^n$ against the hypotheses and the failure of the conclusion;
7. an author response that acknowledges the defect, retains valid parts, and does not claim that the gap’s content is known;
8. a new theorem statement with uniform convergence and a complete $\varepsilon$ proof, including division of the error into three terms;
9. a comment–action–evidence–location–status matrix with at least one genuinely justified open or withheld status;
10. a list of checked impacts and at least two impacts still unknown;
11. a recommendation for V1 and the conditions required before the recommendation can change; and
12. a statement that the entire package remains local and has not been sent or published.

19 Boundaries with B80 and other units

Unit 8 does not teach Python syntax, notebooks, package APIs, visualization, testing, environment management, or numerical-computing techniques; these belong to B80. If a report assesses computational output, Unit 8 uses the reproducible package from Unit 6 and communicates findings rather than teaching how to build the software.

Unit 4 establishes source identities, versions, licenses, and provenance logbooks. Unit 8 merely carries existing identities and locators into notes and reports; it does not redesign the ledger. Unit 5 teaches the architecture of mathematical exposition. Unit 8 does not rewrite a manuscript for general style; it requests and tracks changes justified by comments.

Unit 7 establishes whether a claim is false, its defect type, a reproducer, severity, confidence, and the justified correction. Unit 8 does not repeat that erratum protocol. It takes a diagnosis as input, then teaches how to place it in a report, how an author responds, how an overall recommendation is bounded, and how open issues remain visible. Unit 9, not Unit 8, handles actually authorized submission or community contribution.

20 Sources, provenance, changes, and rights

The original Indonesian prose, seminar-note schema, entry taxonomy, comment contract, report template, recommendation gates, response matrix, synthetic packages, worked cases, exercises, answer guidance, completion task, and rubric in this unit are original O017 material by the O017 contributors, 2026, licensed under CC BY-SA 4.0. This English translation retains that attribution and license.

The Intermediate Value Theorem, existence of a fixed point for a continuous function from $[0,1]$ to itself, the identity-map example, and the theorem that a uniform limit of continuous functions is continuous are classical mathematical results. O017 does not claim to have discovered them. The wording, pedagogical arrangement, proofs presented, review examples, and their audit functions were written specifically for this unit and were not copied from a particular source.

All objects prefixed O017-U08-TALK, O017-U08-MS, and O017-U08-PKT are synthetic and exist only as local materials for this unit. These IDs are not DOIs, publisher identities, or evidence of external works. No external donor is adapted, no external source is quoted, and no real communication has been sent. The CC BY-SA 4.0 license applies to this original O017 material; it does not change the rights in other works that readers may later examine.