Read a conference with a purpose

1. Keep three records separate

Current requirements: Use the official call for your year and track. Record its URL, check date, format, anonymity, ethics and artifact requirements. Published papers cannot establish today's submission rules.

Award classification: Preserve the official category and source. Best Paper, Distinguished Paper, Honorable Mention and a track certificate mean different things. Leave an unverified category unverified.

Paper edition: Record the URL, version or publication date, and whether you read proceedings, a preprint or an author manuscript. Use page or section locators from that edition. Record material differences instead of silently combining versions.

2. Choose a small, deliberate sample

Start with a few papers close to your research method, then include a different method or an unfavorable case. Where available, include ordinary accepted papers as well as award recipients. Write down why each was selected and what the sample excludes. A small mixed sample can generate useful questions; it cannot establish what causes acceptance or awards.

3. Read for the claim and its evidence

For each paper, name the problem, affected party and contribution in your own words. Locate the strongest supporting result and its limits. Record the measurement unit, denominator and conditions. Distinguish an author-reported result from one you reproduced.

Ask what the comparison isolates. A system change may need a baseline with the same implementation and resources; an interview study may need recruitment details and contrasting participant accounts. Explain an absent comparator when appropriate. A component test, a complete system outcome and a proposed benefit need different evidence.

4. Compare questions, not surface features

Put the records side by side. Look for differences in evidence, assumptions and failure conditions. If you count a pattern, state the denominator and selection rule. Keep counterexamples and missing evidence visible. Figure counts, paper length and benchmark choices are observations about these editions, not universal requirements. “Not reported” does not mean “never tested.”

5. Audit one claim in your draft

Choose a central claim. Mark its support supported, partial, open or not applicable, with an exact draft location and evidence reference. Identify the strongest alternative explanation. Decide whether to add a comparison, expose a limitation or narrow the claim. Finish by checking the current official call separately. A completed worksheet is not proof of truth or submission readiness.

Copyable worksheet

VENUE AND RULES
Venue / year / track:
Official call URL / date checked:
Requirements to verify:

ONE PAPER
Title / authors / year:
Award category / official source / verification status:
Paper URL / edition / date:
Why selected / what this sample excludes:
Problem / affected party:
Contribution:
Evidence / page, section or figure:
Reported by authors or reproduced by me:
Unit / denominator / conditions:
Comparator / what it isolates / justified absence:
Limits / negative cases / missing evidence:
Practical implication / what remains unmeasured:

COMPARE AND APPLY
Pattern or contrast / papers supporting it:
Counterexample / selection limits:
My draft claim / exact location:
Support status / evidence reference:
Strongest alternative explanation:
Next concrete change: