know.

CookbookKitchen and tableNo. 40

A cookbook manuscriptrecipe development, testing rounds and headnotes for a proposal

Recipe No. 40 · Kitchen and table

For: A food writer testing sixty recipes for a first cookbook

You get: A document per recipe, every testing round, and a proposal

Time: Twenty minutes after each test

See it in know.sh

Which recipes make the book, as a decisions register: what each replaced or beat, and the testing rounds behind every call.

The Tuesday BraiseProposal

6of 7

Which recipes make the book

Decision, key section, left out of the proposal link, 344 words

Where the sixty stand after the spring rounds. Each line points to the testing that settled it; the full rounds are in each recipe’s own document.

Still to decide. Nine recipes have had one round or none; they stay off this list until they have had two.

Round three as a section, under the same three headings as every round, with a note from Claude that it changed two things at once.

The Tuesday BraiseChicken thighs braised with leeks and cider

3of 7

Round 3, 11 Marchless cider, lid off for the last 20 minutes

Evidence, 1 highlight, 1 note, 486 words

What changed. Cider down from 1½ cups to 1 cup (240 ml), after two testers in round 2 called the sauce sweet. Lid off for the last 20 minutes to reduce it.

What happened. In my Dutch oven at 325°F (160°C), 55 minutes covered, 20 uncovered. Thighs at 185°F (85°C) at the thickest point, well past the 165°F (74°C) minimum; the meat pulled from the bone. Sauce coated a spoon. Leeks held their shape for the first time.

Testers.

  • Tester B, gas oven, Denver: “Still a little sweet. Sauce thin.” Used a 7-quart pot.
  • Tester D, electric oven, Atlanta: “Right. Would make again.” Used a 5-quart pot.
  • Tester F, convection, Minneapolis: skin soft, not crisp, after 20 minutes uncovered.

Contains. Milk (butter to finish).

Claude reads two rounds and the recipe’s revisions, then files a comparison as a new section without touching the recipe.

Claude, connected to know.sh

You

Using know.sh, in the shelf The Tuesday Braise, compare rounds 2 and 3 of the chicken thighs with leeks and cider. What changed in the recipe, what did testers say each time, and is anything still unresolved? File it as a section. Don’t touch the recipe or the headnote.

document_list_sections 7 sections in Chicken thighs braised with leeks and cider

section_get Read “Round 2, 2 March” and “Round 3, 11 March”

revisions_list 9 revisions of the overview since 24 February

Claude

Between rounds 2 and 3 the overview changed in two places: cider from 1½ cups to 1 cup, and “lid off for the last 20 minutes” added.

Sweetness: two of three testers in round 2 said sweet; one of three in round 3. Sauce: thin for Tester B in both rounds. B is the only tester using a 7-quart pot, which may explain it better than the cider does.

Still open: crisp skin (Tester F, both rounds) and the pot size, which the recipe does not state.

section_add Added “Rounds 2 and 3 compared” to Chicken thighs braised with leeks and cider, 8

How it works

By the time a braise reaches print you have made it a dozen times, and the question is why it is this way. Each recipe is a document whose overview is the current version; each testing round is a section; each settled call is a Decision. Round two is one click away when round three goes wrong.

  1. Give your assistant the notebook

    Hand over the notebook photos, draft recipes and testers’ emails, and ask for a document per recipe: the latest version as its overview, each test so far as a section.

  2. Log every round the same way

    After each test, ask your assistant to file the round as Evidence, titled by round and date, under three headings: what changed, what happened, and what each tester reported.

  3. Send testers the recipe behind a password

    Press Share on a recipe, set a password and an expiry, and leave out the round logs, so testers read the recipe and nothing else.

  4. Let the rounds decide the book

    Ask your assistant to gather your calls into a decisions element in the proposal: which recipes are in, which were cut, and the rounds behind each. The headnotes you write yourself.

Try this prompt

Your assistant, connected to know.sh (Claude, ChatGPT or a local model):

Using know.sh, read every testing round in the document [recipe] on my shelf The Tuesday Braise and the revisions of its overview. Make a table of rounds: date, what changed, each tester’s verdict and their oven and pot. Then list what is still unresolved. File it as a new section called “Rounds compared”. Do not edit the recipe or the headnote.

Your assistant, connected to know.sh (Claude, ChatGPT or a local model):

Using know.sh, read the Decision sections in every recipe document on my shelf The Tuesday Braise and add a decisions element to a new section in the proposal called “Which recipes make the book”: one decision per recipe, marked decided, open or superseded, with a date, the recipes it beat, and a reason that cites the rounds by number. Decide nothing I have not decided.

Made with

DocumentSectionRevisionsPublic linkYour AI assistantElements

See also