# Notes

Scratchpad for teaching preferences and working notes. Read this before designing any lesson.

## How to teach me
- Explain in **Hinglish** when a concept is confusing; keep all model English sentences in clean English.
- Lessons must be **short** — target 15 minutes. I open them between client work and stand-ups.
- Every lesson must contain all five strands, because I asked for them explicitly:
  1. **Speaking** — at least 3 microphone drills (say it, get scored, fix it).
  2. **Pronunciation** — native-like rhythm: sentence stress, weak forms, linking. Aim to sound natural, not only correct.
  3. **Vocabulary** — my work words **plus general high-frequency English** (Oxford 3000 bands), with stress marks and audio.
  4. **Grammar** — exactly one point, **named properly** (*present perfect*, *past participle*, *gerund*) and tied to a sentence I actually say.
  5. **Listening** — audio I hear *before* seeing the transcript, then comprehension questions.
- Give me the words as **ready chunks**, not as grammar to assemble.
- Show me **why** a phrase is the one a developer or client expects; I get bored by rules with no payoff.
- Interactive HTML is welcome — I am a web developer, quizzes and drills do not scare me.
- Reference sheets should **print well**; I may stick one next to my desk.

## My working context (use this to choose examples)
- Full-stack developer at **Edone Solutions**.
- Owner has asked me to **manage 3–5 developers** who **do not speak Hindi**, so my instruction-giving has to be English.
- Owner has asked me to handle **customer meetings**; the named client project is **Zaira Furnishing** (curtains, wallpaper, mattresses, upholstery).
- My written English is strong; my speaking confidence is the gap, and my daily speech is Hindi.
- Real practice available daily: stand-ups with my team, calls with the owner, client meetings, WhatsApp voice notes to friends.

## Curriculum plan (keep MISSION.md short; the roadmap lives here)
Five strands run in parallel: **speaking / pronunciation / vocabulary / grammar / listening**.
- **001** ✅ The First 60 Seconds — opener, purpose, read-back, the V/W sound.
- **002** ✅ The Stand-Up — status triangle, task assignment, blockers, present perfect (named), manager word stress, listening.
- **003** ✅ Native Rhythm — sentence stress, weak forms + schwa, linking, flap T, fast-speech listening, general vocabulary.
- **004** ✅ Customer meeting, part 1 — the six beats (open, ask, read back, scope, push back, close), polite pushback on a change request, aspiration (the puff of air), the article named, Zaira Furnishing requirements call.
- **005** Explaining technical things to a non-technical owner/client, plus a 60-second delay report with a new date.
- **006** Developer feedback and code-review English — praise, correction, disagreement, without offending.
- **007** Deadlines and "no" — negotiating dates, saying no politely, chasing another team.
- **008** Small talk with friends and teammates, using the general vocabulary bank (the original everyday goal).

### Track: pronunciation (native-like)
**Chosen intensity: level (b)** — "close to native rhythm in *rehearsed business sentences*". 10 minutes daily + one short video a week, measured by a weekly 60-second benchmark recording. Not free-speech accent work, and no promise of being mistaken for a native speaker.
Daily 10 minutes (see `reference/pronunciation-routine-and-benchmark.html`): 3 min warm-up weak forms & linking → 3 min marking stress on tomorrow's real sentences → 4 min shadowing a native audio. Weekly: one Tim's Pronunciation Workshop video for the feature of the week + one benchmark recording into `./recordings/`.
One feature at a time, each drilled with the mic and the stress widget:
1. **Sentence stress** — only content words get hit (003)
2. **Weak forms + schwa /ə/** in *for, to, and, of, can, was* (003)
3. **Linking** — consonant to vowel, and the /j/ and /w/ glides (003)
4. **Flap T** in American English — *better, water, little* (003, hearing only)
5. **Aspiration** of p / t / k at the start of a stressed syllable — *page, ticket, code* ✅ (004) — drilled with the **paper test** (10 cm, the sheet must move), never after /s/ (*scope, stop, staging*) and never on an unstressed syllable (*production, today*). Evidence: [[RESOURCES.md]] Narkar 2021.
6. **/θ/ and /ð/** — *think, three, this, the* — against the Indian /t̪/ habit (lesson 005)
7. **Vowel length** — *ship/sheep*, *full/fool* (lesson 006)
8. **Intonation** — rising questions, falling certainty, and where a client expects a pause (lesson 007)

### Track: general vocabulary
Oxford 3000 in frequency bands, about 12 items per lesson, always inside a sentence, never a bare list. Work words stay in every lesson too.

### Track: grammar terminology
Every grammar point is given its proper name and promoted into `GLOSSARY.md` once usable.

Grammar queue: present perfect vs past simple ✅ (002) → **content words vs function words** ✅ (003) → articles a/the ✅ (004, as **indefinite article** vs **definite article**) → prepositions of time and deadline (by / until / before) → modals for polite requests (could / would) → first conditional for scope → gerund vs infinitive → relative clauses for describing (*the page that has the price list*).
Listening routine: one BBC 6 Minute English episode (twice) + British Council A2/B1 listening + the listening section inside each lesson. From 003 onward, add one **fast speech** pass per week at 1.25× speed.

## Environment
- **Publishing (asked for in session 4): the tutor does `git add` + `git commit` + `git push origin main` after every session — no need to ask.** Vercel re-deploys automatically from `main`, so a lesson is not "done" until it is pushed and visible at `https://spoken-english-grammar-vocabulary.vercel.app/lessons/0004-customer-meeting.html`. Never report a lesson as published from a local commit alone: verify with `git ls-remote origin refs/heads/main` against `git rev-parse HEAD`, and check the file listing through the GitHub contents API.
- Workspace root: `C:\Altamash-Next-js-Project\spoken-english-learning` (renamed from `zaira-furnishing` at my request; the client's catalog files live here too and are a good source of domain vocabulary).
- Shared components live in `./assets/` — `lesson.css` (Tufte-style typography) and `speak.js` (text-to-speech, script player for listening practice, mic drills, quizzes, word-order builder, sentence-stress widget, timers). A play button can override the script's speed with `data-play-rate="1.25"` — that is how lesson 004 offers a fast replay next to the normal one.
- `GLOSSARY.md` holds the canonical grammar and pronunciation terms. Lessons must use exactly those terms, and a term is added only once I can use it correctly.
- **Microphone drills need a local server**, because browsers block mic access on `file://` pages. From PowerShell:
  ```
  cd C:\Altamash-Next-js-Project\spoken-english-learning
  python -m http.server 8080
  ```
  then open `http://localhost:8080/lessons/0002-the-stand-up.html` in Chrome or Edge. (`npx --yes serve -l 8080 .`
  also works if you prefer Node.) Stop the server with <kbd>Ctrl</kbd>+<kbd>C</kbd>.
- Every lesson body has a normal path too: if the mic is blocked, the drill falls back to "say it, then compare with the model answer" — never a dead end.