L39 FINAL PROJECT - PROJECT RECORD
Brief (A / B / C / D): ___      Started: ____________
AI use: AI allowed — keep a record. Sections 1, 2 and 3 are your own
work, finished before you ask an assistant for any code.

Keep this file next to your code. Fill it in as you go, not at the end.
Lesson 40's write-up is built from it, and the method score depends on
this record showing the order you worked in.

======================================================================
1. SPECIFICATION
======================================================================
Purpose (one sentence):

Who relies on it, and the error they fear most:

Inputs (name, form, example):

Output (exact form, including bad / empty / no-answer cases):

Constraints:

Bounds and stopping conditions (Brief C; others N/A):
  Tools and what each can reach:
  Step limit:
  Other stopping conditions:
  When unsure, the agent:

Assumptions (A1, A2 ...), and who would confirm each:

Acceptance criteria (AC1, AC2 ...), each traced to the brief or to an
assumption:

What I cut to make it fit, and why:

======================================================================
2. SELF-APPROVAL
======================================================================
Checklist completed (L39_spec_checklist.txt):  date/time ____________
Lines answered NO, and why I accepted them:
Stranger-test questions the assistant asked, and what I changed:

======================================================================
3. EVALUATION SET  (written before reading any generated code)
======================================================================
Time I finished the evaluation set: ________
Time I first read generated code:   ________

  #  | kind (normal / boundary / empty / bad / expect-to-fail)
     | input | expected result | how I worked out the expected result
     | criterion it checks
  E1 |
  E2 |
  ...
At least ten items. At least three I expect a first attempt to fail.

======================================================================
4. IMPLEMENTATION
======================================================================
File name(s):
Lines I changed by hand are marked in the code with
  # CHANGED: <one-line reason>

======================================================================
5. VERIFICATION RECORD
======================================================================
Run 1 (first generated version) - paste the full evaluation output:

For each failure, one entry. One change per hypothesis (Lesson 15).
  Item:
  Expected / got:
  Hypothesis (written before changing anything):
  Evidence:
  Category: SPECIFICATION / GENERATION / MY TEST WAS WRONG
           (Brief B: also name the pipeline stage, Lesson 25)
           (Brief C: also the first incorrect step, Lesson 31)
  Change:
  Re-run result (all items):

Checks by hand against the real data or document (which rows,
sections or steps, and what you found):

Final run - paste the full evaluation output:

What I did NOT check, and why:

======================================================================
6. AI-USE RECORD
======================================================================
AI-USE RECORD
Label:            AI allowed — keep a record
Assistant:        (name and version, or "local model")
Date(s):
Prompts:          (paste them, or a link to the conversation)
What I kept:
What I changed:   (and why)
What I verified:  (which claims or results I checked, and against what)
