From 7e9714839c12dbe8b9386f933c8b14d49567ae79 Mon Sep 17 00:00:00 2001
From: Arnaud Bailly <arnaud@pankzsoft.com>
Date: Mon, 3 Nov 2025 19:49:48 +0100
Subject: draft plan

---
 lambda-calcul/haskell/README.md | 58 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 58 insertions(+)

diff --git a/lambda-calcul/haskell/README.md b/lambda-calcul/haskell/README.md
index 096627c..d568ace 100644
--- a/lambda-calcul/haskell/README.md
+++ b/lambda-calcul/haskell/README.md
@@ -6,3 +6,61 @@ A minimal functional language to learn and experiment with type systems.
 * http://www.cse.chalmers.se/~bengt/papers/GKminiTT.pdf
 * https://homotopytypetheory.org/2017/09/16/a-hands-on-introduction-to-cubicaltt/
 * https://github.com/kino3/Mini-TT
+
+
+# Agenda
+
+## Notes
+
+* How do we make a language?
+  * How do you make a baby?
+  * Things we use everyday but don't know how it works?
+* Why?
+  * empower and free oneself
+  * Uxntal -> 100 rabbits
+* How do I talk about my world/domain?
+  * would it not make sense to find better ways to express it?
+  * ensure code complexity does not map 1-1 with domain complexity
+* What would it entail?
+  * reuse "simple" existing language focusing on _understanding_
+  * parser -> interpreter -> pretty-printer
+* If I can make my own language, I can bring the language _where_ I
+  need it or want it to be
+
+* starting point: how to talk to a machine?
+
+## Plan A
+
+* (5') Intro:
+  * How to talk to machines?
+  * not reinvent the wheel -> use existing knowledge
+* Collective coding
+  * (5') play with parser/pretty-printer -> relate internal representation w/ external representation (human readable)
+    * representation for humans vs. for machines
+  * (45') focus on the interpreter
+    * introduce the rules of the λ-calculus
+    * implement evaluator
+      * use the leaderboard
+      * use the web UI for checking/troubleshooting
+  * (20') simply typed λ-calculus
+  * (5') type inference
+* (5') Closing thoughts
+  * we'll be exchanging beers for knowledge
+
+## Plan B
+
+* (5') Intro:
+  * How to talk to machines?
+  * not reinvent the wheel -> use existing knowledge
+* Collective coding
+  * (15-20') play with parser/pretty-printer/interpreter -> relate internal representation w/ external representation (human readable)
+    * representation for humans vs. for machines
+    * boolean algebra (together)
+  * (45') focus on the type-checker
+    * motivated by the fact it's easier to write with types
+    * simply typed version
+    * Problem: can we leverage the types to write stuff?
+    * need to modify evaluator to introduce primitive interpretation
+  * (10') type inference
+* (10') Closing thoughts
+  * we'll be exchanging beers for knowledge
-- 
cgit v1.3.1