Initial Brainstorming Session

Session Notes: From RPG/400 to Zig

Date: November 29, 2025 User: Wesley

1. The Search for the “Blue Book”

Goal: Identify a specific RPG programming textbook from the user’s early career at a textile mill.

• Identified Book: Programming in RPG/400 (2nd Edition, approx. 1995)
• Author: Judy Yaeger
• Publisher: Duke Press / 29th Street Press
• Context: This was the definitive guide for the “Original Program Model” (OPM) before the industry shifted to ILE RPG.
• Action: User confirmed this was the correct book (teal/black cover) and discussed finding a used copy for nostalgia/reference.

2. Project A: The RPG Interpreter (The “Homunculus”)

Concept: A hobby project to write an interpreter for the RPG language using Zig.

• The Architecture:
  • Backend: Use MariaDB/Postgres to simulate the AS/400’s “Single Level Store” and ISAM filesystem.
  • The Cycle: Implement the “Logic Cycle” (implicit read/process/write loop) by mapping F-Specs to SQL cursors.
  • Parsing: Use Zig’s fixed-format string slicing (e.g., line[6..15]) to handle the column-strict syntax of RPG/400.
• Key Insight: RPG was effectively “Proto-Spark”—a data warehouse in a box where schemas (F-specs) and transformations (The Cycle) were tightly coupled to the OS.

3. The Tool: Zig

Why Zig?

• Vs. Rust: Avoids the “Borrow Checker” complexity that frustrated the user previously.
• Vs. Java: No hidden allocations or Garbage Collection pauses. Memory is treated as a flat buffer (perfect for fixed-width text parsing).
• Key Features for this Project:
  • Slices: line[10..20] references original memory (zero-copy) rather than creating new String objects like Java.
  • Arenas: Use ArenaAllocator to dump all memory at the end of a cycle, simulating a “batch job” cleanup.
  • Comptime: Validate column specs at compile time.

Corporate “Superpower”:

• Zero-Install: Zig compiles to a single static binary. It does not require Admin Rights, Visual Studio Build Tools, or complex environment variables.
• Cross-Compilation: Can compile a Linux binary from a locked-down Windows laptop (zig build-exe -target x86_64-linux).

4. Project B: The Pivot to “Momentum”

Observation: The RPG interpreter is a massive undertaking. To build momentum, we pivoted to smaller, finishable “weekend projects.”

The Shortlist:

1. Conway’s Game of Life (Selected)
   • Core Concepts: Double-buffering memory (Current Frame vs. Next Frame), 2D Array manipulation, ANSI terminal rendering.
   • Why: Immediate visual feedback; teaches stack vs. heap allocation.
2. Ad-Hoc CSV ETL Tool (Work Utility)
   • Core Concepts: Stream processing (stdin to stdout), buffered reading, cross-compiling for Linux servers.
   • Why: Solves the “Python is too slow/hard to deploy” problem at work.
3. Factorio Blueprint Decoder
   • Core Concepts: Base64 decoding, Zlib decompression, JSON parsing.

5. Next Steps

• Immediate: Implement “Game of Life” in Zig to learn the syntax and memory model.
• Work: Attempt to rewrite a slow Python utility in Zig to test the “no-admin” deployment workflow.