Skip to content

Commit 1dd12ca

Browse files
hyperpolymathclaude
andcommitted
docs: dogfooding opportunities — C→Zig, V adapters, proven modules, VeriSimDB
Catalogues guinea pig fooding candidates for PanLL: - 5 C/libc calls swappable to Zig NIFs (statvfs, kill signals) - 2 V-lang API adapter candidates (HTTP client, settings) - 4 proven library integrations (SafeHTTP, SafeString, SafeSemVer, SafeHash) - 11 VeriSimDB data categories (everything-first approach) Per dogfooding philosophy: test ecosystem interactions first, strip back to proven value later. Every integration is a potential paper example. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
1 parent 15cd652 commit 1dd12ca

1 file changed

Lines changed: 53 additions & 0 deletions

File tree

docs/DOGFOODING-OPPORTUNITIES.md

Lines changed: 53 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,53 @@
1+
# PanLL Dogfooding Opportunities
2+
3+
Guinea pig fooding candidates — places where PanLL can use hyperpolymath
4+
ecosystem projects instead of generic dependencies, to test interactions
5+
and build evidence for those projects.
6+
7+
## C → Zig Swap Candidates
8+
9+
| Current (C/libc) | Location | Zig Replacement | Effort |
10+
|-------------------|----------|-----------------|--------|
11+
| `libc::statvfs` disk stats | `workspace/sysinfo.rs:107` | Zig NIF for system info (like Burble coprocessors) | Small |
12+
| `libc::kill` process signals | `game_preview/commands.rs:209,212` | Zig process management module | Small |
13+
| `libc::kill` SIGSTOP/CONT/TERM | `llm_coding/commands.rs` (3 calls) | Same Zig module | Small |
14+
| Coprocessor C FFI types | `coprocessor/mod.rs:109-119` | Already Zig-shaped (C calling convention = Zig native) | None needed |
15+
16+
## V-lang API Candidates
17+
18+
| Current (Rust) | Location | V Replacement | Effort |
19+
|----------------|----------|---------------|--------|
20+
| HTTP client to services | `http_client.rs` | V adapter layer (BoJ cartridge pattern) | Medium |
21+
| Settings serialisation | various | V API surface for config | Medium |
22+
23+
## Proven Library Integration Candidates
24+
25+
| Current | Location | Proven Module | Status |
26+
|---------|----------|---------------|--------|
27+
| reqwest URL construction | `http_client.rs` | `proven/SafeHTTP`, `proven/SafeUrl` | 0 believe_me |
28+
| File path handling | `workspace/*.rs` | `proven/SafeString` | 0 believe_me |
29+
| Version strings | config files | `proven/SafeSemVer` | 0 believe_me |
30+
| Any crypto/hashing | future VeriSimDB integration | `proven/SafeHash` | 0 believe_me |
31+
32+
## VeriSimDB Integration (everything-first approach)
33+
34+
| Data Category | What to persist | Priority |
35+
|---------------|----------------|----------|
36+
| Panel state | Open/closed, position, size, scroll position per panel | P0 |
37+
| Workspace layouts | Named layout snapshots, auto-save current layout | P0 |
38+
| User settings | Preferences, theme, keybindings, accessibility | P0 |
39+
| Multiuser perspectives | Full workspace snapshot per team member — "see the world from their view" | P1 |
40+
| Session recordings | Command history, panel interactions, timestamps | P1 |
41+
| Undo history | Full undo/redo stack with branching | P1 |
42+
| Panel usage analytics | Which panels used most, interaction patterns | P2 |
43+
| Collaboration state | Shared cursors, live presence, team annotations | P2 |
44+
| Build/test results | CI output, test traces, build metrics | P2 |
45+
| ECHIDNA proof traces | Tactic predictions, premise rankings, proof search paths | P2 |
46+
| Keystroke heatmaps | Input frequency per panel (accessibility + UX research) | P3 |
47+
48+
## Notes
49+
50+
- This is guinea pig fooding — we're testing ecosystem interactions, not optimising
51+
- Some of these will prove not valuable — that's fine, strip them back later
52+
- The portfolio of VeriSimDB use cases IS the evidence for the project
53+
- Every integration is a potential paper example or demo

0 commit comments

Comments
 (0)