ZASSELECTION - dzuddiyn/ZASS-Zero-to-Architecture-Structured-Sprint GitHub Wiki

ZASSELECTION

Current version: v0.2.0
Status: UX FLOW LOCKED

AI compares. You decide, AI saves.

ZASSELECTION is a selection method for choices where the main problem is comparing alternatives and keeping the resulting decision traceable.

The short onboarding guide lives in ZASSELECTION/README.md. This Wiki page holds the longer operating reference.

Core user surface

[ REVIEW ]      [ SAVE ]      [ HISTORY ]

REVIEW

With new input:

new option / evidence / question
        ↓
REVIEW
        ↓
compare with existing state
        ↓
update Selection Matrix
        ↓
AI Recommendation
        ↓
👉 Your selection?

Without new input, REVIEW means re-view:

  • load the current saved state;
  • show the current Selection Matrix;
  • show the current AI Recommendation;
  • do not change scores without new evidence.

SAVE

If a user selection already exists, SAVE commits the current selection state.

If there is no final selection yet, SAVE preserves the current matrix and information as:

UNFINISHED / DRAFT

When no persistent write integration is available, SAVE should produce an updated ZASSELECTION.md for download/export.

When a real GitHub or future AI-SYNC writer is available, do not claim success until the write returns a factual receipt.

HISTORY

HISTORY opens a compact list:

MY SELECTIONS HISTORY

Work laptop           → Laptop B
AI subscription       → ChatGPT Plus
Citation software     → Mendeley
Rental house          → Taman Putri unit A
New PC                → Draft

A saved item can be reopened so the old matrix, options and decision state become the starting point for new comparison.

PICKS

The quick review technique used by ZASSELECTION is:

P → I → C → K → S
  • 🎯 P — Pin the Problem: define what is actually being selected.
  • 🚧 I — Identify Must-Haves: identify mandatory requirements.
  • 📊 C — Compare Options: build and update the Selection Matrix.
  • ⭐ K — Keep the Best Candidate: show the current AI recommendation.
  • 💾 S — Select & Save: the user chooses; the system records.

PICKS is optimized for fast product review, while the underlying ZASSELECTION method can also support non-product choices.

Visual presentation lock

The quick-review presentation uses these icons consistently:

🎯 P — Pin the Problem
🚧 I — Identify Must-Haves
📊 C — Compare Options
⭐ K — Keep the Best Candidate
💾 S — Select & Save

🤖 AI Recommendation
👉 Your selection?

This is a presentation convention only; it does not change scoring, authority, or persistence semantics.

Selection Matrix

The Selection Matrix is the main comparison surface.

Example:

Criteria Weight Option A Option B Option C
Performance 30% 7 9 8
Upgrade 20% 8 9 7
Price 30% 7 8 8
Practical Fit 20% 8 9 8
TOTAL 100% 7.4 8.7 7.8

Per-option explanations should normally stay short, around ten words, unless the user asks for deeper analysis.

A new option updates the existing matrix rather than restarting the entire selection process.

AI recommendation vs user selection

These are separate:

AI Recommendation ≠ Your Selection

AI may recommend the strongest current option.

Only the user chooses the final option.

An explicit user choice is treated as the current selection and may be committed immediately when persistence is available.

Working State

When the user has not selected yet, ZASSELECTION keeps a temporary Working State:

Current decision
Current options
Current criteria / weights
Current Selection Matrix
Current AI recommendation
Your Selection: Not selected

The user may continue asking questions or add options without being forced to decide.

Working State must not rely solely on a platform AI's long-term memory.

Saved selection format

A saved Markdown record should remain compact but sufficient to continue later.

Example:

# Work Laptop

**Decision:** Choose a work laptop
**Status:** COMMITTED
**Selected:** Laptop B

## Must-Haves
- RAM >= 32GB
- SSD >= 1TB

## Selection Matrix

| Criteria | Weight | Laptop A | Laptop B | Laptop C |
|---|---:|---:|---:|---:|
| Performance | 30% | 7 | 9 | 8 |
| Upgrade | 20% | 8 | 9 | 7 |
| Price | 30% | 7 | 8 | 8 |
| Practical Fit | 20% | 8 | 9 | 8 |
| **TOTAL** | **100%** | **7.4** | **8.7** | **7.8** |

## Quick Notes
- Laptop A — Good battery, weaker overall performance.
- Laptop B — Best balance of performance, upgradeability and price.
- Laptop C — Balanced, but less upgrade flexibility.

## AI Recommendation
Laptop B

## Your Selection
Laptop B

If the user has not selected:

**Status:** UNFINISHED / DRAFT
**Selected:** Not selected

The exact storage representation may evolve, but enough state must remain to reopen the selection without starting from zero.

Current persistence modes

Manual Markdown export

AI chat
   ↓
REVIEW
   ↓
SAVE
   ↓
updated ZASSELECTION.md
   ↓
download / reuse in another AI

This is the universal fallback.

GitHub mode

AI chat
   ↓
REVIEW
   ↓
SAVE / explicit selection
   ↓
GitHub write
   ↓
real commit receipt

This enables persistent, versioned selection state across AI tools that can read and write the same repository.

AI-SYNC

AI-SYNC is a separate transport-layer project and is not part of the ZASSELECTION method.

ZASS / ZASSIMPLE / ZASSELECTION / other projects
                     ↓
                   AI-SYNC
                     ↓
        GitHub / Google Sheets / AI-SYNC DB

AI-SYNC may grow into multiple transport implementations or versions while ZASSELECTION remains stable at the method layer.

Source-of-Truth boundary

ZASSELECTION itself defines how selections are compared, reviewed and recorded.

Current and future storage may include:

  • Markdown files;
  • GitHub repositories;
  • Google Sheets;
  • AI-SYNC databases.

The storage backend is not the method.

Portable methodology ≠ portable automation.

The normal user path should remain simple even if transport and persistence become sophisticated.

Relationship to the ZASS family

Method Purpose
ZASS raw idea → evidence → decision → architecture
ZASSIMPLE lightweight conversational path toward architecture
ZASSELECTION alternatives → comparison → recommendation → user selection

ZASS / ZASSIMPLE / ZASSELECTION structure reasoning and selection records.

AI-SYNC moves and writes information.

Persistent destinations hold the Source-of-Truth records appropriate to each workflow.

Authoritative files