From V1 to V6: Turning FOMO Sun into a real decision tool

Today I audited every shipped snapshot of FOMO Sun and wrote down what actually changed between versions, based on code diffs instead of memory.

Why this matters

When you build fast with AI pair programmers, you can ship quickly but lose narrative continuity. A clean changelog makes decisions reversible and helps multiple coding agents collaborate without stepping on each other.

V1 -> V6 in plain English

V1

We got the core loop alive: ask for sunny escapes, rank destinations, show practical options with links.

V2

This was the iteration layer between launch and the first stable archive: UX and structure cleanup to prepare a more coherent interface.

V3 (stabilization)

We heavily rewired the UI and cleaned project scaffolding (README, deploy config, hygiene files). This reduced friction for shipping repeatedly.

V4 (product framing)

This is where the app started feeling like a product:

• Demo/Live mode
• FOMOscore interaction improvements
• Sunset + tomorrow context
• Better destination and timeline presentation

V5 (location + realism prep)

We added geolocation and polished the score popup behavior. We also added a live weather integration module as preparation for replacing mock paths safely.

V6 (decision quality)

This was the biggest decision-quality jump:

• Net sun after travel was introduced
• Optimal travel radius became explicit
• Origin timeline was added to make the contrast obvious
• Per-destination tomorrow sun data improved planning

What still matters for V1 quality

Even with strong UI progress, trust depends on these fundamentals:

1. Strict API contract validation
2. User-visible error/fallback states
3. Minimal analytics + feedback loop
4. Rate limiting and reliability guardrails

New operating rule (Codex + Claude collaboration)

After every deployment, we now update Notion with:

• commit SHA
• deployed environment
• exact behavior/API changes
• test result
• rollback note

That keeps Claude and Codex aligned on one source of truth and prevents regressions from getting lost in chat history.

Closing

FOMO Sun moved from a working demo to a clearer decision assistant in six fast iterations. The next step is less about new UI and more about reliability, schema stability, and measurable quality.