The Death of Linear Logging

For decades, the standard way engineering teams have documented decisions is by creating an append-only log of events. We write meeting notes, slack messages, Jira comments, and PR descriptions in a chronological sequence. It feels productive at the moment.

But six months later, when a new engineer joins the team and asks "Why did we choose Postgres over Mongo for the user service?", they are forced to spelunk through a massive, disconnected history of chronological logs.

The Problem with Time-Based Documentation#

Append-only logging (like a Slack thread or a Google Doc of meeting notes) optimizes for writing, not reading. It assumes that the context of the present moment will somehow naturally translate to the future. It never does.

When documentation is tied to a specific date and time, it becomes stale immediately.

The Hub-and-Spoke Solution#

Instead of linear logs, teams need a Hub-and-Spoke model.

• Hubs: Evergreen architecture files (e.g., architecture/database.md) that describe the current state of the system.
• Spokes: PRs, Jira tickets, and architectural decision records (ADRs) that link back to the hub.

When a decision is made, you don't just write a meeting note. You update the Hub file. This ensures that the documentation remains a living, accurate representation of your codebase, rather than a graveyard of past decisions.

Dicty was built specifically to solve this problem by automatically reconciling your spoken intent directly into your Hub files.