Scan the System

A scan is the agent’s comparison between the Model and its Requirements and the existing code. What it finds becomes Tasks, for what is missing, and Drifts, for what differs.

Press Scan System

Scan System is at the bottom of the side menu, after a line, and is the only button that scans the whole System. A dialog asks Scan the whole System?; confirm with Scan.

Enlarge screenshot:

1. 1
2. 2

1. Scan System, at the bottom
2. Scan starts it

When Tasks were finished since the last scan of the whole System, the button says Scan System · N, with N the number of Tasks.

Follow the scan

The scan starts without taking the focus. A ghost Scanning System shows in the status bar, and a Scan conversation opens in omp Chat.

Enlarge screenshot:

1. 1
2. 2

1. A ghost while it scans
2. Scan conversation

Each Machine in scope is compared directly with the code, one sub-agent for each Machine when the agent has sub-agents, one after the other when it does not. A Machine with no differences produces nothing. The report says how many Machines were compared and how many produced Tasks or Drifts.

Read the result

What the scan found shows on the Board: a red card in the Backlog for each Drift, and new Tasks in Active. The number on Board counts the decisions that wait for you.

Enlarge screenshot:

1. 1
2. 2

1. The Drift it found
2. Cards to decide

Decide on the Drifts: Resolve Drifts. Run the new Tasks: Run Tasks.

Scans that start by themselves

• During a Run, with the Advisor Scanner on (the default), an advisor checks the code as the agent writes it, so the Run does not scan before or after. With the setting off, the Run scans the Tasks it is about to build and, at the end, what it built. See Run Tasks.
• After a Merge in Solo mode, Trama scans the whole System by itself. See Solo mode and Merge.
• A closed GitHub issue in the Team view can be scanned alone: Scan compares only the part of the Model that issue referenced. See Manager mode and Team.

A scan takes the model set as Scan in Settings. Trama remembers which Model elements it already compared against the same code and the same text, so a Run does not compare them again.