Activities and dependencies

A schedule is built from activities, each a defined piece of work with a duration, resources and a clear start and finish, and the logical links between them. Getting the logic right is the heart of planning: it is what lets the schedule predict the effect of a delay or change.

The most common dependency is finish-to-start (FS): the successor cannot start until the predecessor finishes, for example plastering after brickwork. Start-to-start (SS) allows the successor to begin a set time after the predecessor starts, useful for overlapping repetitive work like laying a pipeline behind trenching. Finish-to-finish (FF) ties finishes together, for example backfilling cannot finish until pipe laying finishes. Start-to-finish (SF) is rare. Lags add waiting time, such as curing before stripping formwork, and leads (negative lags) allow overlap.

Dependencies can be mandatory (physical, like foundations before columns), discretionary (preferred sequences), external (permits, utility shifting, client approvals) or resource-driven (one crane shared). Good schedules have every activity linked to at least one predecessor and one successor, except the start and finish, avoid hard date constraints that hide true logic, and keep lags visible. The network diagram, in activity-on-node form today, shows the logic; durations and dependencies together allow the critical path method to compute dates and float.