Skip to content
Create an account for full access.

Moving away from "waterfalls" with gates

The stages of the life cycle in the waterfall life cycle ended with pre-planned gates, where the results of the previous stage were evaluated by independent experts and a decision was made to continue the project to the next stage (go/no-go). If before the pre-planned review by independent experts within the gate, one of the developers of system parts did not finish the development of their part and the verification of how it does not disrupt the operation of the entire system, the whole project waited for the completion of that developer's work. Gates were specifically designed to identify systemic risks - that is, risks of configuration collisions, unexpected systemic effects/emergence, and unforeseen difficulties in developing individual parts of the system. And if the results of someone's individual work are not included in the gate reviews, there is a risk of not considering some systemic effects, some significant conflicts between system levels in the overall work.

In the early versions of the life cycle of large (primarily aerospace, traditional for systems engineering) projects, there were about fifteen gates, and the project was stopped at all fifteen points for the consolidation of all individual developments into a consistent and free from configuration collisions whole, followed by a review of the stage results by independent experts. As the understanding of the fact that the waterfall model with gates is utopia, the emergence of configuration management computer systems and changes and the reduction of the number of configuration collisions, the transition to parallel engineering, the increase in the reliability of engineering solutions through simulation modeling, the number of gates decreased.

In aviation, there were initially about seven gates, and now there are only three, and not all of them are related to engineering: for example, a decision to terminate a project is made when the design is quite advanced, but there are not enough pre-orders for the new aircraft model to ensure the financial success of the project[1].

The absence of pre-planned gates does not mean that work management is not being carried out. It is conducted, but proceeds through milestones, representing the expectations of achieving a certain state of an alpha in the project by a certain point in time. If a milestone is not passed by its planned time (the state of an alpha does not meet expectations), all possible measures are simply taken for its prompt completion, but this does not stop all other work towards achieving other milestones, as it would be the case with gates.

  1. Altfeld, Hans-Henrich. Commercial aircraft projects: managing the development of highly complex products, 2010 ↩︎