SOAR is a great project and shows how to combine different academic disciplines into reproducible research. It was created with openness in mind and psychological knowledge can be transferred easily into the domain of robotics. The amount of existing model files in the “.soar” format is large and this makes it easy for beginners to write it's own applications.
From STRIPS to SOAR
The STRIPS planning system is known for its contribution to the Shakey the robot project. It was utilized for symbolic planning of the actions. The solver accepts a goal state within the Shakey domain and determines the next movements. Even the STRIPS language evolved into the modern PDDL syntax, it can be called a milestone for AI history, because it explains how to plan if a model is known.
In contrast to the STRIPS systems, the architectures SOAR and ACT* were developed within cognitive psychology. They are simulating human thinking and the researchers are trying to understand the memory of humans. All the actions within SOAR are slowly downed so that to simulate the subject realistically.
Unfortunately, it's not clear what the relationship between STRIPS and SOAR is. Understanding and describing STRIPS isn't that hard. It's a planning system which brings the system into a goal state. On the first look, SOAR has a similar functionality because the SOAR input file contains a STRIPS-like domain model. But apart from planning, SOAR has more functionality.
According to self-definition, STRIPS is an example for narrow AI, which is solving a subproblem (planning) within a robot control system. While SOAR should be an Artificial General Intelligence, which is emulating a human. How can STRIPS be extended into SOAR?