While it is well recognized that different learning styles are suitable for different learners, it is also recognized that constructivist learning leads to acquiring more permanent, deeper knowledge that the learner is better able to apply and connect to other related concepts and skills. The challenge of the constructivist learning lies in preparing a gradual sequence of learning scenarios and tasks that are at the proper difficulty level. Learners need to be given a set of elementary mechanisms they can utilize to explore the topic of study through their own experience. They must be able to search in the conceptual space, discover and play with the relations and dependencies between the concepts. They feel motivated as they discover their meaning and potential for reaching a solution to a particular task they are given to solve. Metaphorically speaking, we put them into the world they are discovering, we tell them how to look around, move around in that world, touch and explore what they see, pass through the doors, climb the stairs, unlocking the tools that make the exploration more effective step-by-step. The skills they learn in this process generalize to other tasks, areas and domains and transform their reasoning process to a more autonomous and creative one. We believe a lot of work remains to be done to understand this process well. In the past, the paradigm of most of the constructivist learning was a simple paper and pencil. The advancement of technological revolution brings unlimited opportunities for new paradigms. Our aim is to study and support the constructivist learning process by providing computer programs that serve as the new paradigm for exploration. Each program is focused on a specific topic. Our field of interest is Mathematics and Physics. In Mathematics, a program provides a set of elementary operations, a set of elementary objects on which the operations can be performed to create objects compositions, or configurations, and a "playground" where the user can explore their meaning, functionality and purpose. In addition, the program typically contains a collection of tasks of gradually increasing complexity that can be solved with the provided toolkit, while instead of providing a single solution trajectory, giving the learner the freedom to find its own way to a successful solution. In Physics, the program typically revolves around some set of simulated Physics experiments. The learner can configure the parameters of the experiment and run the simulation - observing how individual variables influence the outcome of the experiment. In this article, we explore our experiences with building and testing programs supporting the constructivist learning.

Keywords: Software, constructivist learning, math, physics.