Studies on cognitive processes underlying programming learning based on detailed records of learner activities and design of teaching learning environment.

本研究旨在发展一种认知理论，揭示编程语言学习背后复杂的认知过程。在理论基础上，设计一个实用有效的编程技能教学环境。为了实现这些目标，我们收集了详细而广泛的教学过程案例，并将其作为真实的编程学习过程进行分析，其研究程度是前所未有的。具体地说，我们收集了大学编程入门课程中成功和失败的学习案例。我们收集了学习者在完成手头任务时感到困难的案例的视频记录；在不同的学习点上实施的精心设计的测验的结果。我们对这些数据进行了分析，并将其分类为一个结构，以反映编程学习的现实情况，以便我们可以使用它们进行进一步的分析，以及进一步的实施。这种级别的数据收集和分析使我们能够理解编程技能获取背后的复杂认知过程。我们的主要成果之一是编程学习的一般整合理论，该理论基于一个框架，该框架将编程的认知过程作为编程中涉及的许多不同表示层的叠加。这与更普遍的认知机制（例如，推理和专业程度）有关。基于这一理论，我们明确了编程新手学习者的复杂和不稳定状态的现实，就他们的认知过程而言，许多分层表示相互作用。然后我们检查了学习环境的条件，就其对新手级编程学习的有效性而言。这让我们想到了一个基于网络的学习环境，它考虑到新手水平的部分理解特征，现实的学习过程。这种环境的强度取决于它对学习障碍的反应能力；当学习者发现自己遇到障碍时，该系统会引导她自己找到必要的信息。我们用实际运行的类检查了这个环境的有效性，并发现了它的有效性。少

This study aims at developing a cognitive theory that reveals the complex cognitive processes underlying programming language learning. It also aims at, based on the theory, designing a practical and effective environment for teaching and learning of programming skills.In order to achieve these goals, we have collected detailed and wide-ranged cases of instructional processes, and analyzed them as the real programming leaning processes to the extent never studied before. Concretely, we have collected the cases of both successful and failed learning occasions at an introductory college level programming classes. We have collected video recordings of the cases where the learners felt difficulties in achieving their tasks at hand; results of carefully designed quizzes implemented at various points of learning. We analyzed and categorized these data into a structure to reflect the reality of programming learning, so that we could use them for further analyses, as well as for further implem … More entation of better teaching practices. This level of collection and the analyses of the data enabled us to understand the complex cognitive processes underlying the acquisition of programming skills.One of our major results is a general integrative theory of programming learning, based on a framework which grasps the cognitive processes of programming as an overlaying of many different layers of representations involved in programming. This is related to more general cognitive mechanisms (c.f., inferences and degrees of expertise). Based on this theory, we made clear the reality of the complex and unstable states of novice learners of programming, in terms of their cognitive processes where many layered representations interact among themselves.Then we examined conditions of the learning environment in terms of its effectiveness for novice level programming learning. This led us to a web-based learning environment, which takes into account the partial comprehension characteristic in novice level, realistic learning processes. The strength of this environment relies on its ability to respond to learning blocks; when a learner finds herself blocked, the system guides her to find a necessary piece of information on her own. We have examined the effectiveness of this environment with actual classes we ran, and found its effectiveness. Less