Abstract
Learning through practice is crucial to acquire a complex skill. Nevertheless, learning is only effective if students have at their disposal a wide range of exercises that cover all the course syllabus and if their solutions are promptly evaluated and given the appropriate feedback. Currently the teaching-learning process in complex domains, such as computer programming, is characterized by an extensive curricula and a high enrolment of students. This poses a great workload for faculty and teaching assistants responsible for the creation, delivering and assessment of student exercises. In order to address these issues, we created an e-learning framework - called Ensemble - as a conceptual tool to organize and facilitate technical interoperability among systems and services in domains that use complex evaluation. These domains need a diversity of tools, from the environments where exercises are solved, to automatic evaluators providing feedback on the attempts of students, not forgetting the authoring, management and sequencing of exercises. This paper presents and analyzes the use of Ensemble for managing the teaching-learning process in an introductory programming course at ESEIG - a school of the Polytechnic of Porto. An experiment was conducted to validate a set of hypotheses regarding the expected gains: increase in number of solved exercises, increase class attendance, improve final grades. They support the conclusion that the use of this e-learning framework for the practice-based learning has a positive impact on the acquisition of complex skills, such as computer programming. © Springer International Publishing AG, part of Springer Nature 2018.

Abstract
Quizzes are a widely used form of assessment, supported in many e-learning systems. Mooshak is a web system which supports automated assessment in computer science. This paper presents Moozz, a quiz assessment environment for Mooshak 2.0, with its own XML definition for describing quizzes. This definition is used for: interoperability with different e-learning systems, generating HTML-based forms, storing student answers, marking final submissions and generating feedback. Furthermore, Moozz also includes an authoring tool for creating quizzes. The paper describes Moozz, its quiz definition language and architecture, and details its implementation. © Hélder Correia, José Paulo Leal and José Carlos Paiva.

Abstract
Graphs with a large number of nodes and edges are difficult to visualize. Semantic graphs add to the challenge since their nodes and edges have types and this information must be mirrored in the visualization. A common approach to cope with this difficulty is to omit certain nodes and edges, displaying sub-graphs of smaller size. However, other transformations can be used to abstract semantic graphs and this research explores a particular one, both to reduce the graph’s size and to focus on its path patterns. Antigraphs are a novel kind of graph designed to highlight path patterns using this kind of abstraction. They are composed of antinodes connected by antiedges, and these reflect respectively edges and nodes of the semantic graph. The prefix “anti” refers to this inversion of the nature of the main graph constituents. Antigraphs trade the visualization of nodes and edges by the visualization of graph path patterns involving typed edges. Thus, they are targeted to users that require a deep understanding of the semantic graph it represents, in particular of its path patterns, rather than to users wanting to browse the semantic graph’s content. Antigraphs help programmers querying the semantic graph or designers of semantic measures interested in using it as a semantic proxy. Hence, antigraphs are not expected to compete with other forms of semantic graph visualization but rather to be used a complementary tool. This paper provides a precise definition both of antigraphs and of the mapping of semantic graphs into antigraphs. Their visualization is obtained with antigraphs diagrams. A web application to visualize and interact with these diagrams was implemented to validate the proposed approach. Diagrams of well-known semantic graphs are also presented and discussed. © José Paulo Leal.

Abstract
IDEs are environments specialized in support during the development of programs. They contain several utilities to code, run, debug, and deploy programs quickly. However, they do not provide the automatic assessment of programming exercises, which is required in both learning and competitive programming environment. Therefore, IDEs are often underestimated in these contexts and replaced by basic code editors. Yet, IDEs have unique features which are essential for programmers, such as the debugger or the package explorer. This paper presents Raccode, a plugin for assessment of programming exercises in Eclipse. This plugin integrates with Mooshak to combine the diverse capabilities of an IDE, like Eclipse, with the automatic evaluation of exercises, clarification requests, printouts, balloons, and rankings. It can be used both in competitive and learning environments. The paper describes Raccode, its concept, architecture and design. © André Silva, José Paulo Leal, and José Carlos Paiva.

Abstract

Abstract
In ubiquitous streaming data sources, such as sensor networks, clustering nodes by the data they produce gives insights on the phenomenon being monitored. However, centralized algorithms force communication and storage requirements to grow unbounded. This article presents L2GClust, an algorithm to compute local clusterings at each node as an approximation of the global clustering. L2GClust performs local clustering of the sources based on the moving average of each node's data over time: the moving average is approximated using memory-less statistics; clustering is based on the furthest-point algorithm applied to the centroids computed by the node's direct neighbors. Evaluation is performed both on synthetic and real sensor data, using a state-of-the-art sensor network simulator and measuring sensitivity to network size, number of clusters, cluster overlapping, and communication incompleteness. A high level of agreement was found between local and global clusterings, with special emphasis on separability agreement, while an overall robustness to incomplete communications emerged. Communication reduction was also theoretically shown, with communication ratios empirically evaluated for large networks. L2GClust is able to keep a good approximation of the global clustering, using less communication than a centralized alternative, supporting the recommendation to use local algorithms for distributed clustering of streaming data sources.