Tetsuro Kakeshita

Student log analysis functions for web-based programming education support tool pgtracer

We are developing a programming education support tool pgtracer as a plug-in of well-known Web-based Learning Management System Moodle. Pgtracer provides fill-in-the-blank questions composed of a C++ program and a trace table to students. When a student answers a question by filling the blanks, pgtracer collects students answers, elapsed time, evaluation result, etc. as student log. In this paper, we develop seven analysis functions to student log. The student log analysis functions are classified into the analysis functions of a student, those of a question, and those of an answering process. A teacher can analyze achievement level of the students and difficulty level of the problems utilizing the analysis functions. We also perform an evaluation experiment of the analysis function at an actual class and demonstrate the analysis result.

Requirement Analysis of Computing Curriculum Standard J07 and Japan Information Technology Engineers Examination Using ICT Common Body of Knowledge

Information Processing Society of Japan (IPSJ) has announced the Computing Curriculum Standard J07 which is compatible with the Computing Curricula 2005 (CC2005) Series proposed in the United States. Both J07 and CC2005 are composed of five major domains, CS, CE, SE, IS and IT, each of which is developed by a different community so that the relationship among these domains is not clear. In this paper, we analyze each body of knowledge (BOK) of the domains and map them into the ICT common body of knowledge (ICTBOK) which we have proposed in our previous paper. We also analyze Japan Information Technology Engineers Examination (JITEE) whose syllabus is published for each of 12 examination categories provided by the Japanese government. We estimate the degree of importance and the requirement level in terms of the 155 ICTBOK areas for each J07 domain and JITEE examination category by utilizing the mapping. Moreover we estimate the similarity and the difference among them. As a result, the relationship among J07 domains and JITEE examination categories is clarified.

Relationship analysis among curriculum, qualification, BOK and task profile in ICT field

ICT is utilized at many aspects of the society and business. Educational institutions and industries are expected to collaborate to build a consistent ecosystem for education, human resource development and fair evaluation of ICT professionals. In this paper, we analyze relationship among ICT curriculum, ICT qualification, ICT bodies of knowledge and various ICT task profiles. Since they are the major components to develop high level ICT professionals, they should be consistently integrated into the ecosystem. Our relationship analysis will be the first step to promote the mutual understanding of the stakeholders and to develop such a large social system. The analysis is carried out utilizing ICTBOK which we have proposed in our previous paper and i-Competency Dictionary (iCD) which the Japanese government proposed in 2014. We clarified the importance as well as the knowledge and skill level requirements of various activities in terms of ICTBOK domain and area.

The net disk architecture for dynamic load balancing among disk arrays

A disk array is a proposed approach for a high-performance I/O subsystem. It requires dynamic load balancing against varying accesses to prevent some disks from becoming bottlenecks. Thus, we connected the disks by a bus and developed a local migration strategy to migrate data items with a high access frequency within the disk array. However, the bus may become a bottleneck when too many disks are connected to it. To solve this problem, we propose a net disk architecture, which contains multiple disk arrays connected by a crossbar network. Based on the load balancing within each disk array, a global migration strategy is developed to maintain the load balancing among the disk arrays. Therefore, load balancing is established for the entire architecture. Compared to conventional data reallocation techniques that are performed as a background process, dynamic load balancing is effectively maintained even though the disk arrays are burdened with a heavy load by the proposed architecture.

Specification-based component retrieval by means of examples

Software components such as functions, modules and classes should be reused in order to reduce development cost of the application. The situation is essentially the same for databases. Reuse of such components requires a retrieval facility based on its specification. We represent the specification of a component using a set, and propose a retrieval mechanism using a sample containing positive and negative entities. The mechanism enables polynomial time retrieval. Moreover the mechanism provides user assistance for sample construction. We have proposed the notion of element dictionary in our recent paper. A sample which identifies the target set can always be constructed using the element dictionary. We propose two methods to minimize the sample size which corresponds to the sample construction cost. The first one is to retrieve entities from the element dictionary using a user specified query. Another one is based on the weight of each entity. The user selects an appropriate entity from the entity list presented by the system. These two methods can be combined to retrieve the target set using a log n size sample among n sets in most cases.

Fall-in C: a software tool for pitfall detection in C programs

Programming language C has a flexible structure, and its compiler generates compact and efficient object codes. However slight bugs (pitfall) which cannot be detected by the compiler may cause a serious error in C programs. We are developing a software tool Fall-in C to detect such pitfalls. The paper demonstrates the basic features of Fall-in C. In order to enable a programmer to correct the detected pitfalls immediately, Fall-in C is executed within GNU Emacs editor. Pitfalls in C programs are mainly ad hoc. Thus we prepare three pitfall detection methods for the extensibility of Fall-in C: regular expression searching, structural pattern matching and message analysis of external programs. The patterns for the first two methods can be easily added to Fall-in C. Furthermore the message analysis method can be used to integrate several C program checkers such as lint, check and cchk into Fall-in C.<<ETX>>

Transaction sequencing problems for maximal parallelism

The authors define the maximal parallelism scheduling problem and prove that it is NP hard. They examine a special case of the problem and find its nearly optimal solution in the succeeding section. The main techniques for finding an approximate solution are grouping of transactions and defining an order within each group. Moreover they define group of groups (higher level group) and improve approximation using it.<<ETX>>

A web-based assessment tool for various types of self-evaluation utilizing common BOK in ICT

Information and communication technology (ICT) is essential as infrastructure in the modern society. High level ICT professionals are strongly required. Universities, ICT industries and ICT qualification bodies are expected to collaborate to build a consistent system for education, training and evaluation to develop such high level ICT professionals. In this paper, we propose a web-based assessment tool named “cresie”. The tool maintains various types of standard requirement data such as those of college level computing curriculum, various ICT qualifications, and ICT task profiles. A student or an IT engineer can upload their self-evaluation data of achievement to cresie and compare the data with the requirement data. He (or she) can also compare the data with other cresie users. All of the evaluation and requirement data are described in terms of the common ICT body of knowledge (ICTBOK) which we have proposed in our previous paper. The system can also be used for curriculum assessment and requirement assessment for recruiting and evaluating IT professionals.

Utilizing Programming Education Support Tool Pgtracer in an Actual Programming Course

We developed a programming education support tool pgtracer. Pgtracer provides fill-in-the-blank questions to the students and collects student log to analyze students learning process and understanding level. In this paper, we report our experience to utilize pgtracer at an actual programming course as homework assignment. We develop fill-in-the-blank questions corresponding to the course syllabus at each week. Student activities on pgtracer are analyzed to develop questions for the succeeding weeks. We also provide data to the instructor about the activities and achievement of the students for better collaboration between lecture and homework. We received positive feedbacks from both of the teacher interview and student survey about the usefulness of pgtracer as programming education support tool.

Understanding Level Analysis of Students Using Programming Education Support Tool Pgtracer

We are developing a programming education support tool pgtracer. Pgtracer provides fill-in-the-blank questions to the students and collects student log to analyze students learning process and understanding level. In this paper, we utilize pgtracer at an actual programming course and analyze the collected log. There is a negative correlation between required time and right answer ratio of a blank. Although there is a correlation between the mid-term examination score and pgtracer score, pgtracer score is useful to distinguish programming ability of the students. We also perform survey and interview to the students about the usefulness of pgtracer and fill-in-the-blank questions.