DIGITAL LIBRARY
LEARNING BY RESULT-DRIVEN PROBLEM-SOLVING. A CONCRETE APPLICATION TO DATABASE QUERYING
1 Dipartimento di Matematica - Università di Pisa (ITALY)
2 Accademia Navale di Livorno (ITALY)
About this paper:
Appears in: ICERI2009 Proceedings
Publication year: 2009
Pages: 6998-7006
ISBN: 978-84-613-2953-3
ISSN: 2340-1095
Conference name: 2nd International Conference of Education, Research and Innovation
Dates: 16-18 November, 2009
Location: Madrid, Spain
Abstract:
With the general increase of available data, database technology is becoming so pervasive that nowadays it is part of the basic ICT literacy taught at school. On the other hand, huge collections of data demand sophisticated enquiry and skill to extract meaningful information.
Database querying is a typical problem solving activity: it requires both knowledge of the underlying database organization and proficiency with the basic querying constructs such as criteria, grouping and composition. In particular, a complex interrogation may require several intermediate steps, so the ability to break down the task into a hierarchy of linked queries is an important database problem-solving skill that is worth teaching.
Effective teaching should provide both necessary notions and expertise to apply them to solve real problems. A good way to achieve this is to analyse solutions of good examples. In a database setting it means the study of good QBE or SQL coding of significant interrogations.
Personal engagement catalyses student interest and promotes fast learning. Thus, it is crucial to guide the student beyond the simple analysis of teacher’s solutions. We propose a software tool to approach this aim; it relies on a clear distinction between the solution and the results it produces. In database terms, this translates into the distinction between (a suitable hierarchy of) queries and the resulting recordsets.
In our teaching environment, given an informal set of data retrieval tasks, the teacher prepares his own collection of solving queries. Our authoring system produces a student database by substituting the teacher’s solutions with tables containing the corresponding recordsets. In the database proposed to the student, the description of each (sub)problem is available together with the expected recordset; these results supply a complementary and unambiguous description of what each (sub)query is supposed to obtain. The student can see all expected recordsets (even for intermediate solution steps) but, recall, he can inspect teacher’s results only, not solutions.
This approach, based on the breakdown of the original set of database query together with the accessibility of all expected recordsets, offers some benefits that are generally not available in standard problem-solving approaches:
• Teacher’s results provide stimulating feedbacks about the correctness of the student’s solution without forcing him to mimic the teacher’s solution; in fact, the student is asked to obtain the same recordsets with his own solution.
• The student may use teacher’s results from a previous steps as input recordsets to validate the query he has given as solution to a subsequent step. By doing so, all steps become independent of each other and can be singly faced. Consequently, the student is not blocked if he is unable to solve any intermediate step.
• Having all steps independent, the student is also free to tackle the whole task as he prefers. He can approach the collection of sub-problems with different problem-solving methodologies (such as top-down or bottom-up).
The authors have used this teaching approach for some years at the Naval Academy in Livorno. In the paper, we describe the authoring system and report the results of the pilot study.

Keywords:
technology, educational software, learning and teaching methodologies.