Main page
Staff
Projects
Publications
Useful links
Local
|
COMPSER research group has been merged with Learning + Technology Group. These pages will no longer be updated.
COMPSER group has several projects and activities in which new tools
and methods supporting learning are developed. Some of these are
research oriented whereas many others have mostly practical goals.
Some work is carried out in projects whereas many activities are
organized informally as co-operation of COMPSER members and other
teachers in the laboratory.
Projects on developing tools for CS learning environments.
This project is a joint effort with University of Joensuu. The goal
is to develop new tools and methods for automatic analysis of
free text submissions, and applying these methods to text related
to programming, such program specifications and documents.
This project is a joint effort with University of Turku and
Tamper University of Technology. The project aims at improving
dissemination of tools developed for aiding introductory
programming course, especially automatic assessment tools.
Matrix project is a joint effort of COMPSER and the Software Visualization
Group. The project concentrates on developing the Matrix framework for
algorithm animation and algorithm simulation. This framework is the
basic technology behind several applications: TRAKLA2, MatrixPro and
MVT (Matrix Visual Tester), which are developed in their own projects.
The Matrix framework itself is developed by the Software Visualization
Group, whereas COMPSER group concentrates on educational applications.
Members of the group: see project home page.
Theses:
- Ville Karavirta, XAAL - Extensible Algorithm Animation Language.
Master of Science thesis, 2003.
- Jan Lönnberg, Visual software testing. Master of Science
thesis, 2003.
- Ari Korhonen, Visual algorithm simulation, Doctoral thesis
2003.
- Ari Korhonen: Algorithm Animation and Simulation, Licenciate's Thesis,
2000.
TRAKLA2 is a framework for creating and executing automatically
assessed algorithm simulation exercises. The main features of
exercises include randomization of initial data, instant feedback and
grading, and model solutions presented as algorithm animations.
TRAKLA2 also collects data from the user interface operations and
submitted solutions. The framework provides interfaces for smooth
building of new exercises. The simulation and visualization features
are implemented using the Matrix application framework. Members of
the group developing TRAKLA2: Ari Korhonen (head), Lauri Malmi, Ville
Karavirta, Pekka Mård and Jussi Nikander.
Theses:
- Jussi Nikander, Managing Automatically Assessed Exercises in TRAKLA2.
Master of Science thesis, 2005.
- Petteri Torvinen, Tilastollinen analyysi algoritmisten
harjoitustehtäväsovelmien käytöstä (Statistical analysis of
usage of algorithmic exercise applets). Master of Science thesis, 2004.
- Panu Silvasti, Tilastollisen datan kerääminen
algoritmisten harjoitustehtäväsovelmien
käytöstauml; (Collecting statistical data of the usage of
algorithmic exercise applets); Master of Science thesis, 2003.
MatrixPro is a
teacher's tool for demonstrating algorithms on-the-fly. The system
allows demonstrating the behavior of a large number of algorithms
and data structures in terms of algorithm simulation. There is
also an option to include ones own data structures and algorithms that
can be demonstrated in a similar manner than the ready-made library.
Activities in improving methods of computer science
education
Under this title there are a loose set of activities
related to improving educational methods of computer science.
Most activities are being coordinated by the O-group.
Laboratory of Information Processing Sciences started using
Problem-Based Learning (7 steps method) as a method for teaching
introductory programming in 1999. Since that a new form of this method has
been developed as a combination of problem-based learning and
problem-oriented learning. The method has been very successfully
in production use since 2000, and some small modifications have
been carried out since that. These include the use of
concept maps to clarify students' mental models of
programming concepts. The PBL method applied in TKK for learning
programming was originally developed by Lauri Malmi, Esko Nuutila
and Seppo Törmä. Currently Esko Nuutila supervises
the PBL programming course. The group working with
developing concept maps and a new knowledge representation
method called Text graphs includes Esko Nuutila and
Seppo Törmä.
In 2001-2002 a modified version (Tutorless PBL) of this approach
was applied on a test course. The goal of the new method
was applying PBL method on large scale courses, and
therefore the approach reduced tutor attendance considerably
compared with the original PBL method where the tutor attended
each weekly PBL session. The results, however, turned out to
be slightly negative. Some student groups managed to work well
using PBL on their own, whereas for some groups the challenge
was too heavy. A rigorous analysis of student's interaction
recorded during the PBL sessions is being carried out.
Members of the team include Päivi Kinnunen and Lauri Malmi.
Theses:
- Päivi Kinnunen: Interaction in PBL groups,
Licentiate's thesis in educational sciences.
Past projects
This list contains a number of projects and activities which are not
in active status any more. Some tools are still incrementally
developed and some work is carried out in analysing data.
Visual debugger for Java
This project develops a tool for visual debugging of Java
programs on programming courses. The main emphasis of the tool is to
support students learning programming on their first programming course.
Therefore the tool should be simple to use, and not visualize everything
that happens during Java program execution. The tools concentrates
on the following issues: Visualizing program execution in code,
visualizing run time objects and their references to each others, visualizing
method calls and values of variables. Normal control functions
for debugging are supported, as well. Members of the group:
Otto Seppälä and Lauri Malmi.
Theses:
- Otto Seppälä: Visual debugger for Java,
Master of Science Thesis, 2003.
Scheme-robo -- automatic assessment system for Scheme exercises
Scheme-robo is an automatic assessment system for programming
axercises written in the Scheme programming language. It was developed
for the introductory programming course for CS major students, and the
system has been in production use since 1999. Scheme-robo is available
as free software. Members of the group: Riku Saikkonen. At
the moment development of the system is incremental only.
Evaluation of software systems education
In this project a rigorous survey on the software systems education is
carried out. Several different target groups and data sources
including software specialists in industry, professors and teachers or
computer science in finnish universities, software systems graduates
at TKK, students at HUT, work announcements at USA and curriculum
requirements at USA were surveyd to identify which technical
skills are considered important or less important. A triangulation
of the results summarizes the findings.
Members of the group: Sami Surakka and Lauri Malmi.
Theses:
- Sami Surakka: Needs Assessment of software systems graduates,
doctoral thesis. Helsinki University of Technology, 2005.
ETMEDSA - Evaluation of teaching MEthods for Data Structures and Algorithms
The course of data structures and algorithms has been
using automatic assessment of algorithm simulation assignments as a
primary method for teaching basic data structures and algorithms since
1991 with good results. In 2001 we evaluated the performance of this
instructional method in multiple ways together with the Faculty of
Educational Sciences in the University of Helsinki. Members of the
group: Lauri Malmi (head), Ari Korhonen, Patrik Schein, Pertti
Myllyselkä.
Theses:
- Pertti Myllyselkä: Motivaatiotekijät ja
opintomenestys tietorakenteiden ja algoritmien verkko-opiskelussa
(Motivational factors and success in studies in learning
data structures and algorithms in the Internet). Master of Science
thesis.
Visual query system for study register
This project implements a visual query system for the
study register data bases of Helsinki University of
technology (OREK). Such a system can greatly help
constructing new, possibly complicated queries, since
SQL programming is not needed. Members of the group:
Markku Rontu, Ari Korhonen and Lauri Malmi.
Theses:
- Markku Rontu: System for Enhanced Exploration and Querying.
Master of Science thesis, 2004.
OSR is an acronym of OsaSuoritusRekisteri. This
application is used for storing and combining results about the exams,
project works, assignments, etc. that students have to take to pass
courses in the laboratory. OSR is a web-based Java application which
offers also various statistics how the students have passed the
courses or their parts. OSR has been in production use since year
2000, and is currently used by several laboratories in the CS
department. Development work is currently only incremental. Members
of the group: Pekka Mård and Lauri Malmi.
Palautejärjestelmä (web-based feedback questionnairy system)
This tools has been implemented for easy organization of
feedback questionairies in the web. It includes facilities for
defining various types of questions, managing different
questionnairies and summarizing information about the submitted
answers. Members of the group: Veikko Siivola and Lauri Malmi.
KUHA is an acronym for KUrssiHAllintojärjestelmä.
The system is a framework for carrying out and automating course activities
in web. KUHA is not an integrated learning environment. Rather it has
been designed to be used in building open and distributed learning
enviroments, in which course contents are typically on separate web pages,
and standard email / unix newsgroups are used for commmunication between
students and/or teachers. KUHA supports a large variety of course
management services in the web. These include registration for courses and
course parts, dealing out assignment, registering submissions, assessing
submissions, managing student groups, peer reviewing of student's works,
giving out information via different tools, organizing examinations in the
web, etc.
KUHA has been implemented in a joint project of Helsinki
University of Technology and the Innofactor Ltd. The system has been in
production use since Feb. 2003, and is currently supervised by the KUHA
supervisor group: Lauri Malmi (chairman), Kristel Sarlin, Anneli
Lappalainen, Jussi Nikander, Ursula Holmström from TKK and Sami Ensio
and Tuomas Riski from Innofactor. Incremental development work
is carried out.
Very old projects
LEAD project
(Learning Environment for Algorithms and Data structures), 1997-1999
LEAD project developed various tools for managing
information in the web. These included the prototype of the Matrix
system, (continued in the Matrix project) and the Feedback
questionnairy system (also continued as a separate project) and some
groupware. Members: Lauri Malmi, Ari Korhonen, Veikko Siivola, Markus
Kolsi.
TRAKLA, TRAKLA-EDIT and
WWW-TRAKLA (1990-2000)
TRAKLA assesses automatically algorithm simulation
exercises, which are personally tailored for the students. The
original system was based on email. TRAKLA-EDIT works as a graphical
user interface for solving the exercises. WWW-TRAKLA combines these
with other web material of data structures and algorithms. All of
these tools are still used in 2001, although the development effort
now concentrates in the Matrix system. Members: Lauri Malmi, Ari
Korhonen, Markus Kolsi, Juha Hyvänen (1990-1993) and a number of
students of computer science during 1990-1996.
Theses:
- Ari Korhonen: World Wide Web (WWW) tietorakenteiden ja algoritmien
tietokoneavusteisessa opetuksessa (World Wide Web in computer
aided instruction of data structures and algorithms), Master of
Science thesis,
1997.
KELVIN (1992-1995)
KELVIN is a tool for assessing the style of C programs.
Members of the group: Asko Eerola, Lauri Malmi.
Theses:
- Asko Eerola: KELVIN - A System for Analysing and
Teaching C Programming Style, Master of Science thesis,
1995.
BALT (1992-1995)
BALT was a tool for visualizing various balanced trees
and tree-balancing algorithms and for analyzing their
performance. Members of the group: Lauri Malmi
CALPAS (1991-1993)
CALPAS was a computer-aided course of Pascal programming
which was originally developed using Authorware professional tool and
later implemented with Toolbook for Windows environment. Members of
the group: Jyrki Hautaniemi, Lauri Malmi, Erkki Rämö.
|