Thesis Topics
[25.9.2014]
Bruno Rossi
Department of Computer Systems and Communications,
Lasaris (Lab of Software Architectures and Information Systems)
Masaryk University, Brno
2/11
Computer-Aided Ticket Triage (1/2)
● Ticket triaging refers to the assignment of tickets in issue trackers to the most
appropriate developer(s) for their solution.
● The current thesis will deal with the design and implementation a software
application that would assist with a triage and pre-analysis of newly raised tickets.
1
3/11
Computer-Aided Ticket Triage (2/2)
● This thesis will be done in collaboration with a company - data can be only
accessed internally due to its sensitivity. For this reason, the student will be
requested to sign a Non Disclosure Agreement (NDA).
● Learning data is available in the form of an older SQL-based proprietary tracking
system (MRTS), and already existing tickets in JIRA
● The input for the application would be a newly raised ticket in Atlassian JIRA and
the the output should be some, or all, of the following:
– a suggested name of the assignee, and/or a team;
– a suggested priority/importance of the ticket;
– a suggested component, and/or section of source code;
– a list of topically similar (older) tickets;
● Skills requested: some development skills,
interests for machine learning / text mining
1
4/11
Investigation of Design Patterns Grime in
Software Development (1/3)
● Design Patterns are widely known to be useful to better structure and
design software projects for improved future maintenance.
2
5/11
Investigation of Design Patterns Grime in
Software Development (2/3)
● Design patterns can decay over time and are subject to the so-called
"grime" effect → accumulation of code not specific to the patterns that can
also lead to the violation of the original goals for which the patterns were
introduced
Source: Izurieta, Clemente, and James M. Bieman.
"Testing consequences of grime buildup in object oriented
design patterns." Software Testing, Verification, and Validation,
2008 1st International Conference on. IEEE, 2008.
2
6/11
Investigation of Design Patterns Grime in
Software Development (3/3)
● The current thesis is aimed at understanding the "grime" and related
phenomenon in literature, the practical part will deal with an empirical
investigation of the presence and evolution of design patterns grime in
a set of software projects, understanding
● different types of “grime”
● how frequent such occurrence is
● If there are some causes that are more frequent (that is relation with other
changes within systems during development)
Skills requested: development skills, interests in source code analysis
2
7/11
Applying Text Mining Classification for Software
Requirements Prioritization (1/2)
● Requirements prioritization is an important phase in the software
engineering process that leads to the decision about the most important
requirements to be developed
3
8/11
Applying Text Mining Classification for Software
Requirements Prioritization (2/2)
● The current thesis will:
– review the current state of the art in the classification and prediction of
priorities in requirements engineering
– evaluate several techniques (e.g. different classifiers like Multinomial
Naive Bayes, Nearest Neighbour) for machine learning reasoning to
recommend the priority of requirements to support decision makers
Skills requested: interest in machine learning / text mining
3
9/11
A Controlled Experiment for Pair Programming
Enhancement (1/2)
Learning of
concepts
Recall (RE) Navigator, driver, co-located and distributed PP,
pair rotation, etc..
Comprehension
(CO)
Description
of terms
Explanation of the main terms, navigator, driver,
co-located and distributed PP, pair rotation, etc..
Application
(AP1)
Fill a
compatibility
questionnaire
Create
new pairs
Perform a
“Test run”
Application of basic PP concepts with
supervision
Create
“routine”
programs
Frequent
role switches
Application
(AP2)
Application of basic PP to more complex
problems and without supervision
introduction to distributed PP
Work in
more
complex
programs
Distributed
PP sessions
In a previous Master thesis
("Enhancing Learnability of Pair
Programming Practice when
introducing Novices"), a
framework was created to
improve the learnability of
pair programming by
suggesting several practices,
like affinity scores and
constant revisions to better
form pairs so that novice
developers could improve their
approach to pair programming.
4
10/11
A Controlled Experiment for Pair Programming
Enhancement (2/2)
● The aim of the current thesis is to revise the proposed framework and
apply it by means of a controlled experiment in order to understand:
– the effects on students/developers and their feedback;
– possible improvements to the framework;
Tasks will be the design, preparation, realization and evaluation of a
controlled experiment applied when learning agile development
practices
Skills requested: interest in software development process and in doing empirical
research, communication skills
4
11/11
Additional Information
● If you need more information about one topic you can contact me at
brossi@mail.muni.cz
● You can apply for enrollment in one topic through the IS