Supporting exploratory analytics on repository-extracted schema histories by integrating external contextual information (Master thesis)

Παππάς, Αθανάσιος

Data-intensive software systems evolve over time and, as part of this evolution process, so does the schema of any database which is included as an integral part of them. Version control systems store the version histories of open source software projects and the information extraction from these histories can be useful for gaining insights about their evolution. Alongside with the software evolution, new information is posted in different external systems improving in this way the software development experience for example. In this thesis, we combine all the various, heterogeneous, dissimilar sources of information for the history of a schema in one reference model which represents all the aspects of repository-based information. Then, we use the defined reference model to create a system that supports both an interactive and a traditional way to exploratory analytics using the integrated contextual information about the schema histories. Beyond that, we use the same meta-model in order to group the entire lifetime of a database into phases, to which we refer to the term release, and perform a study on how these phases are related to changes affecting the schema of the database. Based on our findings, we can argue that change is mostly absent or kept in small numbers in contrast with few releases collecting a large percentage of the changes.
Institution and School/Department of submitter: Πανεπιστήμιο Ιωαννίνων. Σχολή Θετικών Επιστημών. Τμήμα Μηχανικών Η/Υ & Πληροφορικής
Subject classification: Open source software
Keywords: Εξέλιξη σχήματος βάσεων δεδομένων,Ανοικτό λογισμικό,Αποθετήριο ανοιχτού λογισμικού,Schema evolution,Open source repositorres,Mininy software repositorres
Appears in Collections:Διατριβές Μεταπτυχιακής Έρευνας (Masters)

Files in This Item:
File Description SizeFormat 
Μ.Ε. ΠΑΠΠΑΣ ΑΘΑΝΑΣΙΟΣ 2017.pdf3.46 MBAdobe PDFView/Open

 Please use this identifier to cite or link to this item:
  This item is a favorite for 0 people.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.