Skip to topic | Skip to bottom


NicolasDintzner.WebHomer1.21 - 25 Oct 2016 - 09:50 - NicolasDintzner

Start of topic | Skip to actions

Nicolas Dintzner

PhD Candidate
nicolasdintzner.jpg Office: HB 08.080
Phone: +31 6 23157647
E-mail: N.J.R.Dintzner (atDOMAIN) tudelft (dot) nl
Delft University of Technology
Department of Software Technology
Software Engineering Research Group
Mekelweg 4
2628 CD Delft, The Netherlands

I am a Ph.D candidate in the Software Engineering Research Group (SERG), department of Software Technologies of the Delft University of Technologies. My work is supervised by Prof. Martin Pinzger and Prof. Arie van Deursen.

I work within the Allegio Project, in close collaboration with the Embedded System Institute (ESI) and Philips Healthcare. The Allegio Project aims to improve software development methodologies, tools and strategies in order to accelerate the integration of new technologies in the medical embedded systems. As a member of the Allegio team, I am focused on the evolvability aspect of the sofware architecture.

FEVER: Feature evolution - bigger and better!

We've just completed as study on the evolution of the Linux kernel! The new tool and dataset are available now!

More details on the approach below, but we extended it recently. We now take into account a few more changes than before, and track more modifications at a feature level. Do you want to know how many times a new reference to I2C was created in the Linux variability model ? We can tell you.

We have a complete Neo4j Data of 15 releases of the Linux kernel, broken down by commit and features for your querying pleasure. You can find the dataset here (official TU Delft data repository). (dataset DOI : 10.4121/uuid:c478028a-ac6d-4c45-9e2d-8ad63c7ca75f). Just uncompress the archive, and copy the release that interest you in your Neo4j server (in the /data/ folder). Rename the folder you copied to "graph.db" and you are ready to query.

The tool we used to create the dataset (the FEVER prototype) and the spreadsheets used for our analysis are here, with all its enhancements. You can use the tool to create your own dataset, for complete releases, or commit ranges of your choice. The instructions for installation and usage are inside the Git Hub repository.

FEVER: Feature evolution (old)

FEVER: Extracting Feature-oriented Changes from Commits

We devised an approach we call FEVER: the Feature Evolution Extrator; an approach to automatically extract feature-oriented change information from Git commits. We have compiled all the information related to this study in a Git repository that you can find here. You'll find the source code of the tool and dataset related to our evaluation and findings on the evolution of the Linux kernel. Feel free to browse around!

The source code is available with installation instruction.

Due to size restriction, the dataset is in Dropbox rather than Git. Sorry. You can get here. To access the data, all you have to do is install a Neo4j server download, very easy and fast, and then replace the graph.db folder with the content of the zip files provided here. And you are all set!

Software Product Line Research

Feature-oriented Software Evolution

Our doctoral symposium work on co-evolution patterns got accepted at ICSE 2015. The camera ready version is available here (preprint).

Feature model evolution

As part of my research for the Allegio Project, I study the evolution of large scale product line. In this context, I developed a tool FMDiff to automatically extract feature model changes from the Linux kernel. The paper describing this work can be found here: Extracting feature model changes from the Linux kernel using FMDiff. This work was accepted and presented at the VaMoS 2014 workshop. The publication can be viewed here and the associated presentation is available here

You can access the tool and the dataset we used for our study in our git repository

Change Impact analysis on Multi product line

We developed an approach to evaluate the impact of feature changes on the existing configurations of a multi-product lines. The objective is to assist our industrial partner in assessing the effect of a component change on the external interfaces on a large variable system, modeled using feature modeling techniques. This work resulted in a paper that will be presented at 14th International Conference on Software Reuse. You can access the tool and dataset we used for its evaluation in this git repository: MPL Change Impact

Edit | Attach | Printable | Raw | More topic actions
Revisions: | r1.21 | > | r1.20 | > | r1.19 | Page history | Backlinks
You are here: NicolasDintzner > WebHome

to top

Copyright © 2003-2018, Software Engineering Research Group, Delft University of Technology, The Netherlands