Computational Science and Engineering at TU Wien

Web Page: http://www.iue.tuwien.ac.at/cse/index.php/gsoc/2013.html

Mailing List: http://jwein2.iue.tuwien.ac.at:8080/mailman/listinfo/soc2013/

Projects

• [IoTSyS] A Discovery Service in OBIX Standardization OBIX is an open standard for information exchange in building automation. OBIX defines a standard for messages and end points representation so that enables interoperability between automated devices in smart buildings. The current OBIX standard currently lacks a service discovery mechanism that allows client applications to access OBIX servers without knowing the server addresses in advanced. This project contributes to the IoTSyS project, which uses OBIX as the underlying standard, by designing an appropriate service discovery (SD) method for the IoTSyS gateway. Furthermore, this project also participates in redesigning the IoTSyS OBIX object container to improve the way objects are addressed, stored and retrieved.
• [IOTSYS] oBIX 1.1 Specification WD Implementation IoTSyS uses oBIX (Open Building Information Exchange), an open standard for information exchange in building automation. It provides a generic object model and defines ways to read, write and invoke operations on objects. oBIX also provides certain services, like watch, history and alarm services. The current implementation is incomplete and only implements parts of the watch and history services. In this project, I am going to fully implement the latest oBIX 1.1 Specification Working Draft.
• Eclipse Array Explorer Improvement The Eclipse Array Explorer developed by CVAST at TU Wien University is a powerful tool created to make visualizing arrays and collections in the Eclipse Debugging framework easy. The explorer gives several options to replace the standard Eclipse framework. In its current state, the plug-in is powerful, but it needs to be updated in order to allow for widespread usage and adoption of the plugin by the Eclipse community. I feel that this plugin can easily be updated to become an even more powerful tool able to handle large arrays, collections, and maps.
• GDAL/OGR: Rigorous Support of Vertical Datums within OGRSpatialReference GDAL/OGR is widely used geospatial data abstraction library. 3D spatial data mainly consists of surface location and height. While surface location consider geometrical aspect of the Earth, height is modeled differently due to gravity. Although at some extent vertical coordinate transformation is supported by GDAL/OGR , the transformation chain did not compensate for height information. This project will establish a rigorous 3D transformation chain to support vertical datum.
• Integrating ViennaCL into the Python scientific ecosystem PyViennaCL will provide Python bindings for ViennaCL along with documentation and regression tests. The bindings will provide Pythonic objects and routines that represent the ViennaCL API, at least up to and including BLAS level 3. The bindings will also integrate coherently into the current Python scientific ecosystem, by supporting concurrent use of the NumPy ndarray, as well as aiming to provide a drop-in subset of the full NumPy API using ViennaCL as a back-end where possible.
• MOST: NoSQL Solution for Building Data Warehouses The current implementation of MOST database is in MySQL which includes processing of large amount of data. The data processing is done through the stored procedures. The aim of this project is to implement a java based framework which will support database independent data processing.
• Nomacs Plug-in System Nomacs is an open source image viewing and editing software. The two things that distinguish Nomacs from other image viewer are the synchronization and wide range of image formats it supports.The main goal of my project would be to create a plug-in system that would allow to extend Nomacs by keeping the main program small.
• OpenEngSB: Query transformation for multiple tool models based on Common Concept Model OpenEngSB provides an easy-to-use and easy-to-adapt platform for tool integration. There are two issues to be solved in the current OpenEngSB query transformation. The first issue is to let the users of OpenEngSB easily define and execute query for different data sources. The other issue is regarding the unfamiliarity of users with the SPARQL for ontology querying, as the current implementation of OpenEngSB is using ontology for storing their data. Other issue that is also related but will be addressed as optional is the querying of several different ontology versions. The aim of this project is to provide an implementation of SQL query user interface, transform it into SPARQL queries, and get the combined result data from heterogeneous ontology data sources for OpenEngSB
• OpenPixi: Parallelization of the Simulation Using OpenCL OpenPixi: Parallelization of the Simulation Using the IBIS Framework and OpenCL
• Python Interface for ViennaMesh A Python module shall be developed to allow the use of the ViennaMesh library from the interpreted programming language Python, thus enabling the user to easily take advantage of the power of ViennaMesh from a language that is simple to use and avoiding the long compilation times of C++.
• Skeptik: subsumption-based proof compression algorithms Skeptik is a program for converting resolution proofs produced by solvers into shorter, equivalent ones. To this end various compression algorithms have already been implemented. I intend to enrich this collection of algorithms to produce even shorter proofs while remaining in a reasonable computation time window. I would like to implement three new algorithms. Those are Subsumption, RecycleUnits and a variation of the already implemented RecyclePivots algorithm. The first two are described in the literature, but both also leave room for improvement. While I intend to generalize RecycleUnits, Subsumption needs some adaption in order to have the required efficiency.
• VIM: Integration of Survey Design Facilities into a Gtk2-based R-package VIM Main goal of this project, which is improving the VIM package for R, consist of two parts: adding a modern Gtk2-based replacement (based on the RGtk2-package: http://cran.r-project.org/web/packages/RGtk2/index.html) for the old tcltk-based graphical user interface and integrating support for the survey-package (http://cran.r-project.org/web/packages/survey/index.html). In addition to that other small changes like improving the usability and adding different new imputation methods are planed.