The Information System ICE

ICE stands for Information, Controlling, Entscheidung (in English: Information, Controlling, Decision). It is a web-based information system, developed to support higher education policy decision making. We are planning to approve ICE as free software under open source licence (GPL), so the use of ICE would be free of charge for everybody. The back-end of the system is a data warehouse, in which data of various structuring can be saved. Both individual data and aggregated data can be stored. You can search through the data via the ICE-application server. The main analysis tool is the flexible table generation that provides interactive compiling of user-defined tables out of the existing data stock. Furthermore, tables created with the flexible table generation can be saved as so-called standard tables that can be updated automatically, after importing more recent data stocks.


The debate on protecting personal data which arose in the 1980s led to increasing sensitivity in the population regarding data protection questions and to stricter data protection rules. Both resulted in fewer possibilities of statistical data analysis being allowed: In many fields, statistical analyses can only make use of so-called aggregated datasets. The ICE information system provides a solution that is capable of extracting a maximum of information from fundamentally limited aggregated datasets. At the same time subject to their appropriate availability the system also allows the analysis of individual data records.

At the start of the 1990s, ICE evolved as a commissioned project on behalf of the German education and science ministry (BMBW), which later merged with the research and technology ministry (BMFT) to form the federal ministry of education and research (BMBF). After the ministry converted its operating system from Macintosh to MS Windows, a redevelopment of the ICE system became necessary, in the mid 1990s. An intranet system was developed based on a modern multi-level architecture, with a database at the backend and Java as its central development platform. Thus the system became platform independent and accessible with an internet browser.

At present, the following institutes use information systems developed on behalf of the BMBF by DZHW (German Centre for Research on Higher Education and Science Studies):

  • BMBF: Statistic data portal of the BMBF
  • Federal states: ICEland - joint information system of all German federal states
  • DAAD: ICEwwo - information system of the Wissenschaft weltoffen project
  • Lower Saxony Ministry of Science and Culture: ICEnds - information system of the state-specific information demand of the federal state Lower Saxony

System Characteristics Outline

Web-application. Access to an ICE installation is enabled via a network using a Java capable web browser (such as the open source browser Mozilla Firefox or MS Internet Explorer). In principle, this means that the system can be accessed from any computer on the Internet or intranet which is registered with the ICE server. Access to the system (or to parts of it) can be restricted to authenticated users as required. Where necessary, the system (or parts of it, e.g. libraries of standard tables) can be set up to allow access from the Internet.

Platform independence. On the server-side, the system can be installed under MS Windows as well as under Linux. The client end (user) only needs a Java capable web browser. Browsers are available free of charge for all common platforms (such as MS Windows, MacOS, Linux). The system is also independent in terms of the relational database management system that is chosen: There are installations under Oracle, Informix, MySQL and PostgreSQL.

Flexible data import. Data with any structure and depth of structure can be imported. The system can also be expanded to include new topics. It is possible not only to analyse and interpret aggregated data but also comprehensive individual data with good performance. Similarly, the combined analysis of aggregated and individual data is possible.

Flexible data analysis. The very flexible import of data stocks is mirrored by an equally flexible range of analysis and interpretation modules. The so-called flexible table generation is a tool that can extract any needed parts of the existing data stocks into tables. Analysis using data from various data stocks can also be requested easily: It is no problem to combine information from several data stocks in a single results table.

Flexible data export. Users can choose between several output formats for their research results: XML, (X)HTML, MS Excel, Gnumeric, FOP, PDF. In the background the output is carried out as ICE-Publishing-Framework that is based on Apache Cocoon. The system is based upon XML. XML is a standard which can be used for data exchange and as a universal interface to third party programs (e.g. other databases, spread sheets, graphics programmes, geographical information systems, etc.). By use of XSLT-stylesheets XML-streams will be transformed into the further destined formats. If required, other formats (e.g. LaTeX) can be provided easily by the generation of other XSLT-stylesheets. This procedure allows comfortably layouting (colours, document size, fonts, etc.) for all outputs of an ICE-installation. This is therefore easily, because, when called, all outputs are generated on-the-fly, with all information extracted directly out of the database, and there is no need to access prefixed data files. There is an extra user interface for the common used PDF-format. This interface allows to modify the body structure of single PDF-files (e.g. column and preceding column width, line height, paper size, etc.).

Data harmonisation with an integrated key. All data contained in the system are encoded with a uniform ICE-key. The project team centrally updates and hosts the key-system. This ensures that as far as meaningful for the content various stocks, also possibly from various sources, can be analysed together. Equivalency rules are defined where necessary to make possible comparison of variously encoded data, which do have like content however, (example: subject groups in the staff statistics ? subject groups in the student statistics). The system also recognises key internal hierarchies and places knowledge at the disposal of the user for carrying out sorting functions, for example. (The system "knows" that the University of Hanover belongs to the state of Lower Saxony and to the higher education institution type "university").

ICE-standard tables (with integrated automatic updating). All results tables produced using flexible table generation can be stored as so-called ICE-standard tables in the standard table libraries. These table libraries can also be made accessible to third parties on the Intranet or Internet, and can be searched both by a hierarchical directory structure as well as by a keyword search. The integrated automatic update is a particularly useful feature of the ICE-standard tables: A once generated table, which has been stored as a standard table, can automatically be updated at the touch of a button with data imported into the system at a later point. The user can choose from various update options (e.g. time series addition, time series shift, substitution of the whole table with the latest available data). Using the ICE-Publishing-Framework, it is possible to request these standard tables in various formats.

Performance/High-availability.To ensure a good system performance and high-availability, it is possible to install the ICE-system on a computer-cluster. So user inquiries can be handled by any number of servers and not by a single one.

Support/Hotline. The ICE-project group of DZHW advices on all questions related to the information system (via e-mail as well as via phone). Besides technical questions (network problems, security settings) and questions concerning software-handling (browser, ICE-software), this includes specialist/statistical questions (concerning data and their analysis, key-questions, or the like). For users of ICE-installations that are operated by DZHW, this service is free of charge.

Continuing progress. The ICE-system lies under constant progression. Enhancements and optimisations that are done by the ICE-project group for a client, will be provided promptly for the general public.

ICE is going to be free software/Open source. Endeavors are being made to provide the source code of all ICE-components as free software under open source license (GPL) for the general public. This would allow anyone to use the ICE-system for any means, to study it, to modify it and to publish it in its original or derivative version.

The GPL allows anyone the following four permissions as integral part of the license:

  1. The work can be used for any means and without any restrictions. This explicitly includes commercial purposes.
  2. The work can be copied and redistributed free of charge or for a fee, and the source code must be accompanied. The recipient must get the same permissions e.g. anybody who gets a copy for a fee is authorised to publish it for commercial purposes or for free. A license fee is not allowed. Nobody is committed to distribute copies, neither in general, nor to a specific person but if a copy is distributed, then it must be under this rules.
  3. The operating principles of a work are free for studying, and it is legal to modify it to one's own needs.
  4. Even the derivative version of a work correspondent to Permission 3 can be distributed under the conditions of Permission 2, and the modified source code must be accompanied. Nobody is committed to distribute derivative copies, but if a derivative copy is distributed, then it must be under the rules of Permission 2.