Dr. Reinhold Thurner, Founder and CEO

Software-Tools need a Repository

Many, many years ago I founded Delta Software AG - a company specialized in software tools. We used files and some relational databases to manage the components used by our tools. As the number of users and the size of the system grew we recognized the need for a powerful data infrastructure – a data infrastructure which can handle a large number of components and also the complex relationships between these components.

A Repository - the Integrating Infrastructure

We could not find such a thing and decided to build a repository based on the ISO-IRDS-standard. This was quite a challenge because the memories were too small, computers too expensive and the development of user friendly systems in C was a real nightmare. The repository worked reasonably well but the architecture suffered from these limitations. We shared this experience with some other dictionaries in the field which still exist. 

A Repository in Architecture and Reengineering

I moved on and became responsible for the architecture in IT operations and later in the reconstruction of a banking system.

I encountered the same problem as before: A sort of Shadow-IT with gigabytes of complex data scattered in tons of Excel-sheets, private relational databases and files, files, files. The integration was euphemistically called copy-management. To master all this complexity I used the old repository once more. Under the given circumstances the result was satisfactory – but we were not really satisfied.

New Technology - New Solutions

We asked ourselves:  Should we stick to what we have and pile new features and tools on top of an outdated C-kernel or should we rebuild it from scratch? Not really easy to decide!

We went for a new architecture. We had the knowhow and most importantly new crucial technologies were available:  Big memories, efficient computers, powerful networks and above all Java and frameworks (Eclipse).

From a Repository to a General Purpose DBMS with an ER-Model

The system morphed from the initial idea of a tool repository into something beyond a tool repository, beyond a data dictionary, beyond a configuration database – into a database management system which supports all this type of applications and many more. It took years of research and development to create Metasafe - an entity-relationship based database management system.

Peter Chen spoke once of an "executable conceptual model" which removes the semantic gap - also called impedance mismatch - between the conceptual view and the access to the database.

Metasafe integrates the conceptual models and the instance data into one integrated system. Metasafe provides the necessary set of tools and features a graphical query language for an easy access to the database. I am sure that this solution is good news for people who have to deal with complex data, for software developers and for people responsible for metadata or just for simple user friendly applications.

An Innovation Beyond Narrow Minds

The system is the work of many people who helped, asked, criticized, designed, programmed, tested and used the system in its intermediate stages. I was very fortunate to have met the right people and to be given the opportunity to create the Entity-Relationship database management system. The timing was right - the required infrastructure was available and it was needed to enable users to handle complex data in a user friendly manner.