Speakers

biography | about presentation
Harry Sneed photo

Harry Sneed

ZT Prentner IT GmbH, Vienna, Austria & SoRing Kft.Budapest Ungarn (AT, HU)
The presenter, Harry Sneed, developed the first German testing tools and instigated the first outsourced testing project in Budapest, testing Siemens BS2000 programs on a fixed-price basis in the year 1978. Later he was to develop more than 15 different testing tools from unit test to acceptance test. for Information systems to realtime embedded systems. He has published some 100 papers and articles on Software testing and written 7 books on that subject. In 2011 he was awarded the first German prize for Software Quality and in 2013 he was selected as tester of the year by the ISTQB. He is still working as a Quality and Testing Consultant at the age of 76.

About the Presentation

Testing large databases – a case study

This presentation by a Software Quality Consultant deals with the subject of data quality. In some cases projects have failed because of poor data quality. Such was the case recently at the Austrian Social Security Office where a new system for administering the health insurance of pensioners was rejected because of inconsistencies and invalid data values in the migrated data. The project team had failed to test the quality of the data before releasing the new system to the users. A primary goal of the project was to join the data of several separate insurance providers in a common nationwide data base. As it turned out many of the codes used by the insurance Providers were incompatible. Besides that, many of the data values in the database were simply wrong or invalid. The data had been entered incorrectly from the start or it had become corrupted by errors in production. From the viewpoint of the user it appeared that the software was erroneous, whereas in fact it was not the software which was the problem but the underlying data. To correct the situation users were asked to define integrity rules for all key data items. These rules were then converted into formal assertions which could be automatically validated. The resulting data validity check uncovered over 90% of the incorrect data but it delayed the release of the health insurance system for another six months. Because of that some providers refused to accept it. The lesson learned was that data has to be tested just as the software is tested.

In partnership with

If you would like to be a partner in organizing the conference, please, contact Csilla Kohl (csilla.kohl@hstqb.com).


Our media partners in 2016