uni'wissen 1-2013_ENG

The linguists were at a loss: Their digital data base was suddenly unreadable. The problem wasn’t the data – they were still there – but the new computer system the department had installed, which was no longer capable of read ing the old data formats. The entire local lan guage atlas, the map material, and the transcriptions of audio documents from the years 1974 to 1986 – simply gone. A linguistic treasure trove, assembled in the course of many years of hard work, had suddenly vanished into thin air. Luckily the linguists had the idea of asking for help from the IT Services Department of the Uni versity of Freiburg. Director Prof. Dr. Gerhard Schneider and his team succeeded in retrieving the valuable data: by means of emulation. Emulation refers to the procedure of making the original computing environment on which digital data was stored available again or creat ing a similar environment on which it can be read. Schneider uses an example to explain what this means: “If I want to open a Word docu ment created on an old version of Word on my computer today, I either won’t be able to open it at all, or the program will tell me that the docu ment needs to be converted in some manner.” At best, the computer will still be able to read the information contained in the document, but the original form will be gone: the formatting and representation of the data as well as the socalled macros, the preset functions of forms. That might be inconsequential for casual users, but for scientific research it can be a fatal devel opment. After all, what is the value of data that can no longer be read? Consider the case of the linguists, for instance: They had begun filling up their database in 1993 – using hardware that is now obsolete, and modern computers simply refused to read the old data. The only recourse in situations like these is to “trick” the modern technology into thinking it is actually old. Hardware emulators are special applications that run on current systems and imitate an old system as authentically as possi ble. This allows the users to access and work with the data originally created on the old system. “Emulators bridge the technological gap between outdated and current environments,” explains “No one wants to spend millions collecting data only to find out one day that they are no longer comparable” Klaus Rechert, a research assistant in Schneider’s team. This works with all kinds of data, including digital art. Documenting Processes, Describing Surroundings The idea of emulation goes even further: Data only retain their true value if the means by which they were created is known. “No one wants to spend millions collecting data only to find out one day that they are no longer comparable,” says Schneider. This “comparable” is what he and his team are working on. When generations of PhD students work with the same data, for instance, the procedures and tools they use will change again and again. All it takes is for them to be rearranged, recalculated, or sorted accord ing to new criteria – and hardly anything will be left of the form in which the data were originally created. “It’s like with the game of telephone: with each new user, with each refinement to the system, practical knowledge is lost,” explains Rechert. The computer scientist thus stresses the importance of also archiving this non documented knowledge, this work with the data, along with the data itself. Processes need to be documented, environments described as pre cisely as possible, in order to keep the data available for comparison or general use later on. The emulators make it possible to document the metamorphoses the data has undergone. Not only does this enable the data to be repro duced in the future, it also provides a means of determining the origin of dubious data. “If we know the path by which they were created, we can ﬁnd out more easily whether there is a sys- tematic error that has been carried over through all of the changes to the system or whether it 0 0 0 11111 1 1 Antiquated technology, use- less information: Data stored on outdated computers are often no longer readable by current models. The solution lies in imitating old systems on modern computers by means of emulators. Photos (montage): Carlos Castilla, Maksym Yemelynov, sound- snaps (all Fotolia) 17

uni'wissen 1-2013_ENG

Pages