3 Wegematic 1000 Memories


Index - - Contents - - Previous chapter - Next chapter - - Previous page - Next page

Aarni Perko, cont. 2

Wegematic used vacuum tubes. This meant that it needed continual service and monitoring. Its relays were also sensitive to mechanical faults. In general, the technical service was competent, and Wegematic worked correctly. I remember one occasion when it made an error in multiplication with a special bit combination. The reason was some limiting condition which did not emerge in regular tests. Control sums were used for both data and programs in paper tape and cards in order to facilitate validation of input.

The fixed point representation was a specific problem in scientific computations. The implied position of the binary point was marked in the programming sheets. The numbers were shifted before addition and subtraction in order to get a proper alignment. Surprisingly enough, fixed point calculations were not a very big problem in programming scientific problems.

Testing of programs was inefficient. We sat at the operation unit and in hairy situations one-stepped programs reading binary results from the oscilloscope. Usually the test run ended by the disgusting snarling voice of the overflow indicator. Although we had some trace programs their use was also slow.

Subroutine libraries were scanty. In later stages a matrix package and some kind of assembler were available. The main developers were the University of Uppsala and Chalmers University of Technology in Gothenburg, Sweden. The leading guru was Klaus Appel in Uppsala.

Punched card machines had their own horrors. Many times we had to punch a new card from a broken one. One time I had run a program 20 hours. I forgot one box of cards, and the whole run was spoiled. Programming of tabulators was also tricky and required plenty of experimenting.

Although obsolete already at start, Wegematic had all essential features of a computer of its era. Working with it gave plenty of hard experience in programming and running real applications.

What is different in computing nowadays? All and nothing. Everyone has now at his desk magnitudes of more power than Wegematic had. Fancy programs are available. But still computer nerds work 24 hour days with obscure bugs. Programming and system administration are very inefficient. Some projects run over time and budget, and are finally are thrown into the trash can. Media and formats do not fit. We were afraid of Y2K-chaos. And never, no never, Scandinavian alphabets will go through correctly!

The writer is by January 2000 Professor Emeritus from Jyväskylä University.