The suspected Cobol-module seems to work flawlessly on our test system. At IWIS the system went down. New disk I/O-drivers (PP-program) were compiled and added to the system. No improvement. Where were these differences between their and our NOS/BE testsystem? Did we omit some system changes? Ultimately, we figured that our security protecting the PP was reading more bytes and overflowed its parameter list, an error that is still often exploited by hackers. At IWIS, this happened unexpectedly as their Cobol program went outside his table boundaries. As soon as our code was added to their system, they could start in determining the root cause for all the problems.
Another external Armed Forces user had read something about privacy regulations
those were being discussed in public, something that later became the
We were unpleasantly surprised when the Staatscourant published a note stating
that the computer at our Laboratory contained the complete personnel database
of one of our Services.
The Laboratorium was obviously the "holder" of the database according to the new draft law. As during that period many activists were very active in attacking defense sites, we did not like this type of public exposure....
The members of the Computer group took part in those discussions as well. Systems programming was very pragmatic. The programming group took part in a working committee of the NGI. Those (wo)men were more purists looking at Dijkstra. Others had read "something" in the computer journals.
System programming was largely surprised on day when they retrieved the punch cards for a system interface from the vault. The system interface drove the Calcomp-plotter. It turned out that all cards with comments and description of the interface (self-documentary) were removed from the card deck. As structured programming "quality requirements" required that a subroutine was not larger than a single printed page, one of the colleagues had removed all comment cards from the deck. Now it fitted on one single page...
Another colleague, had heard a bell ringing somewhere far away! He tried the opposite.
How to get as many Fortran statements on one single punch card?
Explanation of what could be a misunderstanding of the principles of structured programming took a lot meetings with them...
To support structured programming in Fortran, the software package FTS (ForTran Structured) was bought and installed at the end of 1979. FTS was a "pre-processor" which recognised constructions like IF- ELSE - ENDIF, DO..ENDDO and CASE-structures. After parsing, Fortran'66 code was delivered. When Fortran'77 became available, a FTS5 was developed in order to generate proper and more efficient Fortran'77 code.