Ice cream zero

Eskalator kosong!

A Malay adjective that you will hear often in Malaysia is kosong meaning zero, empty, nothing. Naturally it's used to describe car parks, fields, buses, and that sort of thing, but it's also used as a modifier to indicate absence. For example teh kosong means black tea, and kopi kosong means black coffee, ais kosong means a glass of ice without water. However the Chinese, ever parsimonious with syllables, had teh-oh and kopi-oh. There is no ais-oh though.

I was reminded of this adjective when I recalled taking a visitor from Singapore around Sydney many years ago. At a Manly ice creamery she liked the smell of the waffle cones but didn't want to have ice cream. She asked the shop if they could sell her just a waffle cone, and they were. I quipped that she was having an ais-krim kosong.

IBM 1130 FORTRAN with localised subroutines in Malay

In 1967, University Malaya acquired an IBM 1130. It may have been the first general mainframe in the country. An academic friend of the family invited me, a schoolboy at the time, to join an introductory course in FORTRAN, intended for other academics, officials and industry to learn about programming and how computing would change the country. It certainly changed my life as it was the start of my interest in computer science.

I could go on about the incredible increase in computing power, about how a Raspberry Pi is a more powerful computer than the 1130 ever was, but this blog isn't about that well-trodden ground. Neither is it about 1130 lore, there are sites for that, like this one.

Rather it's about one aspect of our course which has always intrigued me, and I may never know the answer to. As FORTRAN programmers know, formatted I/O is one of the more difficult areas of the language, and remember the strings were Hollerith constants, the CHARACTER type didn't arrive until FORTRAN 77. So the course presenters gave us a couple of free format input and output subroutines: CALL BACHA() and CALL TULIS(). BACHA and TULIS are read and write respectively in Malay.

If I recall correctly, not only did these routines take both INTEGER and REAL arguments, but were variadic. Probably impossible to provide with a subroutine written in FORTRAN. Now we can think of myriad ways of implementing this in modern languages. But the IBM 1130 had comparatively little memory, measured in thousands of words. My guess is that IBM technical support provided these routines, in assembler, localised for Malaysia. But I may never know.