At a party recently, I met an elderly gent called Bruce. An affable and well-spoken man, he happened to mention that he had worked in telecommunications and computer software from the 1950s onwards. In the US, he was employed by Marconi at a time when the first inter-exchange trunks were being established, and their signalling protocols were being designed. These served to connect local and regional telephone exchanges so that long-distance calls could be made without need for an operator. These days, telephone exchanges are connected with fibre-optic cables; microwave transmission is, I think, still common in less developed parts of the world. But Bruce worked at the time that predated even microwaves, he said. First VHF, then the 200MHz spectrum. Naturally I wanted to know all about it, and when he realised that I was a fellow telecomms type (at least in spirit these days), he was glad to talk. He travelled quite a bit, he said, all over the US, and, later, across Europe.
"Then I worked with the very first mainframes," he said. He spent many years with ICL, that British computing behemoth that fell apart, and was acquired by Fujitsu. Not that the union with Fujitsu was any more successful, of course. PCs were taking over the world, and even though many people at organisations such as Honeywell and Unisys and IBM and Fujitsu could see the writing on the wall, institutional inertia led to these mainframe manufacturers getting increasingly sidelined. But all that was in the future when in the 1960s he began working on COBOL.
COBOL! Who even mentions this programming language these days? The last I ever had anything to do with this was as part of a Programming Languages Lab at IISc, where the instructor was so bored and disinterested that I spent his classes playing tic-tac-toe with myself. That was in 1992. Short years later, enterprising Indian engineers were writing clever parsers and translators to convert COBOL to more modern programming languages in keeping with the large-scale effort on the part of mainframe owners to switch to PCs. They made more money than I can shake my finger at; then they parleyed all that experience in sorting out the Y2K bug for an increasingly panicked world.
Upon close questioning, Bruce revealed that he was instrumental in the internationalisation projects for COBOL. Essentially, the idea was to allow applications to provide seamless interfaces to users in various (human) languages without having to rewrite the underlying code. With all the expertise garnered from this effort, in later life Bruce became the editor for the first Unicode draft. He spent months, he said, going over every line of that 1000-page document. It aimed to standardise computer encodings for every written script on the planet. The ASCII codes for the Latin script could only provide 255 characters, and so in countries such as China where the script consists of a few thousand characters, they developed their own encodings. Without a streamlined system, Chinese computers would suddenly find it impossible to understand, say, a Russian one. I speak extremely loosely, of course. But with Unicode, you could have a uniform encoding for all possible languages - Amharic and Mandarin and even North American Cree. Is Klingon also Unicoded? No, but Unicode is flexible enough to allow it.
"You can thank me for my sleepless nights," said Bruce.
Others have depended on Bruce for his abilities. Even if he doesn't speak or read all the world's languages, he can identify most of them from their writing. At his local barber's, the proprietor, an Italian woman, once showed him a book left to her by her father. He recognised it to be written in the Devanagari script. He asked a colleague - a cellist - in his orchestra if she could read it. She, a Bengali Scotswoman, couldn't.
It turned out to be a Hindi book of poetry published in Gorakhpur. The Italian barber was keen to know if it was a rare book; if so, she might make some money selling it. He had to tell her, rather regretfully, that it was far from rare. The print run had been 120,000 copies.
"Then I worked with the very first mainframes," he said. He spent many years with ICL, that British computing behemoth that fell apart, and was acquired by Fujitsu. Not that the union with Fujitsu was any more successful, of course. PCs were taking over the world, and even though many people at organisations such as Honeywell and Unisys and IBM and Fujitsu could see the writing on the wall, institutional inertia led to these mainframe manufacturers getting increasingly sidelined. But all that was in the future when in the 1960s he began working on COBOL.
COBOL! Who even mentions this programming language these days? The last I ever had anything to do with this was as part of a Programming Languages Lab at IISc, where the instructor was so bored and disinterested that I spent his classes playing tic-tac-toe with myself. That was in 1992. Short years later, enterprising Indian engineers were writing clever parsers and translators to convert COBOL to more modern programming languages in keeping with the large-scale effort on the part of mainframe owners to switch to PCs. They made more money than I can shake my finger at; then they parleyed all that experience in sorting out the Y2K bug for an increasingly panicked world.
Upon close questioning, Bruce revealed that he was instrumental in the internationalisation projects for COBOL. Essentially, the idea was to allow applications to provide seamless interfaces to users in various (human) languages without having to rewrite the underlying code. With all the expertise garnered from this effort, in later life Bruce became the editor for the first Unicode draft. He spent months, he said, going over every line of that 1000-page document. It aimed to standardise computer encodings for every written script on the planet. The ASCII codes for the Latin script could only provide 255 characters, and so in countries such as China where the script consists of a few thousand characters, they developed their own encodings. Without a streamlined system, Chinese computers would suddenly find it impossible to understand, say, a Russian one. I speak extremely loosely, of course. But with Unicode, you could have a uniform encoding for all possible languages - Amharic and Mandarin and even North American Cree. Is Klingon also Unicoded? No, but Unicode is flexible enough to allow it.
"You can thank me for my sleepless nights," said Bruce.
Others have depended on Bruce for his abilities. Even if he doesn't speak or read all the world's languages, he can identify most of them from their writing. At his local barber's, the proprietor, an Italian woman, once showed him a book left to her by her father. He recognised it to be written in the Devanagari script. He asked a colleague - a cellist - in his orchestra if she could read it. She, a Bengali Scotswoman, couldn't.
It turned out to be a Hindi book of poetry published in Gorakhpur. The Italian barber was keen to know if it was a rare book; if so, she might make some money selling it. He had to tell her, rather regretfully, that it was far from rare. The print run had been 120,000 copies.
0 comments:
Post a Comment