A CATALOGUE OF PHARMACEUTICAL DRUG NAMES

Open hugovk opened this issue 5 years ago • 3 comments

Code: https://github.com/hugovk/NaNoGenMo-2018/tree/master/drug-names Output MD: https://github.com/hugovk/NaNoGenMo-2018/blob/master/drug-names/catalogue.md Output PDF: https://raw.githubusercontent.com/hugovk/NaNoGenMo-2018/master/drug-names/catalogue.pdf Words: 50,027 Notes: https://github.com/hugovk/NaNoGenMo-2018/blob/master/drug-names/notes.md

This is a reproduction of historical procgen shared by James Ryan:

extremely early #procgen: Pfizer used an IBM 702 in 1956 to generate 42,000 prospective drug names, which were compiled into a printed book!

'Brain' coins 42,000 Words in 2 Hours - Would Take a Human 269,000 Years

Pfizer seems to find it easier to rout disease that to create names for their new products. In the past the firm employed a philoligist or two who scanned the names of the contents of a drug. Selecting a syllable here and there from the principle components, they patched together a few pronouncable, spellable, descriptive names and submitted them for adoption and trademark.

But the task grows increasingly complicated as new drugs appear, not only Pfizer's but those of a number of renowned pharmaceutical firms. The public has not needed to learn all the jaw-breaking words, as many of them are sold by prescription only. But doctors are stymied at loading their minds with ever more complicated names whose sounds tax their spelling, and whose roots go back to forgotten pharmacology terms. With human brains groping for new word-sound combinations, the forementioned pharmaceutical firm appealed to a business machine corporation for help. It has just arrived in the form of a rare book published by IBM, and of which there are only three in existence.

The 198-page volume was written for Charles Pfizer & Co., Inc., of Brooklyn, pharmaceutical and chemical manufacturer, whose researchers have been having trouble finding new names for its products.

IBM's "electronic data processing machine" was fed an outline of the problem on a magnetic tape: combine 30 groups of word endings like "il," "in," "ite," "ide," and "ane" with various combinations of one and two syllable words. Then, come up with words easy to spell, remember and translate into medical terms.

The IBM machine also was taught to be discreet. It was fixed to automatically eliminate four-letter combinations that wouldn't be proper in a family medicine chest.

Like the IBM machine, the new implementation was "taught to be discreet" and "fixed to automatically eliminate four-letter combinations that wouldn't be proper in a family medicine chest". And unlike the original, they have been added as an appendix.