Newsgroups: ukc.misc Subject: Another reason to be a stickler for correct spelling Reply-To: mg@ukc.ac.uk (M.W.Guy) Distribution: ukc Organization: Computing Lab, University of Kent at Canterbury, UK. Here follows a random list of events which I thought it might be interesting to publicise for the edification of all. Not one of my better articles, I'm afraid. I don't think "edification" means "to turn into an edifice", but it ought to. ----- I just found a message in the mailer's log files, saying: 6/15 10:08:42 local 6579: Unknown action to perform 'distroy' This turned out to be somebody's .maildelivery, which wasn't working because of the spelling error. ----- Back in the good old days, when men were real men, women were real women and hardware looked like the ICL 2960, Jim Darby wrote a news system for version 2.10 of the News transmission protocol. It was the first implementation in Europe and its completion annoyed the incumbent Unix Support Officer, who I will call "Mike" because that was his name. In the news header lines, however, Jim had spelt "Received" wrongly, exchanging the e and the i. Because there was only one occurence of the word "Recieved" in the code (as there should be), to which all interested parties referred, the system was consistent in its misspelling, and the parts worked together. Then Mike tried to interface it to the Unix version. When his Unix end encountered the header line it did not recognise, it went "Bleeyeh" and corrupted its database. This took Mike three days to fix. ----- When Chris and Sean tried to compile a new version of the mailer last summer, the final linking couldn't find a routine called apdm_normalise(). Grepping through the code, there were many calls to the routine, but no routine itself. How come *one* function was missing? To cut a long story short, take the bull by the horns, and aviod cliches like the plague, there was a routine called apdm_normalize() (note the Z) but it was called sometimes as apdm_normalize, and sometimes as apdm_normalise. The compiler on which the program had been developed must only have taken note of the first 12 or less characters of identifiers and treated both forms as the same identifier. ----- There have been lots of cases of spelling and typing errors changing the meaning of comments into lies, or of ambiguous abbreviations in comments. Wheeze. ----- Who can be blamed for misspelling "appropriate" when the "man" program supplied with Unix says "Nothing apropriate" at you N times a day? Sigh. Moral: None prepacked. Draw your own conclusions. Martin