msb@sq.sq.com (Mark Brader) (02/16/89)
I wrote: > In Bell Canada territory, at least, organizations whose names are just a > sequence of letters are alphabetized in the phone book as though each letter > was a word. And the alphabetization is by words: spaces and punctuation > except apostrophes which are ignored, sort together at the beginning of > the alphabet. ... It's not actually quite true to say that the alphabetization is by words. The actual sorting criterion is more complicated. You see, within each letter of the alphabet, all names that begin with single-letter words are listed first... sorted according to how many single-letter words there are at the beginning of the name! Only within that are they alphabetical. This means that the way to get at the front of the Toronto phone book is to begin with a single letter A as the first word, followed by a word that is not a single letter but does begin with multiple A's! The actual first entry in the 1987 book which I have at hand is "A Aacident Towing...", followed by "A Aaron & Son". Only after about 1.6 pages do we get to "A A" (that's a complete name) and "AA Aaba Locksmiths...". No, I'll give that name in full. It's "AA Aaba Locksmiths A Div of Grant's Lock & Safe Co". I bet they answer the phone "Grant's Lock & Safe"! They have a listing that way too. Notice, as I said before, that AA is treated as if it was A A. The names with 3 A's begin about 1.8 pages in, and the 4 A's that would be at the front in the old system, about 1.9 pages in. Have any other phone companies adopted such a style to defeat the standard aaaaaaaaaaaattempt to get the first listing? Note, I don't imagine anyone is interested to know whether the first name in the Los Angeles or New York phone book begins with 6 A's or 7. The point that is supposed to be of interest here is the trick alphabetization. However, while I'm writing I will give the last two names in the 1987 Toronto book. They are "Zzootz Hair Design" and "Zzzyed Himy". Finally, here's an interesting historical note, taken from Section 6.2.1 of Knuth('s The Art of Computer Programming). The concept of listing words in lexical or alphabetical order was invented many centuries later than that of an alphabet with a definite order to the letters. Formerly the sorting was generally only on the first letter, or maybe the first two. For *numbers* using a positional base system, such as the Babylonian one, lexical order is the same as numerical order, and this had been used as early as about 200 BC. But nobody thought to apply the same technique to words... until Giovanni de Genoa used it for his "Catholicon", in the year 1286. He explained the system in his preface and noted that "strenuous effort was required to device these rules ... do not scorn this great labor of mine and this order as something worthless". He was right. Mark Brader "You can do this in a number of ways. SoftQuad Inc., Toronto IBM chose to do all of them... utzoo!sq!msb, msb@sq.com why do you find that funny?" -- D. Taylor [Moderator's Note: Mark, thanks for an *excellent* essay. Here in Illinois Bell land at least, telco reserves the right to refuse a directory entry to anyone if in their (telco's) sole discretion the request entry is an attempt to defeat normal alphabetical listings or otherwise cause an impropriety in the directory. Thus a person or business wishing to be listed as 'A' must produce some evidence; i.e. state business license, etc, showing the actual existence of such a firm. A person or business desiring a listing using a word deemed vulgar or offensive must demonstrate that such a person or business does in fact exist or is in residence where the phone will be installed. P. Townson]