Paul Rubin wrote: > "Michael Tobis" <[EMAIL PROTECTED]> writes: > > The first piece of code that I ever voluntarily wrote was intended to > > solve this puzzle: > > > > Assign the number 2 to 'a', 3 to 'b' ... 27 to 'z'. To each word assign > > the value of the product of its characters. Find the English (or > > language of your choice) word whose product is closest to a million (or > > number of your choice). > > Hey, that was a contest in Games Magazine in the 1980's. A co-worker > and I used a PDP-10 BASIC program to search for numbers near 1 million > with no prime factors higher than 27. The factorizations of those > numbers told us which letters to try to use, and after a while of > fooling around rearranging letters on a whiteboard, we came up with > "curvy", a very recognizable word that multiplies out to 999,856. > > That's the closest number to 1 million (other than 1 million itself) > which doesn't have any prime factors that are too large. We convinced > ourselves that there was no word that multiplied to exactly 1 million, > so we felt we were likely to win the contest. However, somebody > apparently with a computerized word list won with "ixodid". It was > fun to remember this.
The trouble with word lists is when you run across something you don't recognize, like "ixodid", you can't tell if it's a word or an acronym or an abbreviation. Being in the environmental remediation business, I thought "dioxid" (which I assume is related to "dioxin") to be more plausible as a word. And it would seem that Games no longer has contests that simply find words, they add a twist to prevent those like me from having an advantage. See <http://members.aol.com/mensanator/pig_ignorance.htm> I still used my word list database to find valid words to use in my contest entries, but entries are judged, not simply scored. -- http://mail.python.org/mailman/listinfo/python-list