Word Ways


Frank Rubin


It is now approaching two years since I began a computer search for a 10x10 word square. Since my last report in the November 1989 Word Ways, there has been a discouraging turn in events. I furnished statistics about the distribution of letter-frequencies in my wordstock to a researcher at the IBM Yorktown Research Center. He calculated, given these letter-frequencies, that a vocabulary of 289,000 words would be required to produce a symmetric 10x10 word square. I had already projected that when I finished transcribing the Levine list, there still would be fewer than 130,000 words -- less than half the required number.