On Wed, 21 Oct 2015, James Propp wrote:
Does anyone happen to know (or have a way to easily determine) the average length of a Martin Gardner column, as measured by word-count?
I have a copy of the CDROM version of the columns right here. I used pdftotext to convert them to text and wc -w to count the words. The total was 1201144 words. I'm not sure how many columns this covers, but https://en.wikipedia.org/wiki/List_of_Martin_Gardner_Mathematical_Games_colu... says he wrote 297, but I don't think that includes the Flexagon article, which is on the CDROM. In any case, 1201144/298 is 4030.7, roughly. This is definitely an overestimate, because it includes the front matter for the 15 files that the CDROM gathers the columns into, and pdftotext breaks some words into multiple pieces. For example, here's how it renders a sentence from the first file: The resulting folded strip, illustrated by the second drawing in the series, is then folded back on the lines a b and cd [third drazoing], forming the hexagon [ f o u r t h d r a w i n g ]; finally the blank triangle is turned under and pasted to the corresponding blank triangle on the other side of the strip. All that said, 4000 words (about 8 pages of standard typing) is probably a good guess. -- Tom Duff. Password cannot contain continuums bit!