Thank you Jim. I used something like that. As you can see that contains, contains, contains, can get pretty long.
This exercise was to solve a remaining “word” in a cipher type called a Tridigital. A short explanation is each message word is shown as a series of numbers (0 -9), one number per letter in the word, but each number can represent up to 3 letters. that gives 10 possible numbers but one number is used to identify a word space (you don’t know which one) and with the remaining 9 (3 x 9 = 27), one of them represents only 2 letters instead of three (you don’t know which one).
So I had the “word”, 34226870. As this was the only remaining word to decode in the message, I had already identified all but about 12 letters.
Long ago, in the world of 3.5 floppies, I bought a 300,000-word dictionary, and using Panorama, I broke it into word length groups with a secondary breakdown into word with no repeating letters and those with (pattern words). In this case, I opened both word-length eight files (.txt files) appending them together.
With around 12 - 15 letters to worked with, first I used the six-choice Select window and a select BeginsWith "or"ing the six choices (okay, it was Pan6 ), then did it again for the remaining candidate letters with SelectAdditional. Then I removed UnSelected.
From that, I repeated the process using EndsWith - again OR-ing the candidate letters and using SelectAdditional, removing the rest with Remove Unselected. That got my word list down to about 280.
From there, I switched to Formula Find/Select, with the structure A[2,2] = “C” or A[2,2] = “F”, or A[2,2] = “J” … listing the 12 to 15 candidate letters for each position. After each selection I Removed Unselected. It took only one more position (A[3,3] = …) selection to reduce the list to a number I could just review. Once the number of possible words was down to about 100, given the context of the message, it was easy to pick out the appropriate one without having to do further reductions.
The word was “cheering” and I had an idea that might be it from earlier pencil and paper work on the cipher. But when I looked cheering up on my iPhone dictionary, it didn’t have it. (it only had cheerily). So that’s why I went the extra mile to make sure some other word wouldn’t fit better. The context was unusual, “…it’s cheering to know …” rather than a more usual, “The crowd was cheering …” So looking at all the words that could fit was warranted.
I can see now that it’s not so much that the word itself contains a set of letters as it is that certain positions in the word contain one of the letters in the set.
So if I decide to make a selecting function, I’ll need to pass it a word list, and a set of candidate letters for each position in the word.
Panorama is so much fun. I don’t know why everyone isn’t using it.