Legka Но как определить насколько отличаются (насколько похожи) строки:
Super shampoo - 8 oz и Super shampoo 8 oz ?
Есть такой алгоритм Soundex:

"Soundex" algorithm is standard Odell/Russell (1918):
*
* Produce a code of the form "letter, digit, digit, digit"
* using these rules:
*
* 1) Retain the first letter unchanged.
*
* 2) For each succeeding letter, produce a result based
* on the following table:
*
* letter result
*
* B, F, P, V digit 1
* C, G, J, K, Q, S, X, Z digit 2
* D, T digit 3
* L digit 4
* M, N digit 5
* R digit 6
* A, E, H, I, O, U, W, Y (nothing)
*
*
* 3) If two or more adjacent letters produce the same
* result in step 2, ignore all but the first of the
* adjacent letters.
*
* 4) Repeat steps 2 and 3 until three digits have been
* produced or until the source is exhausted.
*
* 5) If less than three digits were produced, right-fill
* with zeros.
*
*
* Notes:
*
* Non-alpha characters are ignored entirely; letters which
* are separated only by non-alpha characters are considered
* adjacent. If the source contains no alpha characters, a
* value of "0000" is returned.
*
* Case is not significant.
*
* Letters which produce (nothing) in step 2 are still
* significant with respect to step 3. That is, two letters
* which produce the same digit are not considered adjacent
* if they are separated by a letter that produces (nothing).
* This is in accordance with the original algorithm.
*