Saddle Points: use row, column references in specification

[dcr8898]

The specification for Saddle Points references matrix points by column, row:

1. "saddle point at column 1, row 2, with value 5"
2. "It has a saddle point at column 1, row 2."

However, the canonical data (and the corresponding tests in the Ruby track) reference matrix points by row, column:

1. "The test cases list saddle points ordered by row ascending, then column ascending."
2. "expected": [ { "row": 2, "column": 1 } ], etc.

While the canonical data does allow implementors to "choose a different ordering logic and adjust the expected output accordingly," the current disharmony between the specification and canonical data (including the current Ruby track tests) are potentially confusing to students.

In the Ruby track, the test results do not identify row/column in the expected results, they simply return two element arrays. Students, who only have the specification to rely upon, may incorrectly structure their responses in column, row format and experience failing tests without constructive feedback for the cause.

Unless this is intentional behavior 😜, I would like to submit a pull request changing the two references in the specification to row, column format in harmony with the canonical data.

Thank you!