Clean up usage of "Tuple"s

[dakusui]

Right now, JCUnit uses a Tuple for an input of test case, a row in a covering array, and a combination of values.
This may prevent us from optimizing a performance of covering array generation.
At least model-wise, it would be preferable to define different interfaces for each of them.

• Tuplet: A combination of values in one Tuple, typically its width is the strength of a covering array and is used in covering array generation and combinatorial joining operation.
• Tuple: A row in a covering array. Its width is the degree of covering array.
• TestInput: Key-value pairs that represents input-part of a test case.

[dakusui]

@qxshower , any suggestion/preference on this terminology?

[qxshower]

Below is my suggestion:

First, you may want to introduce that a covering array is a test suite, and a row of a covering array is a test case.

Tuplet --> Tuple: A combination of values in one [Row], typically its width is the strength of the coverage of interactions in a covering array. It is used in covering array generation and combinatorial joining operation.
Tuple --> Row: A row in a covering array. Its width is the degree of the covering array.
TestInput: Key-value pairs that represents input-part of a test case (i.e., a [Row] in a covering array).

[dakusui]

Thanks.
Let's take your suggestion.
Tuple and Row would become names of interfaces while TestInput(testInput) would be used for a name of a variable.

[dakusui]

This is currently a part of this pull request.: #167
(Although I should have made it an independent one...)