dk96-os / some-algorithms Goto Github PK

Benchmarking provides a way to compare each version of the algorithm.

To improve benchmarking further, there are two pathways:

• Add new data sets that are created from another source, or from a real-world model or example
• Enhance and automate dataset production, meaurement, and presentation of results.

Originally posted by @DK96-OS in #5 (comment)

Explore a variety of pseudorandom test data sets.

Ensure that all algorithms are in agreement for a specific data set.

In order to reuse test cases and compare algorithms easily, an interface will be created.

The method can no longer be static.

If a different algorithm is to be added to the repository, it would be best to create a new module because:

1. Clarity, Organization, and Focus
2. Running Gradle Tasks on one Module (rather than all) is easy
3. An algorithm may have some external dependency, for example if the algorithm input or output is defined using a specific class or interface. Dependencies are imported by a Module.

I would like to try improving an algorithm by structuring the problem in a way that can be divided into smaller independent problems.

Then, it becomes possible to compare methods of implementing concurrency in Java and Kotlin. Some comparisons can be made between Streams and Coroutines.