Hacker News Reader

why would someone right their own tests when all they're trying to do is point out why the current ones are flawed.

To see if it performs better for their particular use case?

You know, choosing something based on how well it fits the problem?