Verifying Tasks

What is verification?

While the purpose of code review is to verify technical correctness, the purpose of verification is to validate the outcome against goals set in task before implementation. It is usually most effective for the person who entered the task in first place to validate its correctness.

For any task the verifyer should assess whether its resolution is reasonable and acceptable for given task; for an implemented task, the appropriateness of the implementation should also be judged.

Verifying tasks usually consists of:

checking the resolution of task
- if it's duplicate, read the task which current task duplicates and decide if they really are duplicates and if there are no additional aspects in the current task
- if it's any other negative resolution, check if you agree with it
- if it's positive («implemeted»), you should check if the version or any newer one is in test environment and then check the application to see if the implementation satisfies your needs
decision
- if you find the resolution and implementation satisfying, you should mark the verification as passed
- if you don't, explain thoroughly what should be improved and mark the verification as failed

Here is why you need to bind tasks with changes – you can always figure out which task is implemented in which version. The simplest way to look it up is to have a glance at version differences.

In case the resolution itself is satisfying, but you found other problems during verification, you should mark the verification passed and add new tasks about the issues you found.

Statistics: Changelogic's current statistics show that about 20% of total verifications fail, meaning there are a lot of inconsistencies in task implementations even after the code review. Thus if you want to optimize your process even further and find bugs earlier, you might want to consider doing task verification during change review.