All right, I had this post in my mind for a while now .  Although it is about test planning , but I’m starting it fairly unplanned so might do it in parts or may be not  {{ After writing most of  the post , I have decided that I will make it “why and how we plan test efforts,that just reflects on what sort of planner i am  }}

Trying to capture thoughts about test planning from my experiences and observations over the last 6 years in the software engineering quagmire.

1. Why do we even need to plan testing ?

To be honest, if have not been able to figure out a precise answer yet . Nowadays,given how insanely complex software engineering projects are,it is getting harder and harder to exercise control and plan on how projects pan out.

Here are some reasons I can think of ::

• To try to prevent looking foolish during or at the end of the release.
• To provide a vision of what is the mission of current testing effort ?
• It is a mandatory testing by product
• Managers feel in more control if they have “a plan”
• Need to project some numbers about test efforts, resources,costs etc.
• “How could the wheels come off?” Risk mitigation and contingency
• What will be tested when ?
• What will be tested how ?
• “You do not have a test plan !??

2.  How do you plan testing ?

Here are a few approaches that I have observed so far ::

• Light weight but alert . Do not plan too much or write bluky plans . Have a precise and correct idea about the mission,time,resources,risks and test strategy. Document it to a basic degree AND make sure at least the test team understands and agrees to it ! Leave lots of room to change and improvisation on the fly as the project moves.Focus is more on execution and quick feedback. We know that things are going to change anyway,so why plan too much at the beginning ?

• We would not start testing until we have a test plan that is so beautiful that we can jerk off by looking at it. My first job was in a company where they used to happen in every project. Bulky plans were documented and signed off and then only the product was touched by testers. Came out useful only in highly controlled and less erratic projects and you have loads of time, of course .

• We do not need a plan or we need one but we would not make one . We will just get together each day/week/iteration and have a look at what happened in the last day/week/iteration and plan for until the next day/week/iteration. This sort of approach makes most control freaks uncomfortable who can not see the end at the start . If a test team has acceptance of this sort of approach , I have seen this working as a charm for agile-ish projects.

What do you think  Dr Strange Love ?

More later …..