Monthly Archives: March 2010

Testing in Agile – seperate or integrated?

I’ve been thinking about how to organise testing within an iteration. The challenge is that activities within an iteration are particularly related to a specific story. However, there are aspects of testing that are larger than a single story (e.g. integration testing). There are testing activities that are related to more than a single iteration (e.g. system and regression testing). I’ve seen models mooted where there is a parallel test iteration. This concerns me as it seems to split out testing (again) and has the potential to remove the emphasis on testing from the development aspects of the product. The risk is that the testers are once again excluded from the intra-iteration activities and lose visibility of what is occurring and what knowledge is being transferred. Another model I have seen is a specific test iteration as part of the overall release plan for the product. But is this not just “waterfall” by another name? What approach do we take that allows us to truly integrate testing into the agile iteration?

My thoughts?

1. include the testers in all core team activities from story elicitation through to unit testing

2. ensure that test automation is considered up front and time is allocated to automating the testing

3. ensure that Integration, System and Acceptance tests are integrated as tasks within the iteration

However, does this mean that we slow down development or do less “new” work during the iteration.

What do you think?

Posted by Sharon Robson