« Summer Events for Agile Coaches | Main | Honing your Skills as an Agile Coach »

07 July 2011

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

allan kelly

I was discussing this very topic with a client yesterday. Also present was Benjamin Mitchell who told us how one of his teams used to mark the card with a yellow dot in the planning session to indicate tests needed to be written. The tester would then write the tests during the iteration and mark the card with another dot (I forget which colour) to indicate they were ready.

Mike Pearce

Great post. I'd never really thought there might be a difference between tests and criteria, but, when pointed out, it actually seems pretty obvious.

Where does BDD fit into this? Is that the Acceptance Criteria, but in the form of tests? Or is it not relevant at all in this context?

Rachel Davies

Thanks, Mike. I hadn't thought so much about the difference until recently. BDD is all about helping people express acceptance criteria and acceptance tests in a way that everyone can understand.

Rachel Davies

Thanks for sharing that, Allan. I've worked with several teams that use sticky dots to indicate the state of a story card. Although nowadays, with Kanban becoming more popular, it seems to be common to move cards into a column on the team board to indicate their state.

sachin kundu

In my team, product owner takes the responsibility of specifying what the acceptance criteria would be before we even take the story for implementation in the sprint. This is an effort to keep ourselves from taking highly ready stories only. The PO comes from a technical background so he is capable of writing the criteria in gherkin but we don't require it to be. Plain english is considered also ready.

Team then demostrates the criteria using functional automated test tools or if thats not possible physical demonstration.

Rachel Davies

Thanks for sharing what your team does, Sachin. I think a lot of development teams would like to work with a PO who has time to do this before Sprint Planning. The only slight downside of having the PO do this alone is that there's a chance conversations with the team don't happen. Often the value of fleshing stories out with tests can help identify ways to build a simpler solution. Perhaps you don't have this issue with such a technical PO.

Chris Chan

The way you described acceptance test seems to be similar to that of a test case. Is that right?

Rachel Davies

Chris, well sort of. Acceptance tests are test cases. However, not all test cases are acceptance tests. Acceptance tests are the subset of test cases that relate to checking that the agreed acceptance criteria have been met.

Cesario Ramos

Nice post.
I usually have a req workshop before sprint planning, especially
with distributed teams, where the Scrum team turns the acceptance
criteria into a couple of test cases or examples. Works great to create a common understanding
across a distributed team.

Scrumcrazy.wordpress.com

Rachel,

While distinguishing between "acceptance criteria" and "acceptance tests" is logical and probably fine at a local team level, the most credible resources on User Stories do not make that distinction.

You mentioned Mike Cohn's book, but he also refers to the "acceptance criteria" you speak of as "acceptance tests". Further, in Ron Jeffries(the co-inventor of user stories) 3 c's article:
http://xprogramming.com/articles/expcardconversationconfirmation/

He refers to "defining" the acceptance tests vs. "implementing" the acceptance tests.

I sometimes refer to "creating" the Story Tests vs. "executing" or "automating the story tests."

Further, I would worry that "criteria" would take the focus off of a test strategy and might lead us back to "the system shall" style prose requirements.

Again, though, I think it is perfectly fine for any team to set up whatever terminology they like. I personally prefer to use industry standard terminology (when it exists formally or by consensus) when I can because I think it helps the global Agile community.

In this case, IMO, Ron and Mike's terminology of "Acceptance Tests" is the consenus standard.

Stephanie

Spot on! Acceptance criteria and acceptance tests are independent entities. The terminology can confuse learners easily. It's easy to forget that stories with acceptance criteria are not contemporary substitutes for funk specs! Thanks for the article

William F. Nazzaro

Great post. This is the very thing that I'm seeing many clients struggle with. Because they use them as synonyms vs. different entities it's driving a poor behavior.

This poor behavior is one that causes them to over-specify the story prior to it being brought into the sprint/iteration.

The comments to this entry are closed.