Incremental approach

State of mind

Learning cycles should be short

State of mind


This practice comes from the fourth principle of SAFe [SAFe 2021-4]. It is the direct impact of the third agile principle described in the manifesto:

                                                           "Deliver working software frequently,
                                                           in cycles of a few weeks to a few months,
                                                           with a preference for shorter cycles".

The incremental approach is to get feedback on what is delivered. The later the feedback, the wider the gap between the idea of the product and what is actually expected, the tunnel effect [Wikipedia 2021]. The further you go without feedback, the more you become disconnected from reality...

Application to test maturity

Learning is a strong objective of testing and the introduction of short learning cycles is the main difference between Waterfall and agile [Rainsberger 2014], so agile practices tend to adopt TDD to introduce this feedback as early as possible. However, although TDD allows rapid feedback on the quality of the code, it does not prevent the developer from entering a kind of cockpit where although his instruments give him always positive feedback, he comes to disconnect from reality, it is still the tunnel effect that is present.

So TDD must be linked to higher level tests, the need.

In agility, these tests at the level of the need are defined in the acceptance criteria of each User Story (US) and applying the TDD approach to the US is found in the "ATDD" practice which is naturally interwoven with TDD [Ambler 2006]. Thus, the ATDD makes it possible to obtain US that conform to the expectations of a US.

However, is it the right product?

Indeed, US are only fractional assumptions during Sprint Refinement on market expectations [Le Lan 2018] [Lawrence 2012]. This is why it is necessary to get feedback from users as early as possible, this approach is called "Lean Startup" [Ries 2012].

Illustration of the Lean Stratup from [Moustier 2020].

The theoretical basis on the intertwining of these interlocking learning loops is called "Double Loop Learning" [Argyris 1977] which teaches that a learning must be guided by a higher order direction [Moustier 2020].

Agilitest's position on this practice

Agilitest is positioned at the level of US acceptance testing and end-to-end testing with different technologies (HTML, fat clients on Windows or iOS, mobile, REST API). Its simplicity of use makes it a lightweight ally for bridging the gap between the hypothesis and the realization of the increment within the sprint.

Related cards

To go further

© Christophe Moustier - 2021