There is a
problem, a modification is suggested to overcome the problem. Stakeholders
become bound up in tide of enthusiasm regarding the suitability and elegance of
the modification, relieved the problem is solved. But is it? It is easy to lose
sight of ensuring that time is not wasted in development and testing by having
to re-develop and re-test because the functional and technical specification
were not subjected to proper scrutiny and diligence in the rush to solve the
problem. Does this situation sound familiar?
The synopsis
of the change is written outlining, functionality, benefits, costs and risks.
The change is approved in principle. A detailed functional and technical
specification is prepared. At this point are the specifications ready for
review, approval and development? Only if the specifications includes a test
plan.
Insisting on
a test plan before the modification is approved for development will inform and
challenge the functional and technical specifications ensuring they are
complete and robust. The stipulation that a test plan must be included becomes
moderating influence on the modification requestors forcing them to think
through the modification before submitting it for approval. It also immediately
raises the quality bar on the specification. As a starting point the user who
requested the modification should be invited to explain how they would go about
testing the functionality of the modification. This basic test premise should
be then developed into a battery of test cases covering unit, edge or envelope
and integration testing.
The
development of a comprehensive test plan for the modification is best done in a
team setting where business analysts, users and developers can build on and
challenge each other’s ideas. Time spent deriving the test plans prior to final
submission and approval will pay dividends in ensuring the veracity of both the
functional and technical specifications saving time later in development, unit
and user acceptance testing. Once the test cases have been developed like any
other project document they should be subject to peer review. This is best done
again in a project team setting, often suited to a walkthrough type meeting
supported by an algorithmic diagram, a few days after the test cases were
originally documented. In most cases a
second walkthrough review separated by a few days proves beneficial in ensuring
quality, the old adage measure twice cut once comes to mind.
Happy implementing
Sobering advice to avoid the post euphoric hangover and mortgaging of the project schedule.
ReplyDelete