Strike the hot iron of change.

The impact bad software can have on lives in incalculable ...

was the strapline I put on my article in September 2023 about the #PostOfficeScandal. The victims, both direct and indirect, are now having their voices heard. Action is finally, belatedly, some might suggest grudgingly, being taken. Quashed convictions and full financial recompense will, hopefully, soon be forthcoming, although nothing can compensate for the lost homes, good standing in communities, childhoods, relationships, livelihoods, businesses, lives!

We don't need QAs or testers ...

is the claim being made by many in the software industry over recent months and, in some cases, years. The #PostOfficeScandal has many contributing factors:

• the honesty and integrity of management at all the organisations involved,
• the obfuscation of evidence during prosecutions,
• the willingness of politicians to listen to the issues raised and hold people to account
• and many others.

However, fundamentally, the Horizon software was not fit for purpose on many levels. Without those defects, the prosecutions would not have been necessary, the lives would not have been ruined, the whole, sorry scandal could have been avoided. Testing of requirements and prevention of defects is possible as I described in my article about Throwing Bad Money After Bad.

Every system released has bugs in it ...

is another argument, which has some validity. As a former developer, I know that I've sent my share of bugs into the world. That is not a reason, though, to render mute the need to test. Neither is it a reason to take the risk that the product owner and developer, however skilled and experienced between them, will always get "it" right.

We have yet to see the defect reports from Horizon and to be able to understand the root causes of the resultant imbalances or how, following a process given to a subpostmistress by the "help desk" resulted in an erroneously negative value doubling, instead of going to zero, or any number of other failures in the system.

Testing just slows releases down

I've heard from more than one project manager or owner over the years, normally when the project is up against it, when gaps in understanding of what has been asked for have been revealed in business testing or, even worse, on release when things have had to be backed out, needing re-work at the last minute, missed deadlines, failed integrations and others.

I've also been part of an organisation that saved their business by bringing not just testers but members from across the organisation into a process intended to remove ambiguities and omissions from requirements/user stories/features prior to development. I've been part of embedding QA and Test resource into 3 Amigos squads which helped transform weekly releases to multiple releases a day. In both these cases we improved the time to market figures, reduced the cost of producing the finished product, enabled to the developers to spend more time developing new features, reduced the context switching required to fix defects, since they weren't there and raised morale across organisations.

Now more than ever ...

the software industry is going to be under scrutiny.

• What were the processes that could/should have stopped the #PostOfficeScandal from happening when the software was produced?
• Why did the system for Covid test and trace fail in its primary objective at huge cost?
• How is it that a lack of "What If Thinking" meant that a simple process to put a flawed flight plan into an alert queue, leaving the rest of the National Air Traffic System processing normally, was missed?

The paying public rightly expects that the software they pay for, rely upon, and use moment by moment daily will have been tested and shown to work before it is released into the world. We don't signoff buildings, drugs, vehicles or any number of other things, failure of which can have fatal consequences in the way we release software into the world, the failure of which can have fatal consequences.

The iron of change is hot

Now, perhaps more than ever, the software industry needs to look to the order in its own house. To recognise the need for all the 'traditional' software development disciplines, which includes QA and Test. To use those disciplines in the way that is most effective to deliver the greatest value to their customer base. To recognise that Quality can lead to business advantage, to be bringers of customers through confidence in the products we release, not simply a "cost centre", perceived as slowing down those releases and adding no value.

I've written my MP and contacted Gareth Davies, Comptroller of the National Audit Office, recommending that an agency be established to "opine on the appropriateness of methodologies for the?development of systems funded from the public purse". I believe that there will be public and political backing for such an agency in the light of the projects listed above.

Additionally I, along with many in the testing community, believe that we have need of a professional society, recognised as an institution akin to the organisations who are members of the Engineering Council, making a recognised "Profession" of our discipline. Many are adopting the "Test Engineer" or similar as a job title. If that is what we wish to be known as, let's do what Civil engineers, Mechanical Engineers and other professions do. Paul Gerrard is working towards creating the Test Engineering Society to "promote the importance of Test Engineering in the wider engineering community and provide a framework for engineers to achieve professional registration."

"The death of Testing" has been announced on numerous occasions. The tragedy of the Post Office Scandal has raised public and corporate awareness of the vital role for QA and testing should play. Let's seize that opportunity to show how a thriving and vital QA and Test profession can deliver huge benefits, industry and society wide.