Category: Software Testing

In late December we heard two sides of the argument about Quality Assurance(QA) and Artificial Intelligence(AI).

Side one from a Financial Executive: ‘AI will do away with the need for testing, testers, and QA’.

Side two from a Quality Assurance Guru: ‘AI will mean more work than ever for QA.

We tend to think it will be somewhere in the middle.

1. AI will automate some tasks that are now done manually. We are already seeing this with some of the tools. This process has been ongoing for a long time.
2. AI will change the way programming is done. As a recent article stated: Some of the programs were very simple; the data, however, was critical and every change in data lead to very different results.
3. Following along with the point from above; AI wil allow us to do things in software that we could not have done before. This is already occurring.
4. However, we will still have to test this to make sure we are using the right data and drawing the correct conclusions.

QA is going to need to know how to understand, find, and load unbiased data.

QA will need to understand and test how the conclusions are drawn, rather than what the actual results are at the time of test execution.

This will impact everything we do from setup, to data provisioning, to execution and reporting. Test planning might escape the maelstorm but we would not bet on it.

Happy 2020’s

Give us a call to continue the discussion.

Photo by Franki Chamaki on Unsplash

We received an interesting article last week from a friend about Technology Trends.

The 10 Best Tech Products of the 2010s were listed. It was surprising how old some of them were. Some technology has been around a lot longer than was remembered!

Of course, the 10 best had to be followed up with the worst failures. Only 5 of those, but they still resonate for the amount of money sunk into them and the scale of the failure in some cases.

The next section was a prediction of the Five technologies for the next decade. A decade seems like a long time for some of these. We expect them to be mainstream within a year or two in some cases. Some are used already.

They are:

• Artificial Intelligence (AI)
• Blockchain
• Augmented Reality and Virtual Reality (AR/VR)
• Big Data
• Internet of Things (IoT)

So what is the impact on Software Testing? Some of these will be a huge challenge for those who are used to dealing with static inputs and static results. Most of these technologies will not work on that principle. The Test Environment will have to change also. We have already had test environments change in the last few years to accomodate many external feeds and applications. Now the test environment will have to expand further in order to accomodate inputs we don’t use now and also the volumes of data needed for the testing.

So what is the impact on Quality Assurance (Process Improvement)? There may be an even bigger impact on how we work. In the past, we could retrieve some mistakes by adding items late in the process. That will no longer be possible with some of the technologies. Failure to consider all items at the beginning of a project could have much larger costs later on. We cannot afford that type of omission in 2020 and beyond.

Photo by Alex Kotliarskyi on Unsplash

Quality Assurance is, as usual, up and down. Some organisations are buying in to Process Improvement, looking long term, and many are realizing that the solutions are not necessarily more testing or another test tool. They understand that the problems lie in how things are done rather than the tools used.

However, many organizations are still adding test tools or adding test phases and layers without adding value to the Software or the Product.

We completed several assessments over the past year and provided seminars aimed at process improvement. We are now up to the stage of implementing the process improvements identified. These organizations are moving ahead and planning for the long term. They can see the value of improvement.

However, on the other hand, some of the people we are dealing with are struggling to complete the necessary testing without considering the long term improvement. They are challenged by the necessity of getting testing done for clients and do not think they have the time for process improvement.

So will this change in the 2020? It seems unlikely in view of the above unless we consider the following:

1. Calculate the real benefit of implementation of process improvement over multiple projects. This is the long term, high level, view of how process improvement benefits multiple projects.
2. Define and gather the appropriate measurements and metrics to determine whether the process improvement is having the desired effect and also to identify further process improvement opportunities.
3. Select and acquire the appropriate technology to support the process improvement effort. Talk to us about a well tested methodology for QA technology acquisition.

Photo by Bekir Dönmez on Unsplash

Judging from what we have heard, AI seems to be the favourite flavour for Automated Testing in 2019. Many of the tools are stating that they are AI enabled or AI enhanced and claiming massive productivity gains as a result.

We are still seeing tools that require a lot of technical knowledge and other ones that have hidden the actual scripting from the users.

One other major change is that most tools are now automatically offering either a cloud based or an on-premise solution to suit every client’s wishes and, more importantly, their security needs.

Tools go through a cycle every few years. For a while we get a different test tool for every possible situation, then someone comes and consolidates a large number of tools into one tool that addresses most situations. Then the cycle starts again. Obviously this cycle is driven by the technology used to build applications and what level of testing is needed.

In addition to the technology that is used to build any particular product, the methodology also impacts the way automation is applied. Any iterative methodology has vastly different needs from a standard Waterfall methodology.

What we are not seeing, is much initial analysis to select the best test tool for an organisation. We are still seeing selection based on features and not necessarily on functionality that is applicable throughout the organisation. We are seeing some ROI for individual automation efforts but that is after the tool is selected and implemented. However, as discussed last week, that calculation is not over the entire organisation or over an extended time period. This means we are losing out on some automation opportunities and completing some based on a false calculation.

So will this change in the 2020? It seems unlikely in view of the above unless we consider the following:

1. Calculate the real benefit of implementation of automation over multiple projects and years.
2. Find an automation tool that suits your situation. There are many good ones around; you just need to find the appropriate one. Talk to us about a well tested methodology for test tool acquisition.

Photo by Franck V. on Unsplash