The annual Infor M3 User Association meeting in The Netherlands is always interesting. A chance to catch up with M3 friends, to learn about what they are doing and where Infor are taking their ever-increasingly impressive technology. And it is impressive for sure, particularly in its connectivity powered by their Ion solution which enables M3 (and other Infor products) to join together into a complete business solution. Of course, where needed it enables the platform to tunnel to other technologies outside of the Infor stable too.
It struck me watching this expanding vista of connected bubbles that with this power must also come complexity, if not in how the technology is set up, but in the possibilities, permutations and options to configure and perhaps misalign. It puts a different emphasis on testing, and perhaps the word validation is better. It is not so much testing each ‘Lego’ brick, they have been well tested, it is validating how they have been configured to achieve the various business goals. This validation demands an end to end, many transactions, cross-module, cross-system approach. It takes a lot of hard work to ensure each tunnel goes to the right place and delivers the right data.
Having made the journey from the UK via the always impressive Eurotunnel which loads its ‘data’ into carefully controlled train packets and delivers it safely to the other end, regardless of the political potholes that have troubled it and will trouble it more in the future, it is, in essence, a much simpler system than that presented by an average M3 environment. But I suspect the “Chunnel” has also undergone a lot more testing as lives obviously depend on its safe operation.
A friend and client Richard Calder presented at the conference how they go about testing their M3 and supporting systems. How they have invested in automation with TestDrive to the point that they run a full suite of tests before promoting any change, around eighty thousand so far. Almost as if lives depended on it. And in a way they do, because jobs depend on it. Reynolds Catering is so tightly connected with their customers that they could not survive half a day without the system. If they failed their customers, they would lose them and that would be commercially fatal. But whilst the smooth operation is a critical need, the investment in software was modest and even more modest in manpower with this being achieved principally though one non-technical team member. This insurance takes them to a place where any change, patch, or upgrade can be implemented with safety, knowing all the right data will appear at the end of all the right tunnels in the complex systems that run the business.
You see, you can’t be too careful, can you? I wish I had that thought when I arose the next morning and managed to persuade myself that a run in the dark before the 400-mile drive home would be a good idea. I might have taken a head torch, or even put my glasses on to assist my less than perfect vision. I might then have spotted the missing foot-sized paving stone next to the curb, even if covered by autumnal discarded leaves. I might have avoided the twisted ankle and stumbling into the path of one of the many cyclists. But all in all, no real harm was done, I’ll be running again in a week and no cyclists were hurt. I got away with it, sometimes we get lucky.
But when it comes to making sure your systems don’t fall over, luck is not a strategy you should be relying on.
Do you ever feel “we are getting away with it”?
Then there is a better answer, look at TestDrive and adopt certainty rather than lucky.