From WinDev to React: What a Migration Tells You About Your Product

A software migration is uncomfortable, and it is also clarifying. When the framework goes, you find out what your product actually is.

I lived this moving self-service kiosks from WinDev to React. WinDev had carried those products a long way. It could not carry them where the market was going.

Why we had to leave

WinDev tied us to Windows at a time when hardware was changing. Modern payment and till integrations were designed for the web. And hiring for a niche technology could not keep up with an international ambition. Staying was a decision too: it meant staying in one market.

What had to survive

The new stack was the easy part of the conversation. The hard part was deciding what could not be lost: the payment rules, the integrations with existing tills, fiscal compliance, accessibility, the way an ordering journey could be configured, and the behaviour when the network dropped.

None of that had anything to do with WinDev. It was accumulated knowledge of how a point of sale really works, and that was the real asset.

How we moved

We did not jump straight there. A first version in a classic V-model, then an interface redesign still on WinDev, and only then the React version, delivered in two-week sprints with pilot clients in the loop. Each step reduced the risk of the next one.

The objection

The argument against any rewrite is well known, and it is a good one: you stop shipping value while you rebuild what you already have. It is only worth doing when the old platform is what stands between you and the market you want. Ours was.

Know what is yours and what you rent. Frameworks go out of date. Knowing how the business works does not.