We also learned that we’re better at designing and building applications than carving out a commercially viable niche for them. But from the very start we realised this would first and foremost be a learning exercise, and here we weren’t disappointed. We learned about npm-hell and the mental contortions Node would demand from people with decades of experience writing apps in Java and ABAP. We learned that CouchDB’s elegant multi-version concurrency control model was really well suited to Meistr’s “Q&A quiz” domain, and underpinned our horizontally-scalable, globally distributed architecture. But it also turned something as simple as a multi-dimensional SQL join into a significant mental exercise, and thus failed the ultimate test of any technology: it did not help the scarce resources (i.e. us and our time) to become more productive and effective. But in the end Meistr showed us how we could build a functional app using static HTML and JS served from a CDN that entirely relied on API servers spread across three continents for all state, logic and persistence.
We would like to thank everyone who contributed content to the public Q&A database, used the app in quiz mode to learn more about SAP technologies, used it in interview mode to ensure they hired the right talent for their teams, or otherwise contributed ideas and feedback.
The Blue T team