Collaborative software lives or dies on a boring question. When one person moves something, how long until the other person sees it? On a shared concept map, that delay is the whole experience. Push it too high and two learners end up fighting over a map that no longer exists.
The kit-build platform faced a harder version of that problem than most. A concept map is not a document. Every learner's map gets compared against a teacher's map for automatic diagnosis, so sync errors do not merely look bad. They corrupt the assessment.
Socket.IO gives a server a persistent channel to every connected client, so it can push changes instead of waiting for a client to ask. For a shared map, that is the difference between a live session and a refresh button.
There is a broader lesson in the choice. The team did not rebuild the learning model to go online. They kept the part that produced the learning results and replaced only the transport. When a method already works, that is usually the cheaper and safer path.
The paper is careful about this. Its evaluation covers potential usability, not measured learning. A prototype that runs smoothly is not evidence that students learn more through it. The authors present it as a foundation for practical use, which is the right level of confidence for the evidence available.
Pinandito et al. (2021). Design and Development of Online Collaborative Learning Platform of Kit-Build Concept Map. Journal of Information Technology and Computer Science.
