Try and maintain a graph model in memory.
Each time validation passes, update the model.
Use the URI (object, step, parameters) as a key for the vertex.
First suggest a next object, then narrow down the keywords, then narrow down the parameters.
This is different from the non-model approach of showing all objects from any app and all keywords from that object.
Try and maintain a graph model in memory. Each time validation passes, update the model. Use the URI (object, step, parameters) as a key for the vertex. First suggest a next object, then narrow down the keywords, then narrow down the parameters. This is different from the non-model approach of showing all objects from any app and all keywords from that object.