High performance navigator for parsing inputs of a message

In a message navigator that exposes inputs of a message as branches of a hierarchical data structure, a method of increasing navigation performance by enabling a load-on-demand feature that creates complete branches for only select inputs in order to reduce runtime setup costs for applications that do not need detailed branches for each input within the message, the method implemented by a computer comprising a processor and a memory, the method comprising:

• receiving a message that includes a plurality of inputs that get exposed to an application as a hierarchical data structure, which represents each of the plurality of inputs as nodes therein;
• as part of an initial processing of the message when creating the hierarchical data structure, using a navigator to parse sub-portions of details about each of the plurality of inputs, wherein the navigator extracts portions of input details for each of the plurality of inputs in determining if such details can be delayed for later, dynamic processing by the application;
• based on details for the plurality of inputs in the message, generating the hierarchical data structure which comprises,
• a first of the parsed sub-portions represented as a node with complete input details about a first input of the plurality of inputs,
• and a second of the parsed sub-portions represented as a stub node that includes partial input information sufficient to represent a second input of the message such that the stub node appears to the application as if it had all elements for the second input already as part of the hierarchical data structure, but wherein the stub node does not include complete input details for the second input in the message such that full processing of the second input is delayed until a later request by the application in order to save setup time in generating the hierarchical data structure;
• after the initial processing of the message receiving a request from the application, which requires further input details associated with the stub node representing the complete input details for the second input;
• and
• based on the request dynamically generating and loading the complete input details of the second input using the partial input information within the stub node such that the complete input details of at least one input are exposed to the application for further processing.