Core vs. Localization

There are two main parts of DS Knowledge Model. First is Core, which contains DS questionnaire data across all specific domains. On the other hand, there is the Localization part, where each localization extends or overrides the core data (or even other localization data). By picking and combining desired localizations, you create final set of questionnaire data, which can be used to create DS plans.

Namespaces

For referencing core and specific localizations we use namespaces. Each chapter has defined namespace as attribute, it should also coincide with the name of a folder in directory structure. All core chapters have namespace "core", localizations must have a different (unique) namespace.

If a localization wants to change a question, it must fill namespace attribute by this question identified by id. Then the question is in the "localization mode" and attributes like addExperts or addReferences can be used, make a change from required to optional, etc.

If a localization wants to add some question, it will contain just normal question as in the Core. It can moreover use insertBefore attribute to place the question before some other question (either from core or localization). When two questions have the same value for this attribute ID (and then namespace) will decide the priority.

Dependencies

Dependencies are introduced by using question changes, preconditions, cross references and question pointers (in insertBefore attribute). It's clear that if you choose a localization, all dependencies should be automatically chosen, too.

A graph of dependencies should be soon available as a "side effect" of the validator tool.