During the definition of the data model and the implementation of the initial prototypes all the information for the objects in ScienceSoft has to be inserted manually by the users. However, this is not practical and may even deter users from registering new collaborations or software. Once the model has been defined and agreed, possibly together any necessary encoding methods, more efficient and automated ways of collecting, discovering and associating information have to be defined. Information can be imported or lnked from external sources or can be used to create new obejctes and associations (for example the contributors of a software product can be discovered from the commits in the source code management system, high-level product dependencies can be inferred from package level dependencies in package repositories, etc.). The relationships can be created automatically or suggested to the users when they login and browse the catalogues. The goal is to be able to ask the minimum amount of information and discover the rest from the base data.

What type of data can be discovered? What information can be pre-registered on behalf of the users and which information requires explicit action or consent? How to import or export data?

Ideas?