-
Notifications
You must be signed in to change notification settings - Fork 1
Guiding principles
Based on the The GenTech Genealogical Data Model The GenTech Genealogical Data Model
- Source Based: All data is required to have a a source cited. Data will not be allowed into the system without a source.
- Source Trust Ranking: Since a source is required for all data, the reliablity of the source needs to be known. For example, a birth certificate will be considered more trust worthy than hearsay. This will be used in the inference engine.
- Exact Data: Data needs to be entered exactly as it appears in the source document. Once the data is entered it can then be standardized.
- Complete Data: All the data contained in the source document should be entered into the system.
An attribute is a characteristic of a some entity in the system. An attribute is not bounded by time and space. The attribute has the following fields: id, value, type (I wonder how they are stored?) (unit?)
An event is something that does have place and time related attributes. As of now, only groups have event attributes: place and date (stored as attributes as mentioned above)
The Conclusions section of the data model deviates from the GenTech Data Model to a greater extent. The GenTech model is based on an Assertion entity which creates a binary tree of all the data. Instead, data that comes directly from a Source is accepted without an intermediate entity like an Assertion. All other data manipulations are done via the Inference entities with the Source based data forming the foundation of all the inferences. The data will still all be in n-tuple trees.
The standardization algorithms an create either Assertions (part of the Research Data Model ) or Inferences (part of the Conclusion Data Model).
Instead of having different type entities for each entity, there will be a single Type System and all other types entities will use it.
- A type is simply a UUID so that types can be created by different users and not clash.
- Topoged.com provides a type clearing house for users to share new types
- Types have more than one representation based on the type, the context and the locale. I18N is handled through types rather than resource bundles (maybe)
- Types can be related to each other. For example, a group type FAMILY can have the list of all the types that are valid members of the group: CHILD, MOTHER, FATHER, etc.