What is STEMMA, in a nutshell?
A private research project for representing micro-history in a transportable fashion (i.e. non-binary, defined schema, locale-independent, and culture-neutral). The goal is more than that of a simple exchange mechanism, or even a representation for long-term preservation; it will serve to create "live" micro-history documents that can be exchanged, opened, and immediately visualised or navigated.
What is micro-history?
Micro-history is a pursuit that looks at the events and interconnections in the fine-grained history of subjects that includes people, places, groups (e.g. regiments, organisations, families), and animals. This is certainly more than 'family trees' as it is a superset of family history, local history, and one-name or one-place studies.
Is STEMMA a data model or a data format?
STEMMA is primarily a data model, which means it is concerned with the relationships between different aspects of the data, including micro-history subjects such as persons, places, groups, and animals; hierarchies, such as lineage; events and timelines; geography; information, evidence, and analysis; text and audio transcription; and general marked-up narrative. There is an XML serialisation format defined on this site, but it is comparatively loose, and not the only possible representation -- the current XML was designed to be as simple to implement and modify as possible during R&D.
Is the informational sub-model mandatory?
STEMMA has notional sub-models termed conclusional and informational. The majority of the research work has gone into utilising the conclusional in conjunction with marked-up narrative. There are aspects of the informational model that have barely been touched so far, and which would need very specialised software. See Our Days of Future Passed — Part III.
Is STEMMA lineage-linked?
STEMMA is a multi-linked model. Family lineage is represented by hierarchies between people, but hierarchies are also used for place/group relationships, and animal lineage. The popular notion of family trees is just one potential visualisation of STEMMA data.
Is there a STEMMA Product?
There is no product available. The research project involves much prototyping of software algorithms and interfaces, and these are used to encode private data, but they do not form a coherent product of commercial quality.
Does STEMMA require a database?
STEMMA is designed to be loaded quickly into a memory and indexed there. Disk-based databases, such as relational ones, are not essential for private data, and cause many problems. See Do Genealogists Really Need a Database.
How is STEMMA related to FHISO?
It is not -- I just happened to be a member of FHISO too. On a more philosophical note:
a) STEMMA is a private research project associated with the representation of micro-history. It is not interested in compatibility with other models. FHISO is currently about standards for sharing existing genealogical data (generally lineage-linked) without loss between compliant products. Its community-driven standards will be freely accessible to the public.
b) STEMMA is more about modelling (getting the data organisation correct), and less about the serialisation (i.e. file format). It currently uses a loose XML-based serialisation, but this is deliberately insular to allow unhindered evolution of the model. FHISO is heavily focused on serialisation issues that can be applied to several existing formats. It is adopting very modern concepts that are already part of the Semantic Web design. It is not currently focused on modelling issues beyond the capabilities of existing software.
c) Both projects want an international and extensible method of sharing and long-term preservation. Neither mandate the use of a database .