Models and diagrams are persisted in XMI format:
http://github.com/AresEkb/anchor/blob/master/anchor.tests/model/Performance.anchor
http://github.com/AresEkb/anchor/blob/master/anchor.tests/model/Performance.aird
It's easy enough to describe another serialization formats (textual or XML-based) for models and diagrams. Also here is a model repository for EMF-based models http://projects.eclipse.org/projects/modeling.emf.cdo
There are some projects for model comparision, merging, transformation, ... All of these tools are intended for use with models (not data based on models).
So one can persist, compare, merge, transform EMF-based Anchor models. But EMF doesn't allow (from scratch) to persist data based on these models.
If I'll need to store data, then I'll make something like this:
1) Create a concrete Anchor model by means of my editor.
2) Transform the model into Ecore-model (Anchors and Ties to EClass, Knots to EEnum, Attributes to either EAttribute or EClass). It will allow me to use EMF Tools to persist data based on the specific Anchor model.
But I don't think that EMF is good enough to store large amount of data. Mostly EMF is intended for models.
Alternatively I can transform the Anchor model into ER-model or into SQL-script (which creates tables in some DBMS). I think they are better for data storage.
In other words EMF isn't intended for data persistence. But it allows to generate (from model) Java-classes, Hibernate configuration files, SQL scripts or something else for a storage tool.
