When explaining the Domino data model to developers with RDBMS background I get a lot of comments along the line of "and that is called a database? There is no way to define a table structure!". Once you take a closer look, you will realize, that Domino HAS a strict data model. Data in Domino is stored as META data. Expressed RDBMS style it could look like this:


(This is a conceptual, not necessarily fully accurate model)
When migrating a Domino application to a RDBMS backend you will be confronted with interesting choices:

• Implement a META data model with attribute/value pairs?
• Treat multi value fields as child tables?
• Keep the DocumentUniqueId as primary key?

You might want to consult the literature about XML to RDBMS mapping to read more about the challenges when transiting between conceptual models (document centric to table centric). Be aware, the questions above are just about the data model, there is much more: document security, handling of RichText, handling of attachments, handling of embedded objects. You also might want to pull some Notes items into the document table (most likely: "form" and "$$" items) and/or add a document access table. Also you might want to pull the first value of an item into the item table to make the data access faster, Your extended model would look like this:



I'm curious about the upcoming Domino7. DB/2 will be an option to store Domino data, thus uncovering IBM's approach how they map Domino to a RDBMS.

No comments.