In Part 1 of this post I covered the NoSQL movement, and described what I think is not ideal there. A short recap would be the following points:
- Lack of standardization – each database provides its own API and query language, that is often tightly coupled with its implementation.
- “No fancy queries” – users need to write their own de-normalization routines outside the database, which is prone to errors.
- Data migration – de-normalized data more sensitive to software updates than normalized data. As a result, more throw-away code needs to be written just to perform data migrations.
In this post I will present what I call NoDatalog, and explain how it can, at least in theory, overcome these three issues.Read More »