I enjoyed your articles, and I look forward to the rest of them, and to trying out the framework you discuss.
I'm neither a DBA, nor a Developer (yet); I work in QA, so please excuse any ignorance on my part. This method of development is fantastic, from a QA perspective, but would it not be a performance nightmare for many applications?
The first article said "The objective here is to produce small to medium database-based applications". The application I work with daily is certainly no Amazon.com. It is a client-server app with around 1200 tables, but most of them are small. A few of them may reach a couple of million records, though. Does that put it outside the scope of the methodology proposed in the article?
Performance, especially of long-running processes, is a constant issue for our largest customers, and we continually have to talk even large, well-funded organizations into upgrading their servers.
I would much appreciate some discussion of performance implications in the future articles of this series.