I prefer an architecture that enables me to clearly identify the 'entry points' of the (UI) application into the dbms. Ideally I would like three schemas: 1 - Application owner schema with all the tables and stored pl/sql that is not directly accessed by the application. 2 - API schema. This schema has all the objects that are accessed by the UI application. 3 - UI application connect schema. Schema 2 would have private synonyms to objects referred in schema 1. And also the necessary object privileges granted from schema 1 to schema 2. Schema 3 has private synonyms to objects in schema 2 + the necessary object privileges granted from schema 2. On Mon, Apr 7, 2014 at 10:11 PM, Jeff C <backseatdba@xxxxxxxxx> wrote: > I wanted to know what everyone's opinion is on separating schemas. Like > having a data only schema, a code only schema and then a schema the > application logs into and executes code from the app schema. Or do you put > all data and code in a single schema and then a separate schema for the > application to run as? > Also do you implement table API's? If you do then this probably nullifies > a lot of these options. > > Let me know what you do. > Thanks. > > > -- Toon Koppelaars RuleGen BV Toon.Koppelaars@xxxxxxxxxxx www.RuleGen.com TheHelsinkiDeclaration.blogspot.com (co)Author: "Applied Mathematics for Database Professionals" www.rulegen.com/am4dp-backcover-text