Introduction Designing database is an art process similar to building a house. Database designers always face the problems of designing a relational database that will be free of database anomalies. These anomalies bring repetition of tuples that delay processing time and occupy memory spaces. Suppose that the value of the attribute BUILDER determines values of the attribute MODEL and PRICE, (BUILDER [right arrow] MODEL, PRICE) and that the value for the attribute MODEL determines the value for PRICE, (MODEL [right arrow] PRICE). Grouping these attributes in relation HOUSE(BUILDER, MODEL, PRICE) has several undesirable properties. First the relationship between MODEL and PRICE is repeated in the relation for each BUILDER who builds a particular MODEL of home. This repetition creates difficulties if a BUILDER who happens to be the last BUILDER of a certain MODEL home is deleted from the relation, then the relationship between the MODEL and its PRICE also disappears from the relation. This is called a deletion anomaly. Similarly, if a new builder who happens to be the first BUILDER of a certain MODEL home is added then the relationship between MODEL of a home and its PRICE will also be added. This is called an insertion anomaly. Suppose that the relationship between a MODEL and its PRICE is changed e.g. the price is increased; then the MODEL and PRICE relationship should be affected for every BUILDER of the MODEL. This is called update anomaly. These anomalies are undesirable since the user is not likely to realize the consequence of the insertion, deletion or updating. The user may inadvertently affect a relationship that was not intended to be modified. Consistency, insertion, deletion and updating are not probe effecting all groupings of attributes. If the relation HOUSE(BUILDER, MODEL, PRICE) is normalized then the consistency and anomaly problems disappear. Normalization is a step by step reversible process of replacing a given collection of relations by successive collection in which the relations have a progressively simpler and more regular structure (Date & Darwen, 2000). The reversibility guarantees that the original collection of relations can be recovered and therefore no information has been lost. Codd proposed three normal forms which he called first normal form (1NF), second normal form (2NF) and third normal form (3NF). A stronger definition of 3NF was proposed by Boyce and Codd and is known as Boyce-Codd Normal Form (BCNF). All these normal forms except 1NF are based on the functional dependencies among the attributes of a relation (Elmasri & Navathe, 1994). First normal form relates to the structure of the relation. It requires that every attribute of a relation be based on a simple domain. The database designers have no problem to know if a relation violates first normal form. They can put the relation into first normal form algorithmically by replacing a non-simple domain by its constituent simple domains. In the second (2NF), third (3NF) and Boyce Codd normal form (BCNF), there is a need for the database designers to know the real meaning and application of database keys such as candidate key, primary key, super key, etc,. Problem Statement Database designers always find it difficult to determine these keys from relational database schemas. It has been difficult to motivate students and database designers to derive primary, candidate, alternative and super keys because they think this area is dry and theoretical. There are many algorithms to determine the database keys but they look abstract for students. Many database researchers indicated that relational database model to derive database keys tends to be complex for the average designers. Failure to determine the database keys at times leads to poor design that can generate database anomalies. The database key algorithms often require extensive relational algebraic backgrounds that database designers lack. …