Normalization helps remove redundancies in Data Design. The redundancies are removed systematically using a three or four step approach. The four important normal forms are 1NF, 2NF, 3NF and BCNF. This tutorial will cover the first three normal forms.

The three normal forms apply to only Dr. Codd's relational model otherwise referred to as the relational data model.

In the 1st normal form (1NF) multi-valued columns are disallowed. If there are multivalued columns such as multiple subjects registered by students in a course registration system or multiple telephone numbers used to call patients in a patients record system they have to be split into multiple records one for each value in the multi valued column. In this form of normalization, every record should be able to be identified uniquely and so would entail the use of the Primary key.

In the second normal form (2NF) in addition to the database being in the 1st normal form, all nonkey attributes should be functionally dependent only on the primary key. In the case of a composite key, the non-key attributes should be fully dependent on the entire composite key. For example, consider the following record system for customers. The following tuple (Customer ID, Order ID, Order Date, Store Location) has the Primary key as Customer ID and Order ID. The two nonkey attributes are Order Date and Store Location. Here the location of the store does not depend upon the primary key or in other words, this nonkey attribute is not functionally dependent upon the primary key. So this table has to be split into two tables in order to make it in 2NF.

Customer -Order (Customer ID, Order ID, Location ID) Location (Location ID, Location Name).

In the third normal form, the database should be in the second normal form and in addition the normal form ensures that redundancy is further reduced by assuring that every nonkey attribute does not have any dependency with any other nonkey attribute or does not have any transitive functional dependency. In popular relational database terminology, this translates to that data model having referential integrity constraints enabled in the design.

As an example of a database not in 3NF is the tuple with the attributes (Customer ID, Name, Address line 1, Address line 2, Address line 3, Pincode). This is not in 3NF as the Address line elements depend only on the Pincode)To make this in 3NF we use two tuples Customer and Pincode with the following attributes Customer (Customer ID, Name, Pincode), Pincode (Pincode, Address line 1, Address line 3). Here the Pincode attribute in the customer tuple is the foreign key in the table which matches the Primary key in the Pincode tuple. This translates to what is termed as Referential Integrity constraint where the foreign key in a table corresponds to the Primary key in another table.

Software, application and web testing are investigations usually conducted to provide stakeholders with the information they need on the quality of services and products they are working on or developing. The tests are also a great means of providing independent, objective views of the application so the business involved can understand and appreciate risks of the implementation of the same.

Most tests on applications have the intention of finding defects and errors to verify whether software is fit for use. The tests can also be about application functionality, usability, compatibility, performance, and security. They are all important aspects of ensuring that only the best applications are offered in the market without the risk of frustrations and disappointments when already in use. Selecting the best testing tool can, however, be a challenge for many, but with a few guidelines, it is very possible to select application tools that are reliable.

Product integration - Software testing tools need to provide integrated modules that support component reusability. Components designed for functional tests should able be able to accommodate other types of testing including load and regression tests. The tools should offer an environment based on easily understanding and common language for all users.

Web testing - Testing tools for application need to have the ability to also support tests within web browser scope. Intranet and internet testing is of importance for applications based on them and should remain portable across the different browsers. The testing tools should be able to automatically adjust to varying load times as well as different performance levels.

Planning and management - When looking for the best application testing tools, it is advisable to stick to those that can manage testing processes and provide testing components that are organized to create meaningful end-user reports. Users should be able to include non-automated procedures within the automated plans and rest results. Robust testing tools allow users to integrate existing results into automated test plans, making it easy to link business requirements to the test results. This way, the evaluation of the application readiness is made easy and is based on the ability of the application to support the specific requirements of the business.

Performance - Testing tools that are reliable make it possible for the users to perform meaningful tests on load and performance so they can accurately measure the performance of the system. The best testing tools should also offer the results of such test in a reporting format that is easy to understand by all kinds of users.

Services - When getting software testing tool, it is important to ensure that the vendor can offer you the support you need during use of the same. There are situations that may require expertise when using the tool and the vendor will be most helpful if they can offer services such as implementation services, extensive consulting, assessment, and training services to the users. The best tools should also be based on structured testing methodologies. They should be engineered to be usable even by non-programmers and application end users.