What is an Artificial Neural Network (ANN)?

- The artificial neural network or ANN (sometimes also called as just neural network) is a mathematical model that has got its inspiration from the biological neural networks. 

- This network is supposed to consist of several artificial neurons that are interconnected. 

- This model works with a connectionist approach for computing and thus processes information based up on this only. 

- In a number of cases, the neural network can act as an adaptive system that has the ability of making changes in its structure while it is in some learning phase. 

- These networks are particularly used in searching patterns in data and for modeling the complex relationships that exist between the outputs and inputs. 

- An analogy to artificial neural network is the neuron network of the human brain. 

- In an ANN, the artificial nodes are termed as the neurons or sometimes as neurodes or units or the ‘processing elements’. 

- They are interconnected in such a way that they resemble a biological neural network. 

- Till now, no formal definition has been given for the artificial neural networks. - These processing elements or the neurons show a complex global behavior. 

- The connections between the neurons and their parameters is what that determines this behavior.

- There are certain algorithms that are designed for altering the strength of these connections in order to produce the desired flow of the signal. 

- The ANN operates up on these algorithms. 

- As in biological neural networks, in ANN also functions are performed in parallel and collectively by the processing units.

- Here, there is no delineation of the tasks that might be assigned to different units. 

- These neural networks are employed in various fields such as:

1. Statistics
2. Cognitive psychology
3. Artificial intelligence

- There are other neural network models that emulate biological CNS and are part of the following:

1. Computational neuroscience
2. Theoretical neuroscience

- The modern software implementation of the ANNs prefers a more practical approach than biologically inspired approach. 

- This practical approach is based up on the signal processing and statistics. The former approach has been largely abandoned. 

- Many times parts of these neural networks serve as components for the other larger systems that are a combination of non – adaptive and adaptive elements.

- Even though a more practical approach for solving the real world problems is the latter one, the former has more to do with the connectionist models of the traditional artificial intelligence. 

- Well the common thing between them is the principle of distributed, non – linear, local and parallel processing and adaptation. 

- A paradigm shift was marked by the use of neural networks during the late eighties. 

- This shift was from the high level artificial intelligence (expert systems) to low level machine learning (dynamical system). 

- These models are very simple and define functions such as:

f: X à Y

- Three types of parameters are used for defining an artificial neural network:

a)   The interconnection pattern between neuron layers

b)   The learning process

c)   The activation function

- The second parameter updates the weights of the connections and the third one converts the weighted input in to output. 

- Learning is the thing that has attracted many towards it. 

- There are 3 major learning paradigms that are offered by ANN:

1. Supervised learning
2. Un – supervised learning
3. Reinforcement learning

- Training a network requires selecting from a set of models that would best minimize the cost.

- A number of algorithms are available for training purpose where gradient descent is employed by most of the algorithms.

- Other methods available are simulated annealing, evolutionary methods and so on.

Explain empirical vs defined & prescriptive process?

To make any process successful it has to be controlled in a way that it achieves its predefined goals or we can say it should be guided along the path of success. This fact holds good for any type of process in this world and so does for the processes in the field of software engineering.

In the field of software engineering, two approaches have been identified for keeping a control over the development and other related processes namely:


1. The empirical process control method and
2. The defined process control method.


In this article the above mentioned two approaches have been compared so that you get a better understanding of both the programs. So let us see how these two approaches control the processes.

The Empirical Process Control Method

- The empirical process control model was defined to exercise or control the process via following some frequent adaptations as well as frequent inspections.

- These inspections and the adaptations both are meant for the generation of the unpredictable and unrepeatable outcomes or results and can be thought of as being imperfectly designed.

- Since for the past many years the various software development methodologies have been known to be controlled by the latter approach i.e., the defined process control method.

- But we all know that every time a certain same output or outcome cannot be expected from the software development processes.

- Therefore, most of the agile software development methodologies are controlled by the empirical process control model and the most famous example being that of the “scrum” agile software development methodology.

- The term empirical process control model itself justifies as the term empirical means the information acquired by the means of experimentation and observation.

- Here the information is achieved by means of inspection and adaptations which serve as a means for the observation and experimentation.

- The process of the empirical control is constituted of a continuous cycle of adaptation of the process as per the requirements and inspection of the process for correct working.

- The empirical control process model has three pillars as we can make out from the definition of the process control model without which it cannot be called as the empirical process control method:
1. Transparency
2. Inspection and
3. Adaptation.

- The first pillar i.e., transparency indicates that the outcomes of the affects of the empirical process control model and the aspects affecting the outcomes should be visible to the programmers and developers who are responsible for controlling the whole process.

- The second pillar i.e., the inspection indicates that all the aspects of the control process should be monitored quite frequently to enable the fast and early detection of the unacceptable variances.

- The third pillar i.e., the adaptation indicates the adjustment of one or more aspects as required of the control process if the software system or application being processed is observed to lie outside the acceptable limits implying that the result will also be unacceptable.

- The defined process control model approach is adopted when the underlying mechanisms of the software system or application are well understood by the programmers and the developers.

The Defined Process Control Model

- The defined process control model can be thought of as a theoretical approach.

- When a well defined set of inputs is given, it is obvious that the same outcomes will be generated every time the program executes.

- With the well understood technologies and stable requirements, one can very well predict a whole software project.

- Even nowadays the empirical process control model holds as the essence of the agile software development processes.

- Empirical process holds good for the complex development processes which encounter difficulty in the production of repetitive outcomes.

Automated Test Data Generation Using an Iterative Relaxation Method ?

Software development is not complete without software testing. It constitutes an inseparable part of software development process. Almost 50 percent of the total funds for development of a software are spend on software testing and this process consumes most of the time of the whole development process.

AUTOMATED TEST DATA GENERATION
Generation of test data for the given test and according to the testing criteria proves to be quite a difficult problem.
- The automated test data generation can relieve much stress of the software testers.
- Generation of test data that is used to make a program follow a given path is the most prominent and an important problem that arises while carrying out the path oriented testing.
- The obtained input is refined by various iterations and another input is obtained.
- The whole process predicates on this obtained input.
- The statements of the program relevant to the evaluation of each branch are executed in each iteration.
- Up on each iteration, a set of linear constraints is obtained which are then solved to obtain the increment values for the input data.
- The obtained increment values are added to the current input value to obtain the input data that is to be used for the next iteration.

RELAXATION METHODS
- The relaxation method provides knowledge about the amount by which the value of each input variable should be modified for the branches on the path in order to evaluate the desired result.

ITERATIVE RELAXATION METHODS
- These can be defined as techniques for solving system of equations.
- Relaxation methods are also iterative methods defined for numerical problems of mathematics.
- They are extensively used for solving system of equations which include the following types:
(A) Linear equations
Relaxation methods are used for solving the linear equations. Problems like that of linear least squares are usually addressed under this category.

(B) System of linear inequalities
Iterative or relaxation methods prove very effective in solving the system of linear inequalities which represent the problems similar to those that arise during linear programming.

(C) Non linear system of equations
These days, iterative methods or relaxation methods have been developed and prove a great help in solving system of equations which are non linear.

- Iterative relaxation methods have proven to be very effective and important methodology in providing solutions for linear system of equations.

- They prove to be effective system of equations that are used to solve partial differential equations based on the model of ellipse.

- These systems of equations are basically used to describe problems related to boundary conditions and values in which the value of the function in the solution is indicated or specified on the boundary of a specified domain.

- If the branch conditions on a path are linear in nature, the iterative relaxation technique either obtains a solution for such a path in just one iteration or it declares that the path is in-feasible.

- We can say that the existing approaches require an unacceptably large number of iterations for longer paths since they use only one branch predicate as well as input variable at a time.

- These methods also use back tracking.

- If the branch conditions on a path are non linear in nature, then it takes more than one iteration to get the desired input data.

- But, the set of constraints that has to be solved is linear in nature and can be solved using the method of gauss elimination.

These advantages make the technique of automated test data generation practical as well as suitable for automated testing.

Some details about Strings and Arrays of Strings in C

Multiple character constants can be dealt with in 2 ways in C. if enclosed in single quotes, these are treated as character constants and if enclosed in double quotes, these are treated as string literals. A string literal is a sequence of characters surrounded by double quotes. Each string literal is automatically added with a terminating character ‘\0’. Thus, the string “abc” will actually be represented as follows:

“ abc \0” in the memory and its size is not 3 but 4 characters ( inclusive of terminator character ).

Arrays refer to a named list of a finite number n of similar data structure elements. Each of the data elements can be referenced respectively by a set of consecutive numbers, usually 0, 1, 2, 3, ……., n. if the name of an array of 10 elements is ARR, then its elements will be referred as shown below:

ARR [ 0 ], ARR [ 1 ], ARR [ 2 ], ARR [3], …… ARR [9]

Arrays can be one dimensional, two dimensional or multi dimensional. The functions gets() and puts () are string functions. The gets() function accepts a string of characters entered at the keyboard and places them in the string variable mentioned with it. for example :

Char name[ 21 ];

The above code declares a string namely name which can store 20 valid characters ( width 21 specifies one extra character ‘\0’ with which a string is always terminated ). The function gets() reads a string of maximum 20 characters and stores it in a memory address pointed to by name. As soon as the carriage return is pressed, a null terminator ‘\0’ is automatically placed at the end of the string. The function puts () writes a string on the screen and advances the cursor to the newline. Any subsequent output will appear on the next line of the current output by puts ().

Arrays are a way to group a number of items into a larger unit. Arrays can have data items of simple types like int or float, or even of user defined types like structures and objects. An array can be of strings also. Strings are multi dimensional arrays comprised of elements, each of which is itself an array.

A string is nothing but an array of characters only. In actual C does not have a string data type rather it implements string as single dimension character arrays. Character arrays are terminated by a null character ‘\0’. So, for this reason the character arrays or strings are declared one character larger than the largest string they can hold.

Individual strings of the string array can be accessed easily using the index. The end of a string is determined by checking for null character. The size of the first index ( rows ) determines the number of strings and the size of the second index ( columns ) determines maximum length of each string. By just specifying the first index, an individual string can be accessed. You can declare and handle an array of strings just like a two dimensional array. See an example below:

Char name [10] [20] ;

Here the first dimension declares how many strings will be there in the array and the second dimension declares what will be the maximum length of a string. Unlike C++, C has some different functions for adding or concatenating strings, checking string length and to see the similarity of two strings. The functions are namely strlen, strcmp, strcat, strrev etc and are included in header file string.h. Strings are used for holding long inputs.

Estimation Techniques - For Web Engineering Projects

Early estimation is required during planning even though there is a huge chance that these are proven wrong. Some uncertainty is unacceptable and some solid techniques and concrete procedures help reduce the inaccuracy of estimates.

Estimating techniques for web engineering projects adapt the agile process model. Along with the agile estimation approach, a modified function point (FP) measure can be used to develop an estimate for a web application. Roetzheim suggests for web application estimation, the information domain values while adapting function points are:

- Inputs include maintenance screen, input screen or form or each tab using a tab notebook.
- Outputs include static web page, dynamic web page script, each report whether web based or administrative in nature.
- Tables include logical tables in database, each XML object if XML is used to store data.
- Interfaces are logical files into our out of the system boundaries.
- Queries are a message oriented interface and they are externally published.

According to Mendes:
The best way to determine the volume of the web application is by :
- predictor variables that are associated with application like page count, function count, media count,
- web page characteristics like page complexity, linking and graphic complexity,
- media characteristics like media duration,
- functional characteristics like code length.

Empirical estimation models for:
- total project effort
- page authoring effort
- media authoring effort
- scripting effort
can be evaluated by using these measures.

Explain Black box and White box testing? What are their advantages and disadvantages?

For complete testing of a software product both black and white box testing are necessary.

Black-box testing
This testing looks at the available inputs for an application and the expected outputs that should result from each input. It does not have any relation with the inner workings of the application, the process undertaken or any other internal aspect of the application. Search engine is a very good example of a black box system. We enter the text that we want to search, by pressing “search” we get the results. Here we are not aware of the actual process that has been implemented to get the results. We simply provide the input and get the results.

White-box testing
This testing looks into the complex inner working of the application; it tests the processes undertaken and other internal aspects of the application. While black box testing is mainly concerned with the inputs and outputs of the application, white box testing help us to see beyond i.e. inside the application. White-box testing requires a degree of sophistication which is not the case with the black-box testing, as the tester is required to interact with the objects that are used to develop an application rather than having easy access to the user interface. In-circuit testing is a good example of a white-box system testing where the tester is looking at the interconnections between different components of the application and verifying the proper functioning of each internal connection. We can also consider the example of an auto-mechanic who takes care of the inner workings of a vehicle to ensure that all the components are working correctly to ensure the proper functioning of the vehicle.

The basic difference between black-box and white-box testing is the areas of focus which they choose. We can simply say that black-box testing is focused on results. Where if an action is performed and the desired result is obtained then the process that has actually been used is irrelevant. White-box testing, on the other hand focuses on the internal working of an application and it is considered to be complete only when all the components are tested for proper functioning.

Advantages of Black-box testing
- Since tester does not have to focus on the inner working of an application, creating test cases is easier.
- Test case development is faster as tester need not to spend time on identifying the inner processes; his only focus is on the various paths that a user may take through GUI.
- It is simple to use as it focuses only on valid and invalid inputs and ensures that correct outputs are obtained.

Drawbacks of Black-box testing
Constantly changing GUI makes script maintenance difficult as the input may also be changing. Interacting with GUI may result in making the test script fragile and it may not properly execute consistently.

Advantages of White-box testing
- Since the focus is on the inner working the tester can identify objects pro grammatically. This can be useful when the GUI is frequently changing.
- It can improve stability and re usability of test cases provided the object of an application remains the same.
- By testing each path completely it is possible for a tester to achieve thoroughness.

Drawbacks of White-box testing
Developing test cases for white-box testing involves high degree of complexity therefore it requires highly skilled people to develop the test cases. Although to a great extent fragility is overcome in white-box testing BUT change in the objects name may lead to breaking of the test script.

Model-View-Controller (MVC) Design Pattern

Model View Controller design pattern is used to support multiple types of users with multiple types of interfaces. The Model-View-Controller (MVC) pattern separates the modeling of the domain, the presentation, and the actions based on user input into three separate classes:

- Model: The model locates all data information for the application. It does not care about how you interpret it or how process it. It divides functionality among objects involved in maintaining and presenting data to minimize the degree of coupling between the objects.

- View: The view manages the display of information. It is responsible for maintaining the consistency in its presentation when the underlying model changes.

- Controller: The controller interprets the mouse and keyboard inputs from the user, informing the model and/or the view to change as appropriate. The actions performed on the model can be activating device, business process or changing the state of a model.

The MVC pattern allows any number of controllers to modify the same model. The strategies by which MVC can be implemented are as follows:
- For Web-based clients such as browsers, use Java Server Pages (JSP) to render the view, Servlet as the controller, and Enterprise JavaBeans (EJB) components as the model.
- For Centralized controller, a main servlet is used to make control more manageable.

What is a System? What are general principles of system?

System is an:
- integrated set of inter-operable elements;
- it consists of group of entities or components;
- interacting together to form specific inter-relationships;
- organized by means of structure;
- working together to achieve a common goal.

In defining the system boundaries,a software engineer discovers the following:
- entities or group of entities that are related and organized in some way within the system, either they provide input, do activities or receive output;
- activities or actions that must be performed by the entities or group of entities in order to achieve the purpose of the system;
- a list of inputs;
- a list of outputs.
Entities that are involved in this system are the applicant, club staff and coach.

General Principles of Systems
- The more specialized a system, the less it is able to adapt to different circumstances.
- The larger the system is, the more resources must be devoted to its everyday maintenance
- Systems are always part of larger systems, and they can always bepartitioned into smaller systems

There are two types of systems, namely, man-made systems and automated systems. Man made systems will always have areas for correctness and improvements. These areas for correctness and improvements can be addressed by automated systems. Automated systems consists of computer hardware, computer software, people, procedures, data and information and the connectivity that allows the connection of one computer systemwith another computer system.

Define Unit Test Case, Integration Test Case, System Test case

UNIT TEST CASES(UTC)
The unit test cases are very specific to a particular unit. The basic functionality of the unit is to be understood based on the requirements and the design documents. Generally, design document will provide a lot of information about the functionality of a unit. The design document has to be referred before a unit test case is written because it provides the actual functionality of how the system must behave, for given inputs.

INTEGRATION TEST CASES
Before designing the integration test cases the testers should go through the integration test plan. It will give complete idea of how to write integration test cases. The main aim of integration test cases is that it tests the multiple modules together. By executing these test cases the user can find out the errors in the interfaces between the modules.
The tester has to execute unit and integration test cases after coding.

SYSTEM TEST CASES
the system test cases meant to test the system as per the requirements; end to end. This is basically to make sure that the application works as per the software requirement specification. In system test cases, the testers are supposed to act as an end user. so, system test cases normally do concentrate on the functionality of the system, inputs are fed through the system and each and every check is performed using the system itself. Normally, the verifications are done by checking the database tables directly or running programs manually are not encouraged in the system test.
The system test must focus on functional groups, rather than identifying the program units. When it comes to system testing, it is assumed that the interfaces
between the modules are working fine.
Ideally the test cases are nothing but a union of the functionalities tested in the unit testing and the integration testing. Instead of testing the system inputs and outputs through database or external programs, everything is tested through the system itself. In system testing, the tester will mimic as an end user and hence checks the application through its output.
Sometimes, some of the integration and unit test cases are repeated in system testing also especially when the units are tested with test stubs before and not actually tested with other real modules, during system testing those cases will be performed again with real modules.

What are Test Case Documents and what is the general format of test cases?

The test cases will have a generic format as below:
- Test Case ID : The test case id must be unique across the application.
- Test case description : The test case description should be very brief.
- Test Prerequisite : The test pre-requisite clearly describes what should be present in the system, before the test executes.
- Test Inputs : The test input is nothing but the test data that is prepared to be fed to the system.
- Test Steps : The test steps are the step-by-step instructions on how to carry out the test.
- Expected Results : The expected results are the ones that say what the system must give as output or how the system must react based on the test steps.
- Actual results : The actual results are the ones that say outputs of the action for the given inputs or how the system reacts for the given inputs.
- Pass/Fail : If the expected and actual results are same then test id Pass otherwise Fail.

The test cases are classified into positive and negative test cases.Positive test cases are designed to prove that the system accepts the valid inputs and then process them correctly. Suitable techniques to design the positive test cases are specification derived tests. The negative test cases are designed to prove that the system rejects invalid inputs and does not process them. Suitable techniques to design the negative test cases are error guessing, boundary value analysis, internal boundary value testing and state transition testing. The test cases details must be very clearly specified, so that a new person can go through the test cases step by step and is able to execute it.
In an online shopping application, at the user interface level, the client request the web server to display the product details by giving email id and username. The web server processes the request and will give the response. For this application, we design the unit, integration and system test cases.

Step 4 To test API : Call Sequencing, Step 5 To Test API : Observe the output

Step 4: Call Sequencing
When combinations of possible arguments to each individual call are unmanageable, the number of possible call sequences is infinite. Parameter selection and combination issues further complicate the problem call-sequencing problem. Faults caused by improper call sequences tend to give rise to some of the most dangerous problems in software. Most security vulnerabilities are caused by the execution of some such seemingly improbable sequences.

Step 5: Observe the output
The outcome of an execution of an API depends upon the behavior of that API, the test condition and the environment. The outcome of an API can be at different ways i.e. some could generally return certain data or status but for some of the APIs. It might not return or shall be just waiting for a period of time, triggering another event, modifying certain resource and so on.

The tester should be aware of the output that needs to be expected for the API under test. The outputs returned for various input values like valid/invalid, boundary values etc needs to be observed and analyzed to validate if they are as per the functionality. All the error codes returned and exceptions returned for all the input combinations should be evaluated.

What are some of the formal approaches used for exploratory testing? Continued...

Some of the formal approaches used for exploratory testing are:

- Identify the break points
Break points are the situations where the system starts behaving abnormally. It does not give the output it is supposed to give. So, by identifying such situations also, testing can be done. Use boundary values or invariance for finding the break points of the application. In most of the cases, it is observed that system would work for normal inputs and outputs. Try to give input that might be the ideal situation or the worse situation. By trying to identify the extreme conditions or the breakpoints would help the tester to uncover the hidden bugs. Such cases might not be covered in the normal scripted testing. hence, this helps in finding the bugs which might not be covered in normal testing.

- Check the UI against Windows Interface etc standards
The exploratory testing can be performed by identifying the user interface standards. There are set standards laid down for the user interfaces that need to be developed. These user standards are nothing but the look and feel aspects of the interfaces, the user interacts with. The user should be comfortable with any of the screens that he or she is working on. These aspects help the end user to accept the system faster. By identifying the user standards, define an approach to test because the application developed should be user friendly for the user's usage.

- Identify expected results
The tester should know what he is testing for and expected output for the given input. Until and unless, the aim of the testing is not known, there is no use of the testing that is done because the tester may not succeed in distinguishing the real error and normal work-flow. The tester needs to analyze what is the expected output for the scenario he is testing.

- Identify the interfaces with other interfaces/external applications
In the age of component development and maximum re-usability, developers try to pick up the already developed components and integrate them. In some cases, it would help the tester explore the areas where the components are coupled. The output of one component should be correctly sent to other component. Hence, such scenarios or work-flows need to be identified and explored more. There may be external interfaces, like the application is integrated with another application for the data. In such cases, focus should be more on the interface between the two applications.

Software Testing - Validation Phase - Alpha Testing

A software prototype stage when the software is first available for run. The software has the core functionalities in it but complete functionality is not aimed at. It would be able to accept inputs and give outputs. Usually, the most used functionalities are developed more. This test is conducted at the developer's site only. In a software development life cycle, depending on the functionalities, the number of alpha phases required are laid down in the project plan itself.

During this, the testing is not a through one since only the prototype of the software is available. The basic installation and un-installation tests and the completed core functionalities are tested.

The aim of alpha testing is :
- to identify any serious errors.
- to judge if the intended functionalities are implemented.
- to provide to the customer the feel of the software.

A thorough understanding of the product is done now. During this phase, the test plan and test cases for the beta phase which is the next stage is created. The errors reported are documented internally for the testers and developers reference. No issues are usually reported and recorded in any of the defect management or bug trackers.

The role of the test lead is to understand the system requirements completely and to initiate the preparation of test plan for the beta phase. The role of the tester is to provide input while there is still time to make significant changes as the design evolves and to report errors to developers.

Validation phase - Unit Testing - how to write Unit test cases

Preparing a Unit test case document commonly referred as UTC is an important task in unit testing activity. Having a complete UTC with every possible test case leads to complete unit testing and thus gives an assurance of defect free unit at the end of unit testing stage.

While preparing unit test cases the following aspects should be kept in mind-

Expected functionality
Write test cases against each functionality that is expected to be provided from the unit being developed. It is important that user requirements should be traceable to functional specifications which should be traceable to program specifications which should be traceable to unit test cases. Maintaining such traceability ensures that the application fulfills user requirements.

Input Values
- Write test cases for each of the inputs accepted by the unit. Every input has certain validation rule associated with it. Write test cases to validate this rule.
- There can be cross-field validations in which one field is enabled depending upon input of another field. Test cases for these should not be missed.
- Write test cases for the minimum and maximum values of input.
- Variables that hold data have their value limits. In case of computed fields, it is very important to write test cases to arrive at an upper limit value of the variables.
- Write test cases to check the arithmetic expressions with all possible combinations of values.

Output Values
- Write test cases to generate scenarios which will produce all types of output values that are expected from unit.

Screen Layout
Screen/report layout must be tested against the requirements. It should ensure that pages and screens are consistent.

Path Coverage
A unit may have conditional processing which results in various paths, the control can traverse through. Test cases must be written for each of these paths.

Assumptions and Transactions
A unit may assume certain things for it to function. Test cases must be written to check that the unit reports error if such assumptions are not met.
In case of database applications, test cases should be written to ensure that transactions are properly designed and in no way inconsistent data gets saved in the database.

Abnormal terminations and Error messages
Test cases should be written to test the behavior of unit in case of abnormal termination.
Error messages should be short, precise and self explanatory. They should be properly phrased and free of grammatical mistakes.

Methods of black box testing - Boundary Value Analysis

Black Box Testing is testing without knowledge of the internal workings of the item being tested. There are four black box testing methods. Graph based testing method and Equivalence Partitioning have already been discussed.

Boundary Value Analysis
Boundary Value Analysis determines the effectiveness of test cases for a given scenario. Boundary Value Analysis (BVA) is a test Functional Testing technique where the extreme boundary values are chosen. Boundary values include maximum, minimum, just inside/outside boundaries, typical values, and error values. Boundary value analysis complements and can be used in conjunction with equivalence partitioning.
For boundary value analysis, the following guidelines should be used:
- For input ranges bounded by a and b, test cases should include values a and b and just above and just below a and b respectively.
- If an input condition specifies a number of values, test cases should be developed to exercise the minimum and maximum numbers and values just above and below these limits.

Advantages of Boundary Value Analysis
- Robustness Testing – Boundary Value Analysis plus values that go beyond the limits.
- Min – 1, Min, Min +1, Nom, Max -1, Max, Max +1.
- Forces attention to exception handling.

Limitations of Boundary Value Analysis
Boundary value testing is efficient only for variables of fixed values i.e boundary.