Directed Testing & DV Notebook SF: Minimize redundant tests to achieve testing goals with a fraction of the cpu-time and licenses. High quality directed tests achieve their intended functional coverage every time they are run. This allows specific features and design changes to be targeted by a list of well-chosen tests.

The DV Notebook SF allows engineers to specify tests in simple test description languages. Coverage annotations can be added on to these descriptions. The software then expands the test population such that it achieves maximal coverage in a reproducible way. The result is system where a single test engineer can create huge test populations of high quality directed tests.

Random Testing: If too many aspects of a test are randomized, re-creating a failure can be difficult. Random-constraint languages often yield uneven probabilities of different behaviors, and teams may spend 100's of CPU hours per day simulating the same behavior over and over.

Directed Testing & DV Notebook SF: Allowing test scripts to determine their own pass/fail criteria saves 100's of engineering hours per week. Directed tests are more predictable, and are therefore a better choice for automatic test scoring. In many cases, the specification of the test is enough to predict the desired result behavior without investing the engineering resources required to create a detailed reference model.

Random Testing: Random testing add project cost in the area of analyzing test results. More engineering effort is needed to look at the results of tests to determine if the behavior is reasonable. Detailed reference models are expensive to create, and often yield false-failure indications early in a project. Over time, test benches are often modified to avoid false-failures, and may ultimately yield false-pass indications, wasting valuable simulation resources while letting bugs slip through.

DV Notebook SF: Automatic generation of directed tests saves the costs that arise with classic random testing and directed testing methods.

The software automatically generates 1000's of directed test cases to satisfy the coverage bins specified by each coverage annotation. These tests are organized in the DV Notebook database, and can be sorted and searched by any criteria. Each test is written once and never modified. Any desired changes to behavior are achieved by creating a different test.

Random Testing: Small test libraries can achieve large-scale coverage by randomizing test behavior. However, a small library of tests may be very sensitive to modifications of the testing scripts. When nightly regression silently change behavior, weeks of testing are wasted.

Directed Testing: With limited engineering resources, hand-coded directed tests often fall short of coverage goals. Engineers create many individual tests and must organize them, usually in some type of file structure. Each test is isolated from the others, so changes to one test do not effect another test.

Random Testing & DV Notebook SF: Automatic generation of different test behaviors saves months from a schedule. Both random testing and DV Notebook SF create sufficient tests to keep unique simulations running 24-hours a day for the life of the project.

Using the DV Notebook SF, testing teams can perform automatic translations from test plan requirements to actual tests. This reduces the need for highly subjective processes such as test plan reviews and software code reviews. It is reasonable to expect the DV Notebook SF to generate 1,000 to 10,000 generated tests derived from each hand-written test. This drastically accelerates the process of driving towards functional coverage completion.

By using the DV Notebook, tests, results, documentation, and analysis are organized and cross-referenced in an easy search-and-click format.

Directed Testing: No modern system can be tested using by-hand coding of directed tests. Though every project can benefit from a detailed test plan and many directed tests, some automation and randomization will always be necessary.

Directed Testing & DV Notebook SF: Directed tests are created to achieve a particular goal quickly and reliably. Randomization is therefore kept to a minimum. This limits the impact of seeds, and insures that every run of that test will achieve the same functional goal.

Random Testing: Random testing reduces effort by randomly choosing values at run-time. However, this makes the test highly dependant on its initial seed value. The same test with different seeds will behave differently. Even if the seed is the same, the slightest modifications to the test program can cause different outcomes.