Electronic Books

Developers tend to drop testing when they are too slow to execute or getting them to work is too
laborious.
The unitary testing is also in the execution context; each test must be independent from the others and
must not rely on a specific configuration. If the tests are interconnected, having one that fails may
compromise an entire chain of tests and render the bug individuation much harder. The context should
not influence the tests because it is critical that tests can be moved with confidence from one machine to
another and with the minimum of configuration possible.
There are areas where Unit Testing does not fit so easily, nominally UI (User Interfaces) and databases.
This is because in the first case the user interaction is required, which tends to remove the automated
advantage of the process. The databases are hard to test for miscellaneous facts: They must be
configured, structures of tables must exist, and data for the tests must be in some way generated and
deleted continually and be in the same exact state when the same test is repeated.


Pages:
318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342