Testing of Existing Software Products
This type of outsourcing peculiars to software development companies with a large base of installed products. Typically, in such companies there's a flagship product, the development of the next version of which employs most of its staff.
Outsourcing of testing of new releases of old versions of software on the circuit shown in the figure, saves a lot of strength, allowing the core development team to focus on creating of the next version of software. However, this can be achieved only with long-term cooperation with the test agency, constantly attracted to the software testing. Otherwise, the cost of learning the next release by software testers will exceed the gain from the use of outsourcing.
When you use the test agencies help you should grant it a maximum of historical information about the product. To track changes made to the product it will require the old reports, lists of errors identified in previous versions, a variety of documents on installing patches and documentation for the changes, including the file readme.txt. Such information is critical important to tracking, fixing old software bugs and search for the new one.
It is advisable to pass to a test agency lists of errors identified in the latest software version (in this case a flow will be added to the diagram in the figure).
The interaction between the developer and testing agency for outsourcing of testing of new releases of old versions of software
It is obvious that many of these errors may be found in previous versions.
Typically error detection in old versions results the output of new releases of software. In mature companies this process is planning.
It is advisable to coordinate plans of the output of release with the test agency. This will help to choose the right policies to ensure the quality and to carry out certain types of tests.
And finally, what makes sense to pay attention: this approach to testing, that is, there are two teams of testers, conducting approximately the same set of tests, and there is probability of competition - who will find more errors.
Since the test agency conducts tests following the main group of testers, it may appear a want to use test results as an indirect assessment of the work of the testing department of the customer. Such a situation should be avoided because it complicates the interaction. It is advisable to devise a mechanism that would allow competition to be transformed into cooperation between the two teams of testers.