Say Uncle, you can give me some best practices pleaaaaaaaaaaase?
The first good practices are:
- Good sense
- Organization,
- Reflections.
Nothing but magic, properly implemented, will guarantee the proper use of a benchmark test. I will expose them as a question to ask the good moments:
Management requirements or requirements:
- Uniqueness?
o This is the only item on this test?
- Clarity?
o Reflect on the tree to establish,
o The need is expressed clearly enough and is there no gray area or ambiguities on the understanding?
o What are the requirements implicit and not listed in the expression of needs (for example), but we must still consider?
- Accuracy?
o The requirement is it enough to be measurable and quantifiable targets via specific test?
- Validity?
o The document led to the requirement is updated and validated by the customer?
- Testability?
o Is it possible to create one or more relevant cases to effectively test all aspects of this requirement?
Managing the card catalog test or the Test Plan:
Fundamental:
- Each sheet test should be linked to one or more requirements or you can say goodbye to the test coverage
- Once the test sheet drafted a trick is to check if all the cards are connected at least once a requirement, otherwise you have holes in your racket!
- Depending on the strategy, the Test Plan should be prioritized according to the functional, organizational or business customer
- The tree test plan must:
o Be easy to understand by functional
o In order of business process / application client (use numerical codes to manage the directory eg
o be determined at the outset between all actors
o 4 to 5 levels MAXIMUM directories otherwise you watch the madness
- Provide care tabs “Details” and “Design steps:
o Enter a clear description of the common people for the purpose of the test card
- Provide care to the tab Design steps:
o Be sufficiently precise and detailed to avoid any ambiguity and interpretation at runtime
o Display data to be used and their origins
o = A STEP ONE ACTION
o Inform the expected outcome (RA) so precise and clear because the result will be compared to the result (RO) during execution
–
Good questions:
- Uniqueness?
o A target = sheet test
- Clarity?
o The stages of the procedure are clearly identified, no-til no gray area or ambiguities on the way forward?
- Data?
o All representative data needed for testing are they identified?
o Have data that are:
consistent
Clarifies
Representative
Prepared
- Coverage?
o The modus operandi of the sheet test is relevant and complete the objective identified?
- Generic?
o No test sheet valued unless a value is hard in the test
Design and execution of test campaigns or Test Lab
- Cash?
o Environment to conduct any tests available?
o The application to test it is installed with the correct version (versioning policy of the client)?
o The access (login / passwords) to applications and other servers valid?
- Preparation?
o The test cards are up to date or she date a year ago?
o Automated test scripts are identified and in the right directories?
- Clarity?
o The names of campaigns and test sheets are quite explicit for your next tier include?
- Data?
o At’on all the data needed to run the cards?
- Coverage?
o The entire campaign covers happens objective identified?
Good practices during the execution of the test sheets:
- Have environments, platforms and networks available
- Availability of the coffee machine effective
- During the execution of steps:
o Follow the procedure in the “description” on LETTER!
o The tests outside the marked paths are BANNED!
o Compare the result (RO) with the field Expected Result (or expected outcome (RA)).
o
Management of anomalies or Defects:
Fundamental:
- Have a workflow defined anomalies that is known, validated, shared and clear. This is for all players
- If yes, do all the players are aware of their roles and responsibilities?
- If the execution proceeds failure:
o Verify that the specified outcome is correct
o Enter the result as clearly as possible
o Check that the anomaly is reproducible integrator or ridicule you and not take into account the anomaly
o Open an anomaly
o Describe the anomaly as clear as possible otherwise your work will not help and you will lose a lot of time in explanation
o ALWAYS associated with a sheet test
- Reproducibility?
o The anomaly is it reproducible under the same conditions?
- Compliance?
o The abnormality has been detected by following the procedure laid down in the sheet test?
o expected results included in the file are updated?
- Uniqueness?
o The anomaly detected is not already open?
To do this, ask your colleagues and your manager is a good start
Use filters and search fields
o Is not there already opened a ticket on a subject almost identical?
Clarity?
- Gravity?
o What is the level of severity of the anomaly?
Critical
Major
Minor
A final practice for the road? Make a copy and paste the above practices in an Excel file, the file named “Checklist guaranteeing the quality of the test repository”. You do not have:
- Go to request and ask a bonus to your manager ;-)
- And submit your client to be finally a real example of industrialization of testing tools
And for the record, know that putting into practice every day these recommendations will make the implementation of this tool CMMI level 3 (see article on CMMI3 and QC).
Commentaires récents