|
|
| Line 1: |
Line 1: |
| − | ** Automated Testing is not quality is not testing **
| + | [Automated Testing is not quality is not testing] |
| − | | |
| − | | |
| − | Automated testing != quality != testing
| |
| − | | |
| − | How to write better tests
| |
| − | | |
| − | Why are the tests passing but if fails on deployment?
| |
| − | | |
| − | Can't test quality into something
| |
| − | | |
| − | testing is not quality
| |
| − | | |
| − | testing is human role
| |
| − | automated testing is for validation
| |
| − | | |
| − | Auto testing
| |
| − | | |
| − | | |
| − | cant detect black text on black background
| |
| − | | |
| − | | |
| − | sapient process of testing
| |
| − | Some experiments on increasing fuzzy logic for testing.
| |
| − | | |
| − | Test automation is part of testing.
| |
| − | | |
| − | Does agile think automated testing is a panacea, bigger than it is in it's role in wider testing.
| |
| − | | |
| − | If writing good tests then get better results, if just filled in, then not.
| |
| − | | |
| − | How get developers to create better tests
| |
| − | | |
| − | Not poss to automate everything... trade off of coverage.
| |
| − | | |
| − | Correlation of manual and automated testing.
| |
| − | | |
| − | Experince of ?? - need to automate 7-9 times to get value.
| |
| − | | |
| − | Frequently devs write poor tests
| |
| − | | |
| − | | |
| − | | |
| − | 1. teach them to write
| |
| − | 2. testers write tests
| |
| − | | |
| − | If poor requirements then auto testing not much value.
| |
| − | Don't agree - tests for stuff that works is useful even not knowing
| |
| − | | |
| − | TDD is not a testing practice, it's a software design practice.
| |
| − | | |
| − | Automate the mundane crap, allows the testers to focus on the new stuff, the edge cases, the human intuitive stuff.
| |
| − | | |
| − | | |
| − | Test automation not replace need for people to be testing, just good for regression testing.
| |
| − | | |
| − | Should regression tests be the same always, or different each time.
| |
| − | | |
| − | Each time there's a defect, write test to test that defect, then fix it. Can't get it right first time, but add tests and improve as you go on.
| |
| − | | |
| − | Infrstructure guy: "Developers don't write good test".
| |
| − | | |
| − | Devs write tests as part of software design.
| |
| − | | |
| − | Can automate almost anythings, if test script, then can automate.
| |
| − | | |
| − | Manual tests just get bigger, so must focus on integration tests. Take long time to write high level tests, but worth it.
| |
| − | | |
| − | What about the client perspective? Does client want automated testing? Is it assumed? Or is it POC and NOT wanted at all? Ask.
| |
| − | | |
| − | Expect vendors to have good development practices.
| |
| − | | |
| − | We do code quicly, cheaply, reliably... pick any two.
| |
| − | | |
| − | Find the trade-off of scripted, ad-hoc, and automated testing. Client expectations.
| |
| − | | |
| − | Business / testers can writer at least test scripts for devs to automate.
| |
| − | Some testers getting better and better at writing automated tests.
| |
| − | | |
| − | Automated testing is easy to pick up... devs can train testers. Diff tools getting better. Eg. Selenium record, click, done = test.
| |
| − | | |
| − | Most NZ testers from non-it background, from business, not tertiary.
| |
| − | Changing with IT grads, not scared of getting into code.
| |
| − | | |
| − | Not all testers need to be technical.
| |
| − | | |
| − | Still see many test teams where no-one is technical.
| |
| − | | |
| − | Test automation good for smoke testing - giving it a once over - to ensure foundation is there for manual testing.
| |
| − | | |
| − | Exploratory testing is human function.
| |
| − | | |
| − | Define the balance up front of automated and manual scripted and exploratory testing.
| |
| − | | |
| − | Tester and BA's work together early to ensure requirements clear, and testable. Devs can write automated unit and functional tests.
| |
| − | | |
| − | Diff phasess of testing aims to find different classes of defects / issues. Role for devs, testers.
| |
| − | | |
| − | Automated testing should test everything in the requirements, as they are anticipated.
| |
| − | Exploratory human testing good for detecting things NOT in the requirements.
| |
| − | | |
| − | BAs, testers, devs must be talking with each other through the proceess early on and continously... all have same goals... quality.
| |
| − | Colocation is best way to get quality - BA, testers, devs.
| |
| − | | |
| − | Are testers gate-keepers? Do they just expose problems and give information, or can they call the shots of go / no-go. Up to each organisation to decide what works well for them.
| |
| − | | |
| − | Colocation or separation / 3rd party? Depends on if client trusts testers will find bugs.
| |
| − | Not just about colation, but communication. Communiation easier with colocation, but not necessary, and sometimes the opposite occurs if not got culture in place.
| |
| − | | |
| − | Sometimes good to have internal and external testers... internal testing deliver to spec and get a level of quaity. External to be independent.
| |
| − | | |
| − | Zero defects is usually a destructive goal. Beware of technical debt. Zero defects is only that not found any more (not that there aren't any).
| |
| − | | |
| − | Devs who write tests usually are more productive... better thought out code.
| |
| − | | |
| − | Good to have walk-thru for use cases once requirements written.
| |
| − | | |
| − | Not really about test automation, it's about people and communication.
| |
| − | Agile issues: need vendor who knows their stuff, and customer who knows what they want... the latter is hard to find. Sometimes waterfall better (including for infrastructure).
| |
