Selenium as a trend
in test automation - traps
          to avoid
                Nataliya Holodiuk
Will talk about


1.   Tool choice
2.   One project experience
3.   Traps to be aware of
CONFUSED?
What?
How?
Return of investments?
Got some trendy tools?
Selenium
    + many browsers & OS

    + directly in the browser

    + JAVA, Ruby, Python, etc

    + comprehensible docs

    + communities
One project story
Given

ecommerce web application written in PHP



      tests in Java with Selenium IDE
                    +
            Jenkins(Hudson) CI
RC
Page Object Pattern
Trend victims
Fail ?

• 85% of time - test suite
  maintenance
• new feature breaks up to
  50% of tests
• spent time and efforts !=
  expected result
• QAs demoralized
• big boss unhappy
Fail!


good tool != great results



                 Classic mistake # 1
100% automation


complete coverage !=
  bug free product


             Classic mistake # 2
UI tests only




          Classic mistake # 3
Late automation

     Interdependent tests

Automation as a way to find bugs

                 Another classic mistakes
BDD
Given I visit “http://google.com”
When I type “DELL” in “SearchField”
And I press “Search”
Then I should see “www.dell.ua” in search results
Why better?

●   clear aims, plan that is easy to follow
●   acceptance tests – user scenarios automated
●   tests on PHP – developers involved
●   using app methods to create preconditions
Even more better?
●   new tests are written before feature committed
●   green build rule
●   more than 1000 effective stable maintainable
    tests in 5 months
●   less time spent on suite support
One size doesn’t fit all
Any questions?


natali.holodiuk


natali.holodiuk@gmail.com
Ever tried. Ever failed.
      No matter.
 Try Again. Fail again.
      Fail better.
                     Samuel Beckett

Selenium as a trend