SlideShare a Scribd company logo

Selenium Users Anonymous

•
•
Dave Haeffner
Dave Haeffner

This document summarizes a presentation on best practices for using Selenium for test automation. The presentation covers 12 steps: 1) Admit you have a problem; 2) Take a deep breath; 3) Try looking at things differently; 4) Pump some tech iron; 5) Find your inner Napoleon and develop a strategy; 6) Break down the wall between QA and development; 7) Learn the terrain; 8) Test less but test well; 9) Keep it lean and optimize; 10) Pay it forward by sharing knowledge; 11) Resources for learning Selenium; 12) Recap of the 12 steps. The document provides additional details on each step and recommendations for learning more.

Related slideshows

TestWorksConf: Experience exploratory testing
TestWorksConf: Experience exploratory testingTestWorksConf: Experience exploratory testing
TestWorksConf: Experience exploratory testing
 
Development without Testers: Myth or Real Option?
Development without Testers: Myth or Real Option?Development without Testers: Myth or Real Option?
Development without Testers: Myth or Real Option?
 
Breaking Dependencies to Allow Unit Testing - DevIntersection Spring 2016
Breaking Dependencies to Allow Unit Testing - DevIntersection Spring 2016Breaking Dependencies to Allow Unit Testing - DevIntersection Spring 2016
Breaking Dependencies to Allow Unit Testing - DevIntersection Spring 2016
 
1 of 41
Download to read offline
selenium Users Anonymous 34 Se by Dave Haeffner
Selenium Users Anonymous
Show of hands How many of you are human? 100% How many of you use Selenium? Most Just getting started in test automation? 40% Comfortable with test automation? Using it regularly? 50% Test automation guru's? 1-2 people Development background 60% Tech proficient -- know enough to be dangerous? 15% Scared to death of tech? Some smiles but no one raised their hand
Selenium Users Anonymous
Exercise 1 What are your biggest pain points when it comes to using Selenium? Or Test Automation in general? In 5 minutes, write them down One issue per post-it Keep it concise Keep it to yourself
Selenium Users Anonymous
Evolutionary paths of Selenium Suites Se-IDE Simple tests captured that can grow into an albatross that is unmaintainable, brittle, and makes you spit nails. Most people (hopefully) realize this is not the way to go and export their tests into a real programming language Se-RC What you used to export your Se-IDE tests into. But not anymore, it's reaching end-of-life. Se-WebDriver Behold the future! Pure browser integration (read: fast and awesome)
Entrees Breaking/brittle/untrustworthy/unstable tests Hard to decipher results Hard to interpret what your tests test False positives Slow Too many tests Ownership Maintenance Out of band communication Works on my machine
Desserts Manual testing as a crutch QA as a bottle neck Defects caught late in the cycle
Selenium Users Anonymous
Exercise 2 You have 5 minutes Come to the front of the room with your post-its from Exercise 1 Post them on the wall Group like with like
12
1 - Admit You Have a Problem
It's not entirely your fault Selenium was Built by Developers for Developers. Some assembly required, oh, and RTFM. Along the way someone built a Selenium IDE to make it easy for mere mortals to capture automated tests, while this got newcomers to use the tool, it reinforced some bad habits -- causing people to build painful and unmaintainable test suits While there is a great community and good documentation, there is not a cohesive, standard "here you guy" guide... yet It can be overwhelming for a newcomer. But most times you duct-tape together a solution to your problem with answers found in blog posts and stack overflow threads.
2 - Take a deep breath
It's Okay Take into account that you are doing something difficult And that you are outside of your comfort zone, and that will likely not change But it's okay, people have gone before you And you have to fail to learn
3 - Try looking at things differently
Try these on for size Zooming Be able to "zoom in" deal with the day to day details while also "zooming out" to see the big picture -- making sure that both the high and low level are aligned Effectiveness = Quality x Acceptance It's much easier to improve quality than it is acceptance. Where are you spending your effort? Personas & Motivation QA's are either very technical or very analytical. Engineers tend to want to solve problems. Understanding these things will help you unlock untapped potential.
4 - Pump some Tech iron
Pump you up I've seen a lot of people who are fearful of technology but they play on the razor's edge of it. Weird. Pump some iron. Learn to code. Or least, learn the broad strokes. It will help you with your "tech comfort". Resources abound -- books, blogs, meetups, classes, conferences, etc. Sharpen your tools
5 Find Your Inner Napoleon, Develop A Strategy
It's less about the tool and more about the people Think big, keep it simple, capture value Envision the future and build a roadmap to it Some example questions to ask Type of app? Data heavy? Rich UI? How does the business makes money? Available Usage metrics? Browsers you care about? Current S/W dev workflow? Release cycle? Resources available? QA Team make-up? Dev Team make-up? Top-down support? Bottom-up interest?
6 Break Down The Wall
Make friends with the Developers -- work together! Choose a technology stack for building a framework that makes sense for you and your team -- this is a topic in and of itself To start, focus on you and your team's context -- e.g. strengths, comfort, confidence, and excitement Also consider your company's current tech stack Put it to the team to try and decide Also, have them give it a fun name -- this stuff matters!
7 - Learn the terrain
Try stuff Commit and make progress Learn what you need along the way to become effective Fill in your knowledge gaps as you go And don't be afraid to ask questions
8 - Test Less
Perform a Pareto analysis. It posits that 80% of the value comes from 20% of the inputs -- and it's correct. Consider this a power law. A good heuristic to follow for this -- 3 buckets. There are 3 things you should care about when it comes to testing: Table Stakes, Money Makers, Back Breakers ● Table Stakes: The simple things that should just work ● Money Makers: How does your business make money/generate value? ● What is risky, prone to breakage, overly complicated? If you get these right, then you are on to something great This requires tight collaboration with business and tech
9 - Test Well
Grow a framework that can be used by the whole team Proper abstraction -- e.g. Page Object Pattern Single Responsibility Principle -- do one thing and do it well Atomic testing -- each test should be able to run and be understood independently Test under the hood -- you can verify important business logic against the underlying API/WebServices Parallelize -- e.g. SauceLabs, DIY w/ SeleniumGrid -- SeleniumHQ has a list of all of the players out there Provide timely, valuable feedback for both Tech & Biz -- e. g. Use CI & sexy reporting Remember that manual/exploratory testing has its place
10 Keep It Lean & Optimize
This is a good way to gauge your success and mine for ways to improve Continually assess how you're doing Are you solving problems or just testing for testing sake? Make changes, try, observe, make changes, try, observe… you get the idea! Once you are doing things that are generating value, optimize them, not before
11 Mailing Meetups Lists Google Blogs Groups Stack Twitters IRC Overflow Quora
12
Learn something that you had to cobble together? Write a blog post See a question on Stack Overflow that you know the answer to? Answer it Have an idea worth sharing? Give a talk a local Selenium/Tech Meetup No Local Selenium Meetup? Start one! Contact ashley@saucelabs.com to find out how
12 Steps Recap Assess reality, always start with why, be strategic, test smarter, pay attention, you've always got a friend, pay it forward This can be a linear progression, but it shouldn't be You can practice any of these steps individually. In fact, I encourage you to You don't need to be an expert to pay it forward You become an expert by paying it forward If you have a strategy and it doesn't seem to be working, be lean and revisit things Always be learning and pumping your tech iron And don't forget to breath But most importantly, and I hope this has been readily apparent. It's less about the tool than you think.
@TourDeDave dave@arrgyle.com
Appendix Recommendations
Sites ● Official Selenium Blog ● Sauce Labs Blog ● Selenium HQ ● Selenium Meetups ● Quora ● Agile Testing Mailing List ● Selenium Users Google Group
Posts ● What Makes A Good Test Suite ● Testing vs Checking ● Get Out of the Quality Assurance Business ● Zooming
Books ● Bridging the Communication Gap ● Specification By Example ● Learn To Program ● Everyday Scripting with Ruby ● The Cucumber Book ● How Google Tests Software ● Continuous Integration

More Related Content

Selenium Users Anonymous

  • 1. selenium Users Anonymous 34 Se by Dave Haeffner
  • 3. Show of hands How many of you are human? 100% How many of you use Selenium? Most Just getting started in test automation? 40% Comfortable with test automation? Using it regularly? 50% Test automation guru's? 1-2 people Development background 60% Tech proficient -- know enough to be dangerous? 15% Scared to death of tech? Some smiles but no one raised their hand
  • 5. Exercise 1 What are your biggest pain points when it comes to using Selenium? Or Test Automation in general? In 5 minutes, write them down One issue per post-it Keep it concise Keep it to yourself
  • 7. Evolutionary paths of Selenium Suites Se-IDE Simple tests captured that can grow into an albatross that is unmaintainable, brittle, and makes you spit nails. Most people (hopefully) realize this is not the way to go and export their tests into a real programming language Se-RC What you used to export your Se-IDE tests into. But not anymore, it's reaching end-of-life. Se-WebDriver Behold the future! Pure browser integration (read: fast and awesome)
  • 8. Entrees Breaking/brittle/untrustworthy/unstable tests Hard to decipher results Hard to interpret what your tests test False positives Slow Too many tests Ownership Maintenance Out of band communication Works on my machine
  • 9. Desserts Manual testing as a crutch QA as a bottle neck Defects caught late in the cycle
  • 11. Exercise 2 You have 5 minutes Come to the front of the room with your post-its from Exercise 1 Post them on the wall Group like with like
  • 12. 12
  • 13. 1 - Admit You Have a Problem
  • 14. It's not entirely your fault Selenium was Built by Developers for Developers. Some assembly required, oh, and RTFM. Along the way someone built a Selenium IDE to make it easy for mere mortals to capture automated tests, while this got newcomers to use the tool, it reinforced some bad habits -- causing people to build painful and unmaintainable test suits While there is a great community and good documentation, there is not a cohesive, standard "here you guy" guide... yet It can be overwhelming for a newcomer. But most times you duct-tape together a solution to your problem with answers found in blog posts and stack overflow threads.
  • 15. 2 - Take a deep breath
  • 16. It's Okay Take into account that you are doing something difficult And that you are outside of your comfort zone, and that will likely not change But it's okay, people have gone before you And you have to fail to learn
  • 17. 3 - Try looking at things differently
  • 18. Try these on for size Zooming Be able to "zoom in" deal with the day to day details while also "zooming out" to see the big picture -- making sure that both the high and low level are aligned Effectiveness = Quality x Acceptance It's much easier to improve quality than it is acceptance. Where are you spending your effort? Personas & Motivation QA's are either very technical or very analytical. Engineers tend to want to solve problems. Understanding these things will help you unlock untapped potential.
  • 19. 4 - Pump some Tech iron
  • 20. Pump you up I've seen a lot of people who are fearful of technology but they play on the razor's edge of it. Weird. Pump some iron. Learn to code. Or least, learn the broad strokes. It will help you with your "tech comfort". Resources abound -- books, blogs, meetups, classes, conferences, etc. Sharpen your tools
  • 21. 5 Find Your Inner Napoleon, Develop A Strategy
  • 22. It's less about the tool and more about the people Think big, keep it simple, capture value Envision the future and build a roadmap to it Some example questions to ask Type of app? Data heavy? Rich UI? How does the business makes money? Available Usage metrics? Browsers you care about? Current S/W dev workflow? Release cycle? Resources available? QA Team make-up? Dev Team make-up? Top-down support? Bottom-up interest?
  • 24. Make friends with the Developers -- work together! Choose a technology stack for building a framework that makes sense for you and your team -- this is a topic in and of itself To start, focus on you and your team's context -- e.g. strengths, comfort, confidence, and excitement Also consider your company's current tech stack Put it to the team to try and decide Also, have them give it a fun name -- this stuff matters!
  • 25. 7 - Learn the terrain
  • 26. Try stuff Commit and make progress Learn what you need along the way to become effective Fill in your knowledge gaps as you go And don't be afraid to ask questions
  • 27. 8 - Test Less
  • 28. Perform a Pareto analysis. It posits that 80% of the value comes from 20% of the inputs -- and it's correct. Consider this a power law. A good heuristic to follow for this -- 3 buckets. There are 3 things you should care about when it comes to testing: Table Stakes, Money Makers, Back Breakers ● Table Stakes: The simple things that should just work ● Money Makers: How does your business make money/generate value? ● What is risky, prone to breakage, overly complicated? If you get these right, then you are on to something great This requires tight collaboration with business and tech
  • 29. 9 - Test Well
  • 30. Grow a framework that can be used by the whole team Proper abstraction -- e.g. Page Object Pattern Single Responsibility Principle -- do one thing and do it well Atomic testing -- each test should be able to run and be understood independently Test under the hood -- you can verify important business logic against the underlying API/WebServices Parallelize -- e.g. SauceLabs, DIY w/ SeleniumGrid -- SeleniumHQ has a list of all of the players out there Provide timely, valuable feedback for both Tech & Biz -- e. g. Use CI & sexy reporting Remember that manual/exploratory testing has its place
  • 31. 10 Keep It Lean & Optimize
  • 32. This is a good way to gauge your success and mine for ways to improve Continually assess how you're doing Are you solving problems or just testing for testing sake? Make changes, try, observe, make changes, try, observe… you get the idea! Once you are doing things that are generating value, optimize them, not before
  • 33. 11 Mailing Meetups Lists Google Blogs Groups Stack Twitters IRC Overflow Quora
  • 34. 12
  • 35. Learn something that you had to cobble together? Write a blog post See a question on Stack Overflow that you know the answer to? Answer it Have an idea worth sharing? Give a talk a local Selenium/Tech Meetup No Local Selenium Meetup? Start one! Contact ashley@saucelabs.com to find out how
  • 36. 12 Steps Recap Assess reality, always start with why, be strategic, test smarter, pay attention, you've always got a friend, pay it forward This can be a linear progression, but it shouldn't be You can practice any of these steps individually. In fact, I encourage you to You don't need to be an expert to pay it forward You become an expert by paying it forward If you have a strategy and it doesn't seem to be working, be lean and revisit things Always be learning and pumping your tech iron And don't forget to breath But most importantly, and I hope this has been readily apparent. It's less about the tool than you think.
  • 39. Sites ● Official Selenium Blog ● Sauce Labs Blog ● Selenium HQ ● Selenium Meetups ● Quora ● Agile Testing Mailing List ● Selenium Users Google Group
  • 40. Posts ● What Makes A Good Test Suite ● Testing vs Checking ● Get Out of the Quality Assurance Business ● Zooming
  • 41. Books ● Bridging the Communication Gap ● Specification By Example ● Learn To Program ● Everyday Scripting with Ruby ● The Cucumber Book ● How Google Tests Software ● Continuous Integration