My first experience with usability testing was on an agile team where the product we were building was being designed with the help of an in-house usability expert. He helped design the user interface (UI) of the application and conduct usability study on the beta version of the software to determine the ease of use of the application.
Though the experience was limited in terms of the interaction we had with the user representatives and the sessions conducted, the feedback we received opened up lots of new avenues for the tester in me around the learnability, understandability and attractiveness of the application I was testing.
Usability has matured a lot over the years. It’s now an essential software characteristic in today’s web and mobile applications. In my article published at the TestRail blog, I discuss ways of performing Usability tests and developing a mindset for Usability in an agile context.
“Selenium Integration with Cucumber for an extended BDD framework”
This workshop will cover • Practical issues faced by most testing teams • Behavior Driven Development – the definition and need • Extending the Agile User stories and acceptance criteria in BDD scenarios • Cucumber as a BDD tool • Integration of Cucumber with Selenium in order to perform functional tests • Demo using Cucumber with Selenium with a real use case • Usage and Benefits of BDD In agile teams
For program details and complete agenda of the event, visit the website
The event will have enthusiasts exchanging ideas on advancing the present and future of Selenium and will bring together bright minds to give insightful talks pertaining to Selenium practice that are solution-focused, and foster learning and inspiration. * Solutions for Practical issues of testing. * Integration of Selenium with other testing tools. * Providing a key meeting place for Selenium Professionals and Executives from leading IT organizations. * A platform to share your research and experience
Looking forward to sharing my experience, learning from skilled professionals in the area and networking with the brightest minds at the event!
Agile poses many challenges to the development team, most of them pertaining to time. Teams are perpetually under pressure to deliver working software at a fast pace, leaving minimum time for anything else. When testing on an agile project, learning how to write lean documentation can save precious time. Furthermore writing lean documentation can help rework efforts by focusing only on what’s really necessary.
The Agile Manifesto emphasizes working software over comprehensive documentation, but most agile teams interpret this wrong and treat documentation as something to be avoided, owing to time constraints. The manifesto states a lesser focus on comprehensive documentation, but some documentation is still needed for the project and any related guidelines being followed. Attaining this balance is a challenge.
Documentation is a necessary evil. We may think of it as cumbersome and time-consuming, but the project cannot survive without it. For this reason, we need to find ways to do just enough documentation — no more, no less.
For example, in a traditional test design document, we create columns for test case description, test steps, test data, expected results and actual results, along with preconditions and post-conditions for each test case. There may be a very detailed description of test steps, and varying test data may also be repeatedly documented. While this is needed in many contexts, agile testers may not have the time or the need to specify their tests in this much detail.
As an agile tester, I have worked on teams following a much leaner approach to sprint-level tests. We document the tests as high-level scenarios, with a one line description of the test and a column for details like any specific test data or the expected outcome. When executing these tests, the tester may add relevant information for future regression cycles, as well as document test results and any defects.
Continuing the discussion on the Hawaii Missile Alert which made headlines in January 2018 and turned out to be a false alarm and ended up raising panic amongst almost a million people of the state all for nothing, (read here for detailed report) I would like to bring back the focus on implications of poor software design leading to such human errors.
Better software design is aimed at making the software easier to use, fit for its purpose and improving the overall experience of the user. While software design focuses on making all features easily accessible, understandable and usable, it also can be directed at making the user aware of all possibilities and implications before performing their actions. Certain actions, if critical, can and should be made more discrete than the others, may have added security or authorisations and visual hints indicating their critical nature.
Some of the best designers at freelancer.com came together to brainstorm ideas for better software design and to revamp the Hawaii government’s inept designs. They ran a contest amongst themselves to come up with the best designs that could avoid such a fiasco in future.
Sarah Danseglio, from East Meadow, New York, took home the $150 grand prize, while Renan M. of Brazil and Lyza V. of the Philippines scored $100 and $75 for coming in 2nd and 3rd, respectively.
Here is a sneak peek into how they designed the improved system :Read More »
Software impacts human lives – let us put more thought into it!
Here is what happened and my take on how software design may have been partly responsible and could be improved >>
Miami state in the US received a massive panic attack on Saturday the 13th of January 2018. More than a million people in Hawaii were led to fear that they were about to be struck by a nuclear missile due to circulation of a message sent out by the state emergency management. The message sent state wide just after 8 a.m. Saturday read: “BALLISTIC MISSILE THREAT INBOUND TO HAWAII. SEEK IMMEDIATE SHELTER. THIS IS NOT A DRILL.”
The residents were left in a state of panic. People started scrambling to get to safe places, gathering supplies and even saying their goodbyes. Some took shelter in manholes, some gathered their kids into the most sheltered rooms in their homes like bathrooms or basements, some huddled in their closets and some sent out goodbye messages to their loved ones.
Turned out it was a false alert. Around 40 minutes later, the agency sent out another message saying that it was a false alarm sent out by mistake!
The questions being asked was – how could this happen and why did it take 40 minutes to check and issue an all clear?
Why Did This Happen?
Investigations into the incident were revealed and the governor stated that “It was a procedure that occurs at the change of shift which they go through to make sure that the system is working, and an employee pushed the wrong button.”
The error occurred when, in the midst of a drill during a shift change at the agency, an employee made the wrong selection from a “drop-down” computer menu, choosing to activate a missile launch warning instead of the option for generating an internal test alert. The employee, believing the correct selection had been made, then went ahead and clicked “yes” when the system’s computer prompt asked whether to proceed.
Analysing the Root Cause
But is the fault only at human level? The software being used for such critical usage also needs to help out to avoid the possibility of such human errors.
After all triggering such a massive state-wide emergency warning should not have been as simple as push of a wrong button by a single person!
Could a better design of the software have prevented this kind of scenario from happening?
As reported, the incorrect selection was made in a dropdown – which lets imagine would look something like this-
After the selection was made, the system sent a prompt and the employee, believing the correct selection had been made, then went ahead and clicked “yes”.
So by this information we can assume that the prompt would have been something generic like
Though it definitely is a human error but isn’t the system also at fault for letting this happen so easily?
Better Design Ideas – More Thought – Improving Your Software
By putting in some extra thought into design of the software we can make it more robust to avoid such incidents.
Here are some things that could have helped design it better –
Do not have the TEST options placed right next to the ACTUAL emergency options!
Have different fields or perhaps different sub menus inside the dropdown as categories.
>> Always have the TEST category of warnings higher up in the list
>>Have the Default Selection in the dropdowneither as BLANK or as one of the TEST warnings and not the actual ones
>>Having the actual warnings section lower down and separated away from the similarly worded TEST warning would ensure lower chance of wrongful selection of the similar named option from the dropdown
The prompt message must be made unique to each scenario and in case of selecting a real warning issue action, the prompt must ask the user to specify the emergency.
>>Make the prompt appear critical with use of color and text
>>A critical prompt must catch the user’s attention and not be similar to the other screens and popups of the system, to avoid the possibility of clicking on it in a hurry.
>>Placement of Yes and No buttons on unusual sides (Yes is on the left which is not typical) avoids the click of the button – also used Red and Green to signify the importance situation. Red is the usual code for danger.
Additional level of authorisation must be added to the scenarios of real emergency warnings being issued. So, for the TEST actions, user may proceed and begin the drill but in case they select ACTUAL warning then the steps take it to another level of authorisation where another employee – a peer or a senior- reviews the action and performs the final warning issue.
>>This prevents erroneous actions and also some possibility of hackers or notorious people issuing false warnings just by gaining access via one user.
>>Define your hierarchy of users or approvals for each case of emergency.
These ideas may sound basic but all these are components of good Usability of the software, its appropriateness of purpose and setting up basic security in usage of the application.
We are just playing around human psychology, easier understand-ability and attention spans.
Let us endeavour to give a little more ‘thought’ to the system
Boris Beizer, in his book Software Testing Techniques (1990) coined the term pesticide paradox to describe the phenomenon that the more you test software, the more immune it becomes to your tests.
Just like, if you keep applying the same pesticide, the insects eventually build up resistance and the pesticide no longer works. Software undergoing the same repetitive tests build resistance to them, and they fail to catch more defects after that.
Software undergoing the same repetitive tests eventually builds up resistance to them.
As you run your tests multiple times, they stop being effective in catching bugs.
Moreover, part of the new defects introduced into the system will not be caught by your existing tests and will be released onto the field.
Solution: Refurnish and Revise Test Materials regularly
In order to overcome the pesticide paradox, testers must regularly develop newer tests exercising the various parts of the system and their inter-connections to find additional defects.
Also, testers cannot forever rely on existing test techniques or methods and must be on the look out to continually improve upon existing methods to make testing more effective.
It is suggested to keep revisiting the test cases regularly and revising them. Though agile teams provide little spare time for such activities, but the testing team is bound to keep planning these exercises within the team in order to keep the best performance coming. A few ideas to achieve this:
Brainstorming sessions – to think of more ideas around the same component testing
Buddy Reviews – New joinees to the team are encouraged to give their fresh perspective to the existing test scenarios for the product, which might get some new cases added.
Strike out older tests on functionalities that are changed / removed
Build new tests from scratch if a major change is made in a component – to open a fresh perspective