Story 2 - Focused On Solving the Customer’s Real Pain Points, than Problem Symptoms – Overcoming the Constraints/Traps Imposed by Tools
A US based insurance company was automating its application testing with HP QTP. The functionality slated for testing was insurance quote generation in the form of PDFs. The customer’s side of testers had trouble automating the functionality, as QTP had no out of the box support for data extraction or validation from the PDFs. This was a critical scenario to be automated as there were many alternates to be tested. It was a tedious task of validating the calculations and values in the table. The automation testers at customer’s side ensured all the ground work by researching on different tools that could help them, and then approaching HP to help them with the PDF validation, yet, nothing worked in their favor. As one another try, they approached SPAN to help them with it.
Description:
We started with a short discussion on this problem and explored the issue. We strongly believe that tool is like a vehicle and automation is the skill like driving. When you know how to drive, it does not matter which car you are driving. Same was the case here; we believe that what is shown to us is the symptom of the problem and not the real problem. Thus, we started looking for the right problem that we are expected to solve. Eventually, we defined the problem that we need to address as “validate the insurance quote data in the form of tables against a set of expected values” rather than the PDF data validation.
With this definition of the problem, we started understanding the technical aspects behind this quote creation. In that case, they were using Apache Formatting Objects Processor (FOP), wherein the application was reading the formatting object (FO) tree and rendering the resulting page to PDF as output. This point helped us in moving forward with the solution. On exploration about the FOP, we found out that it is possible to capture the Area Tree XML, which is an internal representation of the resultant PDF document, in exactly the same layout of pages and contents.
This was enough for us to formulate a solution to the real problem. Our solution was to validate the insurance quote data in the Area Tree XML and to validate the PDF creation so that it depicted the data correctness we have exercised and also the quote creation in a PDF format. We implemented a VBScript that works within the QTP premise, validating the data in the Area Tree XML exactly the way it was required to validate the insurance quote in the PDF format. Further, we validated the existence of a new PDF file to check the creation of the insurance quote.
This solution solved the customer’s problem and the same solution is being used since more than 3 years and it is replicated for similar checks at other applications.
Take Home:
Many times, the symptom of problem is showcased as the problem, but, at SPAN, we take care to find the real pain point and strive to address that, rather than solving a symptom. In this case, the tool did not support extracting the data from PDFs, but the real problem that the customer had was not of the tool support, but of testing the insurance quotes. The solution that we achieved is exactly that.
Read Story 3 at http://spansys.blogspot.in/2015/02/testing-stories-from-spans-trenches_18.html
No comments:
Post a Comment