Important tip: How to raise bugs automatically

In Automation, Tools on August 5, 2017 at 8:03 am

Congratulations! you just automated a suite of end to end tests for your application.  You run these tests for every build, look at the reports and everyone is happy.  Really?

The report contains failed tests too.  You need to raise bugs for that.  What does that involve?

  1. A suitable title
  2. Steps involved to reproduce the bug
  3. Provide information on: OS, Architecture, Browser version, Application Version, CPU information, Memory usage, Disk Usage, CPU usage
  4. Expected Results, Actual results, Screenshots, Logs

Test processes can be inundating  and exasperating for many testers leading to loss of motivation in Testing.  This is a hidden risk.

The aftermath of automated test execution is one such process. However help is at hand when you use the AutomatedBugRaiser







Are we headed towards Prediction Driven Development?

In Automation, Machine Learning, Prediction on December 2, 2016 at 12:49 pm

In a recent tweet about Sanity Testing,  I started thinking about certain Machine Learning tools in the market which seem to predict defects for various change requests submitted.

Now assuming that this defect prediction tool “actually works”,  it would become extremely useful to conduct an extended Sanity (in the least).  This would lead to major issues caused by the change requests to be resolved even before it gets to the hands of a tester, thereby saving money and time.  Changes keep happening up-to the last minute of a Product release, and this could well be another way to develop Software.

Quick Tip: Anomaly Detection to build trust with Customers

In Automation, Bills, Invoices, Machine Learning on November 21, 2016 at 4:05 pm

In a recent incident, one of my friends took his old vehicle for servicing.  Among other things, a small bulb was replaced and the bill for just that part was a whopping Rs 300/-.  The explanation given to him was that the bulb costs Rs.150/- and the labor was Rs. 150/-. Whereas the same would cost nearly half the price a few months ago. Courtesy a history of maintaining invoices, he was able to question the prices.

He was also at the receiving end for some of his other bills, when he saw some extra charges which were never charged in his earlier bills (which he had maintained).  It was later found out, that these were some “errors”.   I wonder how many people would really do all this.

For people who do not store so many bills/invoices and keep track of unusual deviations in them.  They could be ripped off.

Solution: In these days of digital records (including bills and invoices), a machine learning algorithm called Anomaly Detection could help.  A company could basically spot the outliers or anomalies and strive to eliminate them thereby building trust with their customers.  For the customers, they could benefit if certain flags are reported in their bills whenever a large variation in prices is seen in their invoices for many billing products.

Some bills, payslips etc, could contain breakups of various categories, and any missing category or additional categories can also be flagged.

All in all, a win win for both Customers and Businesses


%d bloggers like this: