by-studio - Fotolia

The Met Office automates functional testing of weather warnings sites

The UK’s national weather service automates testing using eggPlant to ensure integrity of weather warnings system

The Met Office has used the eggPlant automated testing tool from Testplant to help it improve the reliability of its National Severe Weather Warning System (NSWWS).

Speaking to Computer Weekly about the test suite, Ian Gentry, test lead at The Met Office, said: “Essentially, we were looking to improve the confidence in testing one of our critical systems.”

The weather warning system feeds weather alerts to the public and emergency services through its web sites, email or through text alerts.

To ensure accuracy of weather alerts, the Met Office created Warnings Manager, a tool to recreate weather warnings in XML format. The Met Office’s IT team wanted to be able to generate warning scenarios in Warnings Manager and then check that these were presented correctly on multiple websites.

Previously, The Met Office ran manual functional tests to ensure that the output presented on the weather warnings sites was correct. Gentry said it used to take a long time to do regression testing and the tests were not repeatable.

Rather than manually check theses websites, The Met Office is now using eggPlant to verify the warnings are presented correctly, with Gentry saying the organisation went with eggPlant “because it wasn’t invasive”.  The testing involves an image comparison test to ensure the output from NSWWS is as expected.

The eggPlant functional testing tool uses a patented image-based approach to user interface testing. According to TestPlant, this allows it to interact with any device just by looking at the screen in the same way a user would.

The automated testing has enabled The Met Office to slash the time spent previously on regression testing from half a day to 10 minutes, allowing the IT team to be significantly more productive.

Read more about The Met Office

“Apart from saving time, we can do tests much quicker and do more of them,” said Gentry, adding that this is moving The Met Office along the path from waterfall to agile software development, which in turn means more changes. As such, automating testing is set to become a key part of the organisation’s software development process.

The project coincides with the 30-year anniversary of the Great Storm of October 1987, the most severe weather disaster in recent history. Following this, reforms to the Met Office’s severe weather warning systems were instigated, leading to an increased volume of weather warnings, with the NSWWS service being just one of these reforms. 

Read more on Software development tools