Unique in the software testing industry, eggPlant actually sees how applications are performing by watching and interacting with the software under test – engaging with its screen images as a seeing, human user would!
Over 500 organizations worldwide rely on robotic tools from TestPlant to test software for industries including defense, life science, finance, IT, telecoms and media. And its premier test tool, eggPlant, broke new ground with a unique, highly accurate, way to automate software testing. Unlike traditional object-based test programs, eggPlant looks at, and tests, screens as an expert human eye does – yielding remarkable precision and ease. In some important areas however, OCR was needed to simplify the process and enhance results. But only ABBYY's FineReader Engine was up to the task.
Completely platform independent, eggPlant works over a VNC protocol to control the system under test. It can be trained to spot bugs just as a human would, tuned to detect colors, operate in dynamic environments using Silverlight, Flex and Flash and be adjusted for customized screens. But although eggPlant had several methods for dealing with text, none would work with mobile phone applications. This meant that TestPlant risked missing out on a huge, rapidly growing, market.
For TestPlant, the solution was plain to see: Integrate Optical Character Recognition (OCR) into eggPlant, and enable it to read text directly–no plug-ins or intermediary steps required.
Spurred by market potential – and the desire to make eggPlant a more intelligent and productive tool when testing applications with more text than images – TestPlant began testing OCR engines. But according to Jason Noble, Director of Product Development, early results weren't encouraging: "The quality rates of our integration tests were disappointing. When testing, you can't have failures, the test tool has to work every time. Especially in regression testing, where compounding an error can badly impact development cycles."
Having already evaluated several vendors, Noble found ABBYY. "The ABBYY® FineReader® engine stood out. It was the most impressive because it worked across the operating systems of several different mobile phones. I got the free trial version and tested it against several different mobile phone screenshots – all of which it translated accurately."
Satisfied with accuracy, TestPlant began implementation. Integration of the FineReader Engine took three months from start to launch and was performed on three platforms: MacOS, Linux and MS Windows – a process that also benefitted from ABBYY's commitment to customer support.
With its new OCR capabilities, eggPlant provides further testing functionality for its customers particularly in the mobile platform. Plus, FineReader engine extends eggPlant's capabilities in dramatic new ways. It can now read and find virtually any text displayed by the system being tested – without sacrificing the advantages of a pure image-based approach. In addition, eggPlant remains completely agnostic to the underlying technology, interacting with the software it tests exactly as a seeing, and reading, human user would.
eggPlant with FineReader engine is now a more intelligent, intuitive and productive test tool. In addition to seeing and recognizing objects on-screen purely by their appearance, eggPlant can now read text and identify elements by their text labels without any need to communicate with the underlying system, objects, or language used to construct them. Plus, eggPlant's visual approach is easier for users because it parallels the actions of a human tester. Thus, it works out-of-the-box with any application on any system, from desktop to web to server to mobile device. But that's just the beginning, according to Noble: "Our team looks forward to the future, where we'll identify more common ground to collaborate with ABBYY."
OCR enhances eggPlant's ability to test tough-to-automate technologies that leave users of other tools scrambling to fill in the gaps.