Do you think it is time for QTP to die? Paul Hammant posted on his blog that he thinks so. He makes some good points, but I disagree.
His first point discusses developers not using the tool. In my experience, most developers have rejected the tool when offered or failed to use it once it was delivered. Not because it was too expensive (concurrent licensing can ease the installation burden), but because they did not want to take the time to automate their tests. Even in this modern day, many development shops still are not writing unit tests! In the ideal world, all applications would have undergone a battery of automated tests before turning over the project to a testing team, but that is not the world most of us live in.
His second point discusses new lightweight testing tools that are either free or come at a low-cost. The terms “free” and “low-cost” should not be taken lightly, as the cost to actually implement and use a tool is never free! You have to be trained on the tool, gain experience on how it is used, and troubleshoot software issues with the tool itself. Some of these tools are not actively supported or lack a lot of the training material to get testers up to speed quickly. One of the biggest benefits to QTP is that you have “one tool to rule them all”. A single tool to handle testing a variety of technologies should not be underestimated. Sure, it may take a little time to adjust to testing a WPF-based application in QTP if you’ve only ever tested web-based applications, but at least you maintain the familiarity with the base tool and underlying automation concepts.
The remainder of his article primarily focuses on the growing support of Selenium. I have not personally used Selenium and have only recently looked at the product, but my initial thoughts were far from positive. The first webinar I attended about the technology did more to talk about what Selenium cannot do instead of what it is capable of. I have since attended additional webinars on the tool, and have yet to see a good presentation on how you implement this technology. Perhaps I have been spoiled by the IDE that QTP provides or the tight integration with Quality Center / ALM. To gain anything close to the QTP/QC/ALM combo, you have to start purchasing additional, 3rd-party tools. In the end, remember that investing in Selenium is also an investment in a single technology for web-based testing. If you ever need to automate .NET or Win32 applications, you will have to seek out, learn, and implement an entirely new technology.
Does QTP have it’s faults? Of course it does! The whole reason we initially developed Test Design Studio was to cover some of those faults. The fact remains, however, that QTP (even with it’s flaws) is a far more robust and mature product than any of the “free” or “low-cost” tools we see on the market today. Combine that with Test Design Studio, and you have a winning combination!
What do you think? Feel free to post your thoughts in the comments.