It is interesting how test tools seem to move through cycles. We mentioned in a previous post how many test tools were listing AI as their differentiator. This year we seem to have encountered two different items:
Codeless – several test tools that have come to our attention recently are claiming to be codeless. Either they are driven by capture of the transactions and generate a script from that capture or they use keywords to create a script. Either way they seem to have removed the need to work at the scripting level.
Driven by a design – in a somewhat similar fashion they use a design or a diagram to drive the scripts and conceal the actual scripting again.
We have asked a few questions about these tools (the same ones that no doubt occurred to you)
- How can they be modified (especially when dealing with changing requirements in a fluid project)?
- How can they be maintained across several releases so that the investment is repaid (positive ROI)?
- How can they be controlled (for those clients who need an auditable trail)? Some of our clients would not be happy with just the actual scripts landing in a controlled repository. They would also want the script design keywords or diagram to be controlled as well. This way they have the whole story.
We have received varying answers to these questions. Some seem prepared, others do not.