NFRs Considered Harmful

By: on August 27, 2020

I’ll make this short: Defining Non-Functional Requirements (NFRs) is dangerous. That’s the tweet.

An NFR is broadly defined as a ‘quality’ of the software, rather than what it ‘does’. So ‘the software shall add two numbers’ is a functional requirement, while ‘the software shall run within two minutes’ or ‘the software shall be maintainable’, or ‘scalable’, or whatever, is deemed ‘non-functional’. I submit that this is a dangerous distinction to make when running a software project.

In the large, everything is functional (This has been formalised into Hyrum’s Law: “With a sufficient number of users of an API, it does not matter what you promise in the contract: all observable behaviours of your system will be depended on by somebody.”)

I’ve noticed that if you call something ‘non-functional’ it immediately becomes deprioritised, a tier two requirement. What people hear is “this would be nice to have but we don’t think anyone important actually cares much about it.”

As an example from Wikipedia, take performance requirements. They say that ‘software shall run in time X’ is a non-functional requirement. Well, it’s not important until it is. If there are no deadlines, there is no problem. Is it even a requirement? But if your software has to process a bank’s worth of securities in two hours, it is a requirement. ‘Non-functional’, my horse.

The danger of reflexively saying ‘it’s performance, that’s non-functional’ is the team doesn’t pay attention to it until too late. Just eliminate the category.


Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>