July 1, 2011 – Last week’s INCOSE International Symposium was refreshing. The sessions and events offered great opportunities to network with other industry professionals interested in systems and not just software, and I joined some useful tutorials in systems engineering. I’m still distilling my thoughts on “systems engineering vs. software engineering”.
Participating on the “Is Requirements Engineering Really Necessary?” panel with Brian Berenbach, Mark Sampson, and James Hulgan was great fun. We don’t have the official session survey ratings yet, but we drew an audience of several hundred who never ran out of questions for us. My top 3 takeaways from our discussions are:
- Emphasize activities, not titles. The more stakeholders and team members who understand and can use Requirements Engineering methods effectively, the more the system and business will benefit. RE advocates have to remember, though, that most systems engineers aren’t, and don’t want to become, “requirements engineers” or even “requirements analysts”. They are committed “control systems engineers”, “electronics engineers”, “software system engineers”, or “power systems engineers” who are passionate about their own domains of expertise. To be most useful, training in requirements elicitation and analysis should be aligned to their domain worlds, instead of expecting systems engineers to align with the world of RE.
- System requirements need systems thinking, too. How formally requirements are managed should depend on the risks and consequences – not all requirements are “created equal”. Likewise, how requirements are documented should depend on who they are being documented for – the audience who needs to understand and use them. With today’s increasingly complicated systems and escalating time-to-market pressures, the same old mountain-of-text-documents approaches don’t scale; we need to adapt, and start ’system-engineering’ how we handle our requirements to fit the needs of the business and the system.
- No silver RE bullets. Requirements engineering isn’t a panacea that can solve any and all problems in a system. Requirements aren’t mushrooms to be “gathered” for analysis. They’re more like truffles that need to be carefully searched-for and unearthed. RE techniques can help you find the truffles and ensure that key needs aren’t overlooked. And RE can help you analyze and manage needs to ensure that requirements are well-defined, prioritized, verifiable, and necessary. RE can’t guarantee that you’ll never miss a requirement, include extraneous features, or misinterpret an important aspect. Using a mixture of senior and junior staff can help: experienced people are guided by the pain that came from overlooking key requirements or quality attributes in the past, and junior people can help the team avoid “expertosis”, by questioning assumptions and asking “why?”.