There is probably something to be said for certifying software engineers.
Edit: After attending the talk at the IITP AGM I’ve changed my stance on this issue—Novopay probably doesn’t fit software engineering, per se, so much as IT. I’ll write a short post explaining when I get time.
I was at the Engineering Professions Forum last weekend, where in one session a student pointed out that part of IPENZ’s lack of relevance to those in electronics and computers was that being a chartered professional engineer (CPEng) doesn’t apply to them. To that, someone interjected, “One word: Novopay!”
The student’s observations were correct: almost all electronics and computer engineers couldn’t care less about CPEng. But the interjector’s point was valid. In civil engineering, disasters analogous to Novopay are exactly what the CPEng quality mark is supposed to help prevent. So I started to wonder: would it have helped?
First, let’s be clear. The question is not, “could the problem have been avoided if Talent2’s engineers were chartered?” That wouldn’t work. While software is taught as a discipline of engineering in New Zealand’s universities, in practice it has virtually no place in IPENZ or CPEng. The question is much more abstract: “could the problem have been avoided if it was standard practice for software engineers to be chartered?”
When do people care?
Even in civil engineering, enthusiasm for CPEng is hardly universal, but computer engineering is particularly dismissive. In an acknowledgement of this IPENZ includes in its CPEng FAQs the question, “Is CPEng of value to the telecommunication and IT fields?”
I suspect the real distinction is not by discipline, but by the nature of work. Fundamentally, the need for professional quality marks arises because you wouldn’t know any other way whether to trust a practitioner. That is mainly the case when an engineer is hired by laypeople: that is, contracted to provide a service, rather than employed to produce a product.
When you contract someone to design a building, you can’t tell whether their practice is bad before it’s too late. But when you buy a phone, you’re perfectly capable of telling whether the phone works. Even if you can’t, for example with electric power leads, standard for products make more sense than standards for designers. You don’t need to know how good the engineer was; you just need to know that the thing you’re buying works.
This explains why registration matters for accountants, doctors, lawyers and teachers. It explains why enthusiasm for CPEng is greatest in engineering consultancies. It partly explains why in electronics and software, people don’t care.
Why Novopay is relevant
But here’s the thing. Novopay was a contracted job. It was a service, and it was hired by laypeople (the government). Talent2 is not a software company. It’s an HR company that sells software as part of its “payroll solution”. But Novopay, being an extremely large nationwide system, is understandably a contract of its own.
In some civil fields, government policies dictate that consulting engineers must be chartered. This isn’t the practice in software. Civil engineering is more consequential—when bridges fall, people die—but information technology is increasingly a core backbone of society, and even if they don’t always have potential to kill, they certainly impact people’s lives. (Not being paid isn’t fun.)
Furthermore, in reality, software services are often contracted by laypeople. This isn’t always the case: when you buy or download software such as Microsoft Office or Angry Birds, it’s a product: lots of people have the same thing and you can tell whether it works. But businesses often require large-scale IT solutions, and contract companies such as Datacom (who ran the payroll before Talent2) or IBM to do it.
In a reflection of its grounding in civil engineering, IPENZ has not really been scratching its head about the IT disasters of the last year. The Institute of IT Professionals has—and indeed the talk at its AGM is about whether the IT profession should, too, move towards chartered accreditation. Whether CPEng or chartered IT professionals would better cover Novopay is a discussion in itself; I leave it aside but suffice to say that it is also debatable whether software engineering is legitimately “engineering”.
Before an inquiry into Novopay’s failings has been done, we won’t really know whether technical incompetence caused the problems (as opposed to, say, management or training of school payroll staff). So I admit it’s a little premature to speculate on whether CPEng (or chartering IT professionals) would actually have helped avoid Novopay. But I take the interjector’s point. It’s at least plausible to say that, if it was incompetence, then professional certification might have pushed standards high enough to avoid this, or at least helped the Ministry of Education discriminate between those who were up to the task and those who were not.