[trustable-software] Reposted from LinkedIn: Who can we trust for software?

[Paul Sherwood]

On 2016-09-15 09:58, Paul Sherwood wrote:
<snip>
> I'm re-posting this here, in case it triggers any further 
> discussion...

I'm hearing quite supportive comments and interesting private 
discussion around these topics, but as some folks have remarked, not 
everyone is comfortable speaking out in public because of potential 
career impact, NDA restrictions etc.

Ultimately I think this is one of the key barriers that we need to 
overcome - how can we best draw people out to talk about the realities 
of what works and what doesn't, in the private/proprietary worlds that 
deliver hard systems software?

Anyway, quoting one of the comments I received on linked-in

> What changes specifically are you suggesting within the framework of 
> what already exists?
> Aren't there already FAA, FDA certifications for software 
> development?
> Are you suggesting licensing software engineers apart from the 
> corporations they work for?
> Are you suggesting stricter requirements for state or national 
> certifications for corporations/firms that produce software?
> By what authority?
> By what standards?

I'd be very interested in others' perspectives on potential answers. My 
own simplistic view is that, even though I've been involved in 
assessment/recruitment of thousands of engineers over three decades, I'm 
not aware of any certifications that would actually give me confidence 
in a person's knowledge and understanding of how to 
design/develop/deliver software and systems that have no safety net.

This may be a gap in my knowledge of course, or things may be different 
in other regions. But broadly I'm starting from the principle that if 
there was an established and generally accepted solution, I'd have 
already heard about it.

I do value some proof of a technical/CS degree, masters or doctorate - 
but even then, the actual experience may turn out to be great in theory, 
mostly irrelevant in practice.

br
Paul