I wouldn’t start from here, if I was going there…

So concludes the punchline of a very old joke. At its core is the observation that where you start a journey can dramatically affect your ease of passage. But how often when starting a new technology development do we start with the familiar tool kit and methods, rather than assessing what’s most appropriate for that particular innovation?

Despite being first described over 15 years ago, the Internet of Things still has a distinct whiff of hype about it. The Gartner Hype Cycle 2014 has IoT at the “Peak of Inflated Expectation”, with 5-10 years before reaching a productive plateau. So where does that leave embedded software engineers trying to decide how best to develop IoT devices?

Probably confused.

There are well established approaches to developing embedded software in different industries. These reflect the underlying assumptions, technology constraints, and regulatory environments within each industry. However, the IoT doesn’t fit neatly into any pre-existing industry sector. In fact it spans them almost by definition. So which set of protocols and design decisions do you follow?

What is needed is a map of the Embedded Software technology landscape.

A (mostly) humorous map of Linux distributions [there isn’t an existing one, humorous or otherwise, for Embedded Software]

Something that provides the engineer considering a novel technology with a clear framework to look at a specific innovation and plot a course through to successful deployment. It is likely to require moving the starting point from ‘here’ to ‘over there’. It is likely to involve tools, techniques and methods that aren’t familiar. However, there are experts out there (many in the Bristol & Bath area), accepted ways of working, even standards and verification techniques to be sure that your code ‘does what it says on the tin’.

Its something we’re working on as a Special Interest Group within HBB. It’s early days, and it’s expressly ‘bottom up’. It’ll be untidy at first, it may be untidy at the end. But we plan for it to be useful.

If you’d like to be involved, come along to the Intelligent Testing Conference, we have a workshop on the next iteration just after the main conference. If you can’t make next Thursday, please drop me a line or consider joining High Tech Bristol and Bath CIC to help set the direction and final output.