This year I’ve become increasingly interested in how (and if) the Agile project management methods that are widely used in software can be applied in projects involving hardware.
The essence of Agile, I believe, is that a design or development task is not specified in full detail at the outset (though there must be a clear aim). Instead, a team works closely together to try and find a plausible solution. They meet regularly to review progress and plan the next steps, and to share out the tasks amongst the team members. When they reach a deliverable it is tested, not against a previously-agreed spec but by potential users. Then improvements are made and tested again until the customers are satisfied.
The paradigm for an agile project is a software team designing a new user interface. The task can’t be specified in full detail beforehand but once a plausible solution is found it can easily be built and tested by a group of users. Since there is not much specialisation in software skills tasks can be allocated among the team quite efficiently.
Obviously, you can’t design a car or a plane or a spectrometer in this way. Or can you?
A number of companies have been experimenting with agile methods in hardware projects – or at least in parts of them. Not many such projects have been taken to completion yet but the indications are that there can be significant improvements to speed and efficiency, as well as in staff morale.
But it’s not clear how to recognise tasks that could benefit from an agile approach and those that need the full project planning process. Nor how best to integrate such tasks within a larger project.
R&D Today would be really pleased to hear about and share peoples’ experiences in using Agile in hardware projects. There’s much still to learn.