Sean Barker runs a Technical consultancy on Enterprise Integration and Information Sharing for Products
Some time back long term data sustainment lost the attention of my main funders, so last month - when I retired - I thought I'd have a chance to return my attention to it properly. Unsurprisingly, my immediate attention has been taken up with VAT returns, no longer putting off redoing the kitchen and the magnificent local buzzards circling low overhead.
Attending to the right thing is not straightforward. A colleague once said that a managing director should have nothing on their desk, otherwise that becomes locus of their attention rather than the future of the company. Look at any Computer Aided Design (CAD) demo and your attention will be drawn to a complex 3-D part rotating to show itself off. However, the future in CAD lies in the change from drawings to models, something only hinted at by the dull text boxes at the display's edge and which did not grab your attention. Although early CAD did replaced the drawing board, the modern CAD model replaces the model shop where skilled craftsmen would construct wooden mock-ups, For example, Airbus created a full scale mock-up of the A340 wing, said to cost a million pounds, but saving several times that through design improvements. Now model-based CAD saves the cost of building those physical mock-ups.
Because the shape of an object is easily grasped, CAD provides a paradigm for model-based engineering. However, model-based approaches are increasingly important for other disciplines. For example, Systems Engineering is taking up model building using SYSML. SYSML is introduced as a series of diagrams, each showing a different aspect of system design, such as structure, behaviour or timing. But these diagrams are simply slices through a complex, multi-dimensional model that allows the development of system level prototypes.
Model-based engineering has profound implications for long term data sustainment. It is not possible to simply take a snapshot of the data that records the model, as that data means nothing without the software that interprets it. The data for a square may be the co-ordinates of its four corners, but without the software to join the dots, it remains just four dots. The mechanical CAD world has spent a lot of time and effort making sure that its data is linked to well defined formulae so that the CAD software can reproduce the model shape exactly. Some five or six years ago, this investment was quoted as $250 million dollars, and there has been more work since. They also spent - and continue to spend - a great deal more time and effort making sure that the CAD software suppliers implement the specification in the same way so those models are identical when they exchange models or, equivalently, update software.
The immediate savings from CAD models gains the attention of executives, whose attention is mostly focused on making a profit. However an executive will change jobs every few years, so the importance of long term sustainment - which improves the bottom line of their successors - has a harder job gaining their attention.
And there is a lot in engineering to attract their attention. I have just finished a short book, Aircraft as Systems of Systems: a Business Process Perspective [1], a rapid romp through aircraft engineering from designing a simple component to integrated vehicle-health-monitoring-maintenance-logistics-depot-and-fleet-management. One of its objectives - prompted by a discussion with a group of chief engineers - is to explain to executives why engineering is not a simple, linear process - why an engineer's GANTT chart is more shopping list than a promise that they will do something by a given date for a given cost. And much as I like to talk about long term data sustainment, and important though the topic is, I found that I only had room for two pages out of a hundred to cover both the business need and LOTAR [2], the major aerospace project to standardise long term sustainment across the industry.
Why is LOTAR significant? Why is sustaining of models different to archiving documents? And what else is needed to maintain aircraft designs over seventy years? These are questions to occupy my attention once the kitchen is redone and the VAT returns are in. I feel another book coming on (except when distracted by the buzzards). And thank you for your attention.
References:
1] Aircraft as a Systems of Systems: A Business Process Approach Sean Barker, SAE International, Warrendale PA ISBN-Print 978-0-7680-9402-2 - other formats available. (see also books.SAE.Org)
2] LOTAR, “LOng Term Archiving and Retrieval - LOTAR,” accessed December 14, 2017, http://www.lotar-international.org/.