Recently the old adage measure twice, cut once has been very relevant to a personal project, so I thought it would be worth while writing a few notes on what may to most of us appears the flippin’ obvious, but can still catch us all out every now and then.
What could go wrong?
Currently we’re having some building work done on our house, in fact so current I’m finding it difficult to write this over the noise. At some point in the distant past the main contractor came to measure up some existing steelwork in order to order a new precast concrete slab which will rest on top. As far as I can recall:
He jotted some dimensions down in a little book and we were all happy.
A few weeks ago the slab arrived and only when they tried to put it in place did they realise it was too deep. It turns out they had given the wrong dimension to the people casting the slab, It should have been the total depth required minus the depth of frame e.g. 150mm – 100mm = 50mm, but the slab had been cast 100mm deep. These might not be the exact dimensions but you get the idea. By the way, that’s about 6″ – 4″ = 2″ if you’re still using imperial.
(There’s another question we could ask about whether a 50mm deep slab would be strong enough but that’s a different issue)
What went wrong?
Well we don’t know exactly, but i wager that it wasn’t clear what the various dimensions recorded were, and when they came to order the slab, they got confused and used the wrong ones. But in any case, the mental picture they had of how the parts would integrate and what they needed to do was different form the reality.
What’s the impact?
On us: At the moment just frustration and some inconvenience. Thankfully this bit of the build isn’t on the critical path and there are no other scheduled jobs waiting on it.
On the main contractor: Time and Cost, they made the mistake so they have to put things right.
What should they have done?
Draw a little sketch! I’m not just saying this with hindsight. I’m constantly amazed how tradsmen try to explain things to me using nothing more than the spoken word and a few wavy arm movements. If you think that’s reasonable, what if I tell you I’ve had the same experience with Chartered Engineers building railways, aircraft and ships! A quick sketch, annotated with the measurements, would have put the data in context and reduced the risk of misinterpretation. Additionally, other people (including me) could have double checked it and any mistakes or misunderstandings likely discovered.
By the way, I’m not suggested railways, aircraft and ships get built without any drawings, just that there are lots of peripheral (but still important) conversations where the drawings, or even informal sketches, aren’t used.
Why didn’t they do a sketch?
They either didn’t anticipate the risk or decided the consequences weren’t serious enough to mitigate them. Nobody likes to do more than they have to do and drawing a sketch didn’t seem necessary. Although note; I’m only talking about a hand drawn sketch here, you could make an argument that they should be creating formal engineering drawings, but producing a quick sketch seams like it would have greatly reduced the risk for very little extra effort.
Why is it so difficult?
Because we can’t easily identify the tipping point from simple to complex and/or from low risk to high risk. Imagine you regularly run 5K (you can do your own imperial conversion here) and someone suggests you do a 6K run. You’ll probably decided all you need is a few more carbs, water and a little extra willpower. Now imagine someone asked you to run a marathon (just over 42K / 26 miles). You’d be well aware that you’d need to build up to it with a careful considered training plan, a healthy diet and some new running shoes. But at what distance is simply “trying harder on the day” not going to be enough – 7K, 10K, 20K, 30K? Where’s the tipping point?
All’s well that ends well then?
Faced with the existing steel work and a newly cast concerete slab that are incompatible, the main contractor decided to commission new steelwork. Along came a steel fabricator to inspect the job. As far as I can recall:
He jotted some dimensions down in a little book and we were all happy.
You’re probably way ahead of me here…
Today the new steelwork arrived and we discovered it was exactly the same dimensions (depth) as the old steelwork i.e. no improvement.
This project is starting to make me look like an incompetent client!
What has all this got to do with Model-Based System Engineering?
Well, the ‘sketch’ I keep mentioning is a simple model, and not only that it’s an appropriate model for the task at hand. This sketch (or series of sketches) should inform the viewpoints of all the stakeholders – the main contractor, the steel fabricator, the concrete slab maker and the client (me). If it seams common senses to have at least a sketch for some home renovations, then what sort of model and views do you think you need if you’re building a railway, aircraft or a ship? And how do you know if their form is appropriate?
In MBSE the various views (diagrams) are all linked to a common underlying model. We can interrogate this model to check that it is is complete and consistent rather than having to compare and contrast all the different views. Also, if we need to make a change we only have to make it once and in one place.
Still convinced conversation and wavy arms are the way forward? Or is it better to:
Model, Validate (measure twice) then Construct (cut once)
P.S. I’m going to do my own sketches for my next meeting with the main contractor.
