Fail fast

Hey everybody what is going on? Welecome to the Roger in Sabbatical blog, where I talk about not working, not doing projects, but also work, projects, and a good deal about kayaking.

Today was make-a-crossbow day. I know that might sound like a normal Tuesday to most of you but to me it's an unusual project, and as such packed full of unusual skills that I have yet to fail at. In preparation I roughed out the furniture in advance using some spade bits, a half-inch wood chisel and a whole lot of swear words so today was really the mechanical bit and assembly.

Or so I thought.

Right off the bat let's provide some context. I live in the beautiful city of Cambridge. And in Cambridge can you buy polyester string? You know, cheap string. cheap, plasticy, synthetic string. You know the kind. It's usually barbie pink or bright blue but if you are really lucky there will be a sandy coloured variant that looks like its from GetDate().currentDecade. 
"But this is Cambridge" We've got organic natural cotton string. Hand fed free range balls of puffy white expensive string. Who sells cheap synthetic string? bloody nobody.
So after hiking all morning through town until I found some, and then a quick march to buy some paracord that I'll need. Plus breakfast pancakes and coffee, and that takes it to technically after noon, so as you can tell the disaster that follows isn't going to suffer from having too much time.

And Cambridge is nice, and the unsightly kayak wounds on my thumbs are scabbing over nicely and will scar up just so, and I'll be out on the river again in no time. 

But anyway. I was most of the way through forming the prod of my crossbow, right before learning how to twist a Flemish style bowstring, and figured I could shape the thing a little more with a cold-cut chisel. This isn't a wise decision, as the material stress caused a fracture just beyond the recurve that will "fuck it right up" when put under stress. So a couple of hours wasted forming the thing that I've now wrecked through a bad choice of tools. 
However, all is not lost, as I proceeded to stretch the thing in my hands to see how much flex and restitution it had and discovered a flaw in my curve that ultimately is a weakness that would have prevented it from ever being functional.

What I suffered from here is a lack of early testing. I'd not QA'd each step of fabrication and was relying on user-acceptance-testing (UAT) to tell me that the system as a whole doesn't work after then darn thing was built.
Fail-Fast is the (software) engineering principal that you should expose and uncover defects as soon as possible in the development cycle to prevent additional work being wasted. Essentially prevent you throwing good money after bad and allowing you to take one step back before your two steps forward.

To put a civil engineering slant on the proceedings, which is a frequent source of metaphor in software, I'm building in the garden and my visualisation wasn't as good as digging a foundation and looking it it in place. Stopping while it's just a square hole and redesigning now is a lot cheaper than building it and moving walls. The civil engineering metaphor only works with software so far, as one is an architecture task and the other is a construction one.

With the foundations its not too early to step back and redesign. The poor crossbow prod has to be deleted and remade, but it was a cheap enough process that throwing one away in the guise of education is fair enough. Tune in tomorrow to see if the forgehouse survives the redesign process.