Just a Five Minute Job - A Home Improvement Metaphor for Tech Debt

One of the most interesting panel discussions that I participated in during Agile + DevOps 2018 centered around Technical Debt and what we can do about it. We found ourselves using a metaphor of home improvement projects a few times - especially the idea that you don't always know how big the task is when you start. With that in mind, I'd like to present a short recap of a recent "five minute" easy home improvement fix I found myself in.

The Problem:My girls share a room which contains one of the air returns for the heater. The grate covering this duct was no longer staying in the wall. This had led to at least one feline excursion into the ductwork, and now required that a trashcan or other large object block the grate to keep the cats away.

The Estimate:
This isn't a difficult fix. All I needed to do was grab my drill, zip in a couple of drywall anchors, drive the screws back into the wall, and done. Five minutes. Max.

The Reality:
First off, the grate cover is the wrong size for the hole. What I really should have done was get a new cover, or enlarge the hole, but one of those things would require spending extra money, and the other would require measuring and cutting additional bits of drywall. Both were out of scope.

After grate removal. Can't expand the hole to the left because there's a wall stud there.

My first fix was exactly as specified in the estimate. That worked fine, right up until I realized that the screws were now too short to properly open the drywall anchors. Okay, I'll get some new screws.

Then the screw on the right drove the drywall anchor straight through the sheet of drywall and into the five inch gap in the wall that I didn't realize was there.

Feline help from Romana the Time Kitty did not decrease the task duration.

No problem, I can go grab a couple of 2x4 ends, trim them up, fill the gap between the walls, and drive a 3" wood screw in there.

Which I did. Total time, between trimming 2x4 scraps, swapping out screws, removing busted drywall anchors, and oh yes, removing a helpful cat: forty-five minutes. That's 9x longer than my original estimate.

Yeah, this drywall anchor isn't going to work. I don't know why the original grate was set up with that screw so close to the edge of the hole.

To bring this full circle, it's the same issue with technical debt. You don't know what you don't know. Even so, technical debt must be handled. If I'd known I needed to trim up some lumber to fill a gap, etc, my original estimate would have been a lot more realistic. As it is, sometimes you just have to pull a task off the board, set aside 10x whatever your time estimate is, and hope this task is actually what it seems. Ever time, the set of unknown unknowns in your infrastructure will (hopefully) decrease. Every time I dig into a fix in this house I learn a bit more about it, and shrink my list of unknowns a little more. Eventually I'll know this place as well as the builder. I hope.

Introspection - Doing Things Half-Way

When my family moved to your new location, we got a rental house in one of the city's historical districts. Built sometime around 1904, it's a classic home with a lot of, let's call it character.

The problem with character though, is that it takes a lot of work to maintain properly. And while this is a rental, it's pretty clear that not only has it not always been maintained well, but sometimes people came in with some big ideas, and didn't really follow through.

For example, the Second floor used to have carpet. How do I know? Because the baseboards are a half-inch or more off the floor, set up for carpet that someone ripped out at some point to expose the nice hardwood floorboards underneath.

Of course those floorboards are also pretty scuffed and scratched from decades of use, to say nothing of the carpet laying on top of it for who knows how long. They really need a major resurfacing (if they can be salvaged at all), and the baseboards need to be redone and lowered to eliminate the gaps in the flooring. Then you'd need to repaint.

But for the person who ripped out the carpet, that's where they stopped. Why did they leave the job half-done? Who knows. Maybe they ran out of money, or time, or interest. Maybe all three. I may never know.

What I do know, is that living in a constant reminder of what half-done looks like has been motivational for me. It's a problem I've been guilty of far too often myself. Major hobby projects get half finished, and left, to eventually be either redone, or tossed, depending on the project in question. I should put the coffee down, I'm not a closer.

But I'm getting better. When I think about starting a new game, adding another project to my overloaded TODO list before I remove two, I walk up the stairs, and take a long, hard look and the floor, and the gap between the baseboards. And I remind myself of what not finishing what I've started looks like.

It Seemed Like a Good Idea at the Time - Turret Fighters

I'll (hopefully) pick up my Night Fighters series proper next week. This week, however, I'm going to focus on one of the weird sideshows from the beginning of WWII aviation, and an aircraft that was pressed into service as a night fighter for a while.

It must have seemed like a good idea at the time. In the late 1930s, light- and medium-bombers were outpacing the late model biplane fighters of the day, and it was generally believed that speed, combined with massed turret fire, would allow bombers to attack enemy targets without need for fighter cover. Of course, this thought didn't stop British Fighter Command from asking something along the lines of "Well and good for the enemy, but we can't have our fighters getting shot to pieces. How are we going to intercept the enemy bombers?"

One of the answers went something along the lines of "Well, if turrets are good for bombers, then they'll be great on fighters! Do that!" And so, approximately, were born the Boulton Paul Defiant for the RAF, and Blackburn Roc for the Royal Navy.

Unlike the Defiant, the Roc also carried provision for
underwing weapons, and intended for ground-attack
duties as well.

Blackburn Roc. Ugly in a uniquely British way.

Still not a good fighter, but Merlin engines make everything
at least somewhat better.

Boulton-Paul Defiant

On the surface, it was a concept that solved a couple  of problems. The gunner freed up the pilot to focus on interception and flying duties, while the turret allowed for attacking enemy bombers from underneath, where fewer guns could be brought to bear.

Unfortunately, in actual combat, the turret fighter concept proved cripplingly vulnerable to single-seat enemy fighters. The extra weight of the turret and lack of forward armament put the turret fighters at a serious disadvantage against lighter, faster, more maneuverable opponents. Defiants did have some early success, primarily against German pilots who mistook them for Hawker Hurricanes and initiated attacks from the rear. Lessons subsequently learned, squadrons equipped with Defiants soon began suffering significant loses.

As a stop-gap measure until the Bristol Beaufighter became sufficiently operational, Defiants were re-assigned to night fighter duties. They saw some successes in this role, equipped with some of the early airborne radars, but were soon replaced by the Beaufighter, and eventually the Mosquito. Surviving Defiants were re-assigned to serve as trainers and target tugs,

Like any number of military innovations throughout history, the turret fighter was born from ideas that seemed good on paper, but failed in reality. The sole surviving Defiant is on display at the Royal Air Force Museum in London. No Rocs appear to have survived.