Spare a Thought for the Next Guy
By Ed Johnson
I just had a new ISDN phone line installed at my house. It was an unexpectedly entertaining event, and provided the opportunity for some reflection on the similarity between the problems faced by software developers and those in other occupations.
The installation was performed by a technician who introduced himself in a Yugoslavian brogue as “Ranko.” Ranko looked at the existing phone outlets, declared that it would be a straight forward job and should take 30 - 45 minutes.
I sat down to read a book and let Ranko go about his work.
Things seemed to be going well for him until 15 minutes into the procedure when I heard some veiled mutterings coming from the kitchen. Putting my book down to listen more carefully, I heard Ranko talking to himself in angry tones - “What have they done? What have they done to poor Ranko?” (he had an unusual habit of referring to himself in the third person).
Curious, I sauntered into the kitchen on the pretence of making myself a cup of coffee.
I found Ranko still muttering away and staring in angry disbelief at the display of some instrument. Before him were a half dozen cables spewing out of the wall like so many distended plastic intestines set loose from the house’s abdominal cavity. Ranko asked if he could get up into the ceiling cavity of the house, and I assented – pointing him in the direction of the access cover. He strode outside to his van and reappeared in my front door a few moments later with a step ladder under one arm. I returned to my reading while he pounded around above me. Shortly I heard a few exclamations of “Bloody hell!” followed by more thumping. After a brief pause, there came a series of “You bloody idiots!” / “Bastards!” two-shots in rapid succession, punctuated by some unnecessarily loud pounding of feet upon the ceiling joists. Underneath, I listened with growing amusement, choking back laughter with one hand over my mouth.
For the next 10 minutes or so I was lost in my reading, and looked up in surprise to find Ranko standing in front of me looking slightly disheveled but rather proud of himself.
“I have found the problem” he declared proudly, and proceeded to explain. It appeared the previous residents of the house had self-installed one of the telephone extensions in my house. Rather than daisy-chain the additional outlet on from another outlet, they had simply spliced into the phone line up in the ceiling cavity and run cabling from the splice point to the new outlet. This was easier for them than daisy-chaining, as it halved the number of times they had to run a phone cable through a wall cavity.
But for future technicians, it meant that any wiring changes of the type Ranko was attempting would necessitate access to the ceiling cavity where the splice-point was located. If done in daisy-chain style, as is regular practice amongst phone technicians, the wiring changes could’ve been done without having to ascend into the crawl space above. The job took nearly twice as long as what Ranko initially estimated, because he had the unexpected tasks of diagnosing the problem with the existing installation, determining the location of the splice and then working around it.
Sound familiar?
Ranko has experienced the same problem that maintenance programmers face every day. We estimate the duration of a maintenance task based on some assumptions about the nature of the artifact we will be altering. We begin the maintenance task, only to find that those assumptions don’t hold, due to some unexpected shortcuts taken by those that came before us. Then we have to develop an understanding of those shortcuts, before we can perform our maintenance task. And if we chose to work around the shortcut rather than fix it, future maintenance programmers will have the same problems. And so a short- term expediency made by a programmer long ago becomes the burden of every programmer that follows.
And the very need to make assumptions at all stems from the absence of any information about the morphology of the existing system. Those that hack into the phone line are not of the nature to document their efforts, nor to keep existing documentation up to date.
So next time you’re under dead line pressure and have your fingers poised above the keyboard ready to take a shortcut – spare a thought for the next guy who will have to deal with that shortcut. He might be you.