Fool's Gold, An Introduction

Ryan James Spencer

Developer's are a big target for what I call "fool's gold". It's the hope that a piece of tech can solve all of our problems that keeps us going with the bait of new tech. Solutions tempt despite us knowing better. An experienced software developer realizes that everything has strengths and weaknesses which we call "tradeoffs", but plenty of developer's don't realise this yet or are in denial. This article is an introduction to the concept that plenty software and services are sold as panacea but anything sold as panacea should be considered with caution.

Let us discuss two ends of an argument first. There is the camp of choosing boring tech and running less software. This camp says that cognitive load and operational costs are distractions for teams whose primary focus ought to be the product they are building that makes them profit. By running less software you are curbing the desire to have a bijective mapping between problems to solutions where the relations are each their own distinct solutions. Think about it this way, if you have N many devs and M many distinct solutions for your problems, you have three particular cases to consider

  1. N > M: devs can't be experts except for some subset of the total pool of technology. Ditto (and more importantly) maintenance and operations. More than one dev has to be allocated per tech to make this work (pigeon hole principle).
  2. N = M: every dev can own a particular piece of tech and grow with. Devs may get bored and want to congregate on other pieces of tech. If this happens you wind up with N > M, effectively.
  3. N < M: Devs can congregate around pieces of tech without much fuss. They have freedom to pick what they like most (within a certain degree depending on the delta M-N). Maintenance and operations is bearable as the whole team can participate and not have to spin plates.

Then there is the camp of constant agitation. A company goes under if it's not constantly pushing to optimise for end users and reducing costs. Some even argue you can only pick one of these two optimisations. Enterprises claim to avoid this complication because their golden goose is sitting pretty, but the reality is that any revenue generating organisation has to constantly push themselves into the future to compete. Technology is an enabler, it allows teams to move faster by automating away toil, easing collaboration friction, and a productive team means they can, hopefully, deliver user experiences that delight or at a cost that is beats the competition.

All of this shuffling around comes at the cost of churn and bloat. Companies try to sooth this issue by either by hiring more devs and/or performing lots of migrations. Software rots, which can mean different things to different people, but I see it as the eventual ineffectiveness of a piece of software as improvements are found in competing solutions. That is to say, even though your software may not actually be getting slower, it will definitely feel slower in the context of all neighboring solutions getting faster. This is but one example yet other things like security exploits, support for a particular version of a language or library, and so on all work to disempower your application or system.

The reality is that we cannot simply pick one camp to be part of as professional developers. We are paid to help companies continue to live and be better than they were before we joined, all within the confines of the ethics and legalities we are bound to. We cannot sit still but we can't move too much! Some have suggested things like a novelty budget to support keeping the platform largely stable while pursuing new ways of handling constantly arising issues.

This article isn't meant to attack particular companies or pieces of tech or practices despite those practices that carelessly hold on to the past or endlessly throw it out for the new. The core intent is to encourage critical thinking and consideration of tradeoffs when problem solving. Weigh your options! You cannot make decisions purely by reading and watching, nor can you make a choice by trying everything as there simply isn't the time. Certain devs learn to do cheap experiments at work or at home but this can have the pitfall of comparing a toy project a success that may not handle the scale of an industrial grade application. This sort of scrutiny is part of the weighing process.

If anyone tells you they're going to make you rich, they're getting rich off of you. Take marketing with a grain of salt and throw out the fool's gold!