My current blog is epistem.ink. This one is here just for archival purposes.
Imaginary Problems Are the Root of Bad Software
There are many factors which can be a catalyst for bad software: from the tools being used, to team communication, to the personal stake developers have in its success, to the testing methodology.
I propose that there is one problem chief among them, an impetus for bad software from which almost all others take root: imaginary problems.
Most complicated or broken software is not designed to be overly complex or dysfunctional. It’s just designed to do something other than its intended purpose.
Let’s say you’re a podcast host who wants a custom website where you can sell your promotional products, make advertising money without a third party cutting in, and, most importantly, deliver podcasts, videos, and blogs to your audience.
The requirements for your little web-app might look something like this:
- Fast load time in North America, with real-time podcast streaming and downloads
- Doesn’t crash or freeze in the first 15 minutes for 99.99 percent of users, preferably never crashes or freezes
- Integrates well with Google Adwords and maybe some other third-party ad providers as well, if there’s time
- Dynamically links to the latest products in my Zazzle shop and, if possible, gives recommendations to users based on the content they’ve consumed
- Integrates with Facebook live player. If it’s easy to create an alternative solution for streaming that doesn’t require Facebook, even better
You give these specs to a team of contractors, and you chat about them a bit. It seems that everyone is on the same page. Yet, when they return with the Minimum Viable Product two months later, your face turns red. You’ve just wasted $15,000 on a piece of garbage; you want your money back.
The first time you open the app, the screen freezes. You ask how to select what kind of ads should be allowed to run on the site and are pointed to an ugly, hard-to-understand custom user interface (UI). Half the links to your merchandise on Zazzle are broken or missing images, and the Facebook livestream is laggy!
But the development team is confused at your anger — rightfully so, from their point of view — because they’ve gone to hell and back for you.
They’ve put their heart and soul into creating this app, and it has some amazing features:
- A state of the art recommendation system
- An algorithm generating the transcript of all your streams, in real time
- Your front page loads in sub 200ms times all over the world
- A streaming protocol and client build almost from scratch, in case you don’t want to rely on Facebook live
- A service that allows you to easily integrate over 20 ad exchanges
The problem is that you thought you requested a core product with a couple of extra features, if they were easy enough to implement. Meanwhile, the dev team heard something else. They heard about some exciting challenges they could tackle… and a slew of boring, basic features they couldn’t be bothered to test properly or care about.
Even worse, you didn’t communicate directly with the devs — you communicated through a game of Telephone. You spoke to a sales guy, who held a meeting with some middle management chap, who wrote some business specs and gave those to a PM, who wrote some technical specs and gave those to a team lead or architect, who then, at last, began to design the product with his team — each one of them putting a bit of his own twist on it along the way.
Imaginary problems are often more fun to solve than real ones. Extremely intelligent people play competitive games, construct and solve math problems, and write books that aim to answer abstract questions about the human condition, all of them for free. A mediocre programmer, however, will probably charge you a fair amount to build a simple Android app. That’s not because mediocre programmers are harder to find than geniuses, but because the former activities are all fun, while the latter can be quite boring.
Most programmers want to get paid and have fun at the same time. Of course, the definition of “fun” is different for everyone, but for many engineers, it boils down to tackling interesting and challenging problems that are within the realm of solvability.
Give a somewhat intelligent person too many boring tasks that are impossible to automate and you will eventually drive him mad. The human brain however, after billions of years of evolution, is quite talented at keeping its sanity. Much like victims of childhood hardship or abuse can find escape in fantasy books, victims of enterprise programming or freelance web development can find their escape in solving imaginary problems.
The amount of imaginary problems a software engineer can create for themselves is a function of their imagination and of the difficulty of the real problems they’re supposed to solve.
It should be noted that this issue isn’t unique to developers. Management, sales, HR, support, legal, and even accounting departments have their own unique ways of creating imaginary problems. They try to involve themselves too much in a decision, when their presence at a meeting is just a formality or wasn’t requested at all. They overemphasize a minute problem that is related to their role, or hire teams much larger than necessary to illustrate their importance.
When problems are dumb, intelligent individuals will find a way of coping.
But imaginary problems aren’t just the result of bored developers. They’re also the result of long chains of communication.
When I first began taking on freelance clients, I couldn’t afford to be particular. This means I’ve had email chains lasting for over a 100 exchanges, discussing insignificant details about internal MVPs. I’ve had people change every single requirement on a project within the span of a week. I’ve had clients ask questions such as “Could this be ICO-ed?” or “Can we add some A.I. in here?”
Granted, most clients are savvier than that, but, even still, they often lack a bit of the knowledge necessary to articulate or construct some of their requirements. That is fine, as part of my job as “the computer guy” is to help people figure out what they do and don’t need based on their use cases. But it can become much harder to determine what’s needed when there are a few layers between you and the client.
Requirements get changed because someone either misunderstood an intention or because someone was trying to cope with that aforementioned boredom
Most companies like having a sales guy who pitches potential customers, negotiates prices, and outlines possible features. They also have a people person to discuss more in-depth requirements and details with the customer, usually another sales guy, but with a slightly different title. Then there’s the internal chain of command, various levels of management, and possibly some hierarchy, within the technical team.
When a list of client requirements goes through so many people, even if those people have the best of intentions, some things will inevitably get lost in translation. Sometimes that change happens because the original requirement made no sense, or sometimes requirements need to be redefined. The sales guy might have told the client, “for only 39,999 extra we can do this on the Blockchain.” But that leaves everyone who encounters the requirements down the line wondering what the definition is of “doing it on the Blockchain.”
More often than not, requirements get changed because someone either misunderstood an intention or because someone was trying to cope with that aforementioned boredom, trying to make his job or the work of his team more interesting and impressive.
Through all of this, the original requirements — the real problems that have to be solved — get lost. They are replaced with imaginary problems and with voids, and you’ve got plenty of people ready and willing to fill those voids with their own imaginary problems, because the problems they have to solve are boring, and filling the voids gives them a way of coping.
Overcomplexity and natural selection
There can often be an even darker reason for the existence of imaginary problems: problems can help a team or a company grow, and can even become an integral part of its function.
“People who are bred, selected, and compensated to find complicated solutions do not have an incentive to implement simplified ones.”
— Nassim Nicholas Taleb
Have you ever heard about those three web engineers who figured out that secure online banking is actually quite an easy problem to solve? They developed some flawless banking software from scratch, using a functional design methodology and memory safe languages, then started migrating major banks to their amazing infrastructure.
Probably you haven’t heard of them, because they don’t exist. There are, however, plenty of teams of thousands of developers, who are unable to grasp simple concepts such “rollbacks,” perpetually creating banking software.
The storage and transfer of numbers is not a particularly hard problem. Indexing the whole content of the internet and providing relevant results to natural language queries, in sub second times, is a hard problem. But just a few smart guys managed to solve that problem.
The persistent problem for online banking is that the banking ecosystem has become really good at preserving its own money-grabbing hierarchy. Its leaders are corrupt leeches who prey on society — but the leaders in an organization are just a symptom of its members.
I wouldn’t suggest that most underling workers for banks are evil or malicious in any way. Far from it. They are usually friendly lads, working to provide food, shelter, and an education for their families. But their chief incentive is not to fix the banking software, it is to stay employed. Losing your job in today’s economy is no joking matter for some; in the banking industry, a big mouth or too much initiative is an easy way find yourself in front of a disciplinary committee.
So banking systems remain the same — not because the systems are efficient, but because of inertia. This inertia comes in the form of working on imaginary problems in order to avoid fixing real problems — real problems which, once pointed out, would threaten the jobs of other people. To focus on these real problems could lead to getting fired, or, in the case of some particularly nasty “institutions” like Goldman Sachs, getting a few life-ruining brown envelopes sent to a few FBI officers and prompting a strange suicide.
“It is difficult to get a man to understand something, when his salary depends upon his not understanding it!”
— Upton Sinclair
The C-suite ignores the fact that their upper management workers spend 90 percent of their time on “client meetings” that involve tropical islands and million-dollar budgets for “other expenses.” Upper management, in return, turns a blind eye to corruption in C-suite.
Because middle management encourages them to live in their Wolf of Wall Street fantasies, upper management ignores middle managers who buy eccentric offices and hire themselves three secretaries and a dozen interns.
Because line management doesn’t complain about their dictatorial power fantasies, middle management ignores the fact that line managers, instead of cutting costs, spend their time working on PowerPoint presentations about “Improving our Agile Methodology.”
Because the team leaders don’t seem to notice the fact that their superiors can’t even use Excel properly and only hit the office every few weeks, line managers ignore the team leaders and architects talking about “next generation interfacing between our systems using JRPC and microserviceization using Hibernate and Spring” when they should be getting those bloody Mysql queries to take less than a day.
It’s a vicious cycle of solving imaginary problems, from the CEO who doesn’t realize that stealing another 30 million won’t make his dad love him to the user-experience intern who doesn’t realize that redesigning the “submit” button using Angular-Material-Bootstrap 19.13.5 won’t make the fact that they store passwords in plain text (and use them as part of the auth cookie) go away.
But everyone needs to keep solving the imaginary problems, because if they stop creating and solving these problems, if they start focusing on the real problems, they might realize the whole system is broken. They might realize Debra has been sitting in that corner, staring at uptime graphs of the internal server farm for 10 years, despite the fact that the company moved to AWS five years ago. They might realize 99 percent of their job is to perpetuate the existence of someone else’s job. And that’s a hard realization to digest—impossible for most, I dare say. So, instead, most find a way of coping.
If you enjoyed this article you may also like:
Published on: 2019-04-29