Software Failure: Why Problems Abound and Multiply Like Gerbils

March 19, 2024

This essay is the work of a dumb dinobaby. No smart software required.

I read “Why Software Projects Fail” after a lunch at which crappy software and lousy products were a source of amusement. The door fell off what?

What’s interesting about the article is that it contains a number of statements which resonated with me. I recommend the article, but I want to highlight several statements from the essay. These do a good job of explaining why small and large projects go off the rails. Within the last 12 months I witnessed one project get tangled in solving a problem that existed 15 years ago. Today not so much. The team crafted the equivalent of a Greek Corinthian helmet from the 8th century BCE. Another project infused with AI and vision of providing a “new” approach to security wobble between and among a telecommunications approach, an email approach, and an SMS approach with bells and whistles only a science fiction fan would appreciate. Both of these examples obtained funding; neither set out to build a clown car. What happened? That’s where “Why Projects Fail?” becomes relevant.

Thanks, MSFT Copilot. You have that MVP idea nailed with the recent Windows 11 update, don’t you. Good enough I suppose.

Let’s look at three passages from the essay, shall we?

Belief in One’s Abilities or I Got an Also-Participated Ribbon in Middle School

Here’s the statement from the essay:

One of the things that I’ve noticed is that developers often underestimate not just the complexity of tasks, but there’s a general overconfidence in their abilities, not limited by programming:

Overconfidence in their coding skills.

Overconfidence in learning new technologies.

Overconfidence in our abstractions.

Overconfidence in external dependencies, e.g., third-party services or some open-source library.

My comment: Spot on. Those ribbons built confidence, but they mean nothing.

Open Source Is Great Unless It Has Been Screwed Up, Become a Malware Delivery Vehicle, or Just Does Not Work

Here’s the statement from the essay:

… anything you do not directly control is a risk of hidden complexity. The assumption that third-party services, libraries, packages, or APIs will work as expected without bugs is a common oversight.

My view is that “complexity” is kicked around as if everyone held a shared understanding of the term. There are quite different types of complexity. For software, there is the complexity of a simple process created in Assembler but essentially impenetrable to a 20-something from a whiz-bang computer science school. There is the complexity of software built over time by attention deficit driven people who do not communicate, coordinate, or care what others are doing, will do, or have done. Toss in the complexity of indifferent, uninformed, or uninterested “management,” and you get an exciting environment in which to “fix up” software. The cherry on top of this confection is that quite a bit of software is assumed to be good. Ho ho ho.

The Real World: It Exists and Permeates

I liked this statement:

Technology that seemed straightforward refuses to cooperate, external competitors launch similar ideas, key partners back out, and internal business stakeholders focus more on the projects that include AI in their name. Things slow down, and as months turn into years, enthusiasm wanes. Then the snowball continues — key members leave, and new people join, each departure a slight shift in direction. New tech lead steps in, eager to leave their mark, steering the project further from its original course. At this point, nobody knows where the project is headed, and nobody wants to admit the project has failed. It’s a tough spot, especially when everyone’s playing it safe, avoiding the embarrassment or penalties of admitting failure.

What are the signals that trouble looms? A fumbled ball at the Google or the Apple car that isn’t can be blinking lights. Staff who go rogue on social media or find an ambulance chasing honed law firm can catch some individual’s attention.

The write up contains other helpful observations. Will people take heed? Are you kidding me? Excellence costs money, requires informed judgment, and expertise. Who has time for this with AI calendars, the demands of TikTok and Instagram, and hitting the local coffee shop?

Stephen E Arnold, March 19, 2024

Written by Stephen E. Arnold · Filed Under Business process, Management, News

Comments

Comments are closed.

Search the site
Subscribe to Beyond Search
Feature archive
News archive

Stephen E. Arnold monitors search, content processing, text mining and related topics from his high-tech nerve center in rural Kentucky. He tries to winnow the goose feathers from the giblets. He works with colleagues worldwide to make this Web log useful to those who want to go "beyond search". Contact him at sa [at] arnoldit.com. His Web site with additional information about search is arnoldit.com.