In today’s fast-evolving tech world, the pull towards new tools and technologies is strong. Sometimes, it can be hard to decide what to get involved with and which ones to leave behind.
Let's explore this a bit further!
Let's check out some dilemmas, biases and misconceptions that influence the way we look at and choose tools.
Developers like me tend to chase the latest trends. This makes sense. Learning new tools offer fresh solutions, new perspectives and in general broaden our arsenal of technologies that we have at our disposal.
We often see new tools can have some advantages, such as being more innovative, offering better integration with other new technologies and in general giving you a competitive edge. And, yes, they also look good on LinkedIn...
On the contrary, older tools can be perceived as outdated, unmaintained, less cool and more boring. In short, something we want to get rid of.
But it this really a smart move?
Let's look at some things that influence how we choose, perceive and interact with tools.
When you start with a new tool, you probably first take a look at its website. Here, typically, you will find short "getting started" tutorials to follow. These are great to get started.
However, they generally show the "happy path", a use case that is simple, straightforward, and will most likely work on the first try. You need to be careful here, as this may hide real-world complexities that occur in a real project, giving a false sense of ease and confidence.
Proofs of concept projects are essential to explore whether certain technologies should be considered for a specific problems. They help in assessing how a tool fits into your existing tech stack and problem scope.
The problem here lies in the cost and time involved are real considerations, especially because these PoC implementations are just a test that will be thrown away again. Also, problems can still occur in the future since a PoC generally does not consider all of your existing technologies, pipelines, APIs, databases, environments etc.
Another pitfall is the so-called "sunk cost fallacy". This basically says that once you have invested time, money or other resources into learning or buying a new tool, you might feel the obligation to use it.
Once invested, it can be hard to abandon a project, leading to irrational decision-making to justify prior effort.
Abraham Maslow, a psychologist, famously said that
"If the only tool you have is a hammer, everything looks like a nail."
This expresses the danger of wanting to use your favourite tool for everything, even it is not the right one for the job.
Using niche tools that only one team member knows can create bottlenecks and knowledge silos. If the person in the know is run over by a tractor, all that knowledge is lost if it is not properly documented.
Also, if you don't plan on getting run over by a tractor, you need to decide whether you want to be the single person in the team with tool specific knowledge. In this case, all questions will come to you, leading to time loss and context switching on your part.
To overcome this, you would need to train the whole team and create documentation. This is a considerable effort and time investment.
Humans prefer known options and often avoid uncertainty. When finding a new tool and you present it, it often happens that the people in the audience are sceptical at first. So even to convince your fellow team members can be hard work.
Choosing the right tool requires understanding the "known unknowns" and the "unknown unknowns."
Known unknowns are things that you are aware of, but you don't know exactly when they will occur.
An example could be
“If we have 10,000 requests per second, I don’t know what will happen.”
Here, it is clear that the system will break at one point.
Unknown unknowns are completely surprising things that can happen:
“I had no idea enabling this option would cause memory leaks…”
Things like that can break your anticipated solution completely because there was no way you could have prevented that beforehand.
Dan McKinley explains in Choose Boring Technology, that
"For new technologies, [the unknown unknowns] are significantly greater."
This is a major reason why old tools can be way more reliable.
Or, to put it another way, you know under which circumstances they become unreliable and can work around it in advance.
There are some more things to consider when looking at old tools.
When you've been using something for years, you know all its quirks, shortcuts and internal workings. That muscle memory is worth something, and it often means you and your team can work faster with an older tool.
Older tools tend to play nice with other tools, especially when they follow proven standards. Sticking with tried-and-true options can save you from compatibility headaches or workarounds. This highly depends on your environment and other technologies you use, though.
If you don't have to school the whole team, buy a new solution or burn resources by working around unknown unknowns of new technologies, you save money in the end. This is definitely something to consider, especially when talking to management.
Established tools usually have tons of Stack Overflow answers, tutorials, and battle-tested solutions to common problems. That's like having a massive support team at your fingertips! Often times, there is less information about very new tools which can lead to more hallucinations when using GenAI tools, as mentioned above.
Here's a interesting theoretical concept: the longer something has been around, the longer it's likely to stick around. This is meant for things that have no natural end of life. If a tool has survived for 20 years, it's probably solving some real problems pretty well. That staying power means something.
I tested and work with a lot of popular AI tools such as ChatGPT, Microsoft Copilot, GitHub Copilot, Google Gemini and Anthropic Claude. These have several strengths but also some limitations which play a role in this context.
Especially the third point is the reason that AI is often most effective with well-documented, older tools.
Here, many questions have been asked, many use cases have been described and there are less unknown unknowns.
At the end of the day, picking the right tools is really about knowing yourself and being smart about your choices. Don't get caught up in what everyone else is doing or chasing the latest trends without proper thought.
Sure, it is comfortable to stick with what you know, and it's equally tempting to jump on the latest tech bandwagon, but remember: the best tool is simply the one that gets the job done for you.
Sometimes the perfect solution isn't the fanciest one, and that's totally okay.
This article is based on my talk "Old Tools, New Tricks".