Jack Franklin

Letting tools make choices

In my first post of the year about technical choices I'm making in 2020 I finished with a point about letting tools make choices for you:

I've become a fan of putting more of that burden onto the technologies I'm choosing such that I can focus in on the actual application. I would be surprised if the tools that catch my eye in 2020 aren't ones that follow this pattern.

This has been a trend for me and my work over the last year or so and I wanted to use this post to expand on it.

Holding onto control

I would never have written a post like this a few years ago. One of my main decisions when deciding what tools to use was how much control I had over the entire surface area of the tool and how much I could customise it to my every need.

My logic at the time was that if I could configure my environment and tools to precisely what I wanted, I would create this zen-like application that was set up to enable me to be the most productive I could be.

Time spent configuring

My approach wasn't correct because of two things that are true of all software:

  1. Software is never finished.
  2. Software is never perfect.

I could never get to this perfect setup of tools because there would always be something left to do! Anyone who has worked on an engineering team knows this; teams create sprint boards and backlogs and those backlogs inevitably end up with far more items in than your team could ever achieve. I remember sitting down one evening after work to focus on a side project and losing the best part of the evening trying to get two different tools that I'd chosen to use playing nicely alongside each other. I finished for the night and realised that I'd made no progress. I didn't even need those tools to work together to allow me to make progress, but I was so concerned about having the perfect setup that I forgot about building the actual application.

Once I had everything playing nicely, one of the tools would have an update which broke something and I'd repeat the process all over again.

Shockingly, that project never saw the light of day (and there are many more like it).

Losing hours to extreme ESLint configuration

The amount I valued control really became apparent on another side project where I probably spent the first two hours just configuring ESLint. I can't tell you how long I debated in my head between Style A or B, all while having no actual project code and basing my decision off dummy code I was writing to test my ESLint setup!

At this point as well I was the only developer, so why it really mattered to me that I stuck to an incredibly strict set of rules I'm not quite sure. Yet again I'd scuppered a side project before it had even started.

I still use and value ESLint in my toolchain but apply far fewer rules than in the past.

Working on a team

When I got my first professional job out of university I joined a team who already had a set of conventions for their code that newer members of the team were expected to stick to (although we had a good culture where anyone could suggest changes/new additions). It shocked me how much I struggled with this! I would sit at my desk not wanting to write code or even trying to avoid certain language features because it would frustrate me to do it "wrong". Looking back on this it's very embarrassing to admit and silly that I got hung up on it.

Deciding what to work on

Once I'd worked in an engineering team for over half a year it dawned on me that every person in every team has far more they'd like to do than they can actually do. Regardless of company size, team size, an individual's role or experience, there is simply too much to do.

Once I realised this I began to think about what I value most and what I want to spend my time doing. At work, I like building things that people use. I like finishing off a nice UI to help users with a particular problem. I like building tools that other engineers use that helps them be more productive. I like improving our acceptance tests after a bug makes it into production so it can't happen again. That's what I like doing and it's how I have the most impact on the company, my team, and the people who are using the products we build.

I decided that any work that takes me away from the core of my job and what I like doing was not worth dedicating multiple hours to. I still care about it and still work in this area (as I said above, we use ESLint at work to help us) but I spend far less time than before. On side projects I'll tend to chuck my favourite ESLint config in and be done with it so I can focus on the actual project itself.

Tools that make decisions make me more productive

The best example of a tool that I've come to love is Prettier. It's far from the first code formatting tool out there but it popularised the idea of automatic code formatting for frontend development. Suddenly I went from strictly controlling all aspects of my code formatting via thousands of ESLint rules to having one tool that made a set of choices that I couldn't change even if I wanted to! Prettier is opinionated and it's what I love the most. I install it, set it running and it formats my code how it thinks it should be formatted and that's that. I don't like every decision Prettier makes with code, but I like that its made that decision for me.

The trend of zero config tools was perhaps overhyped slightly - there will always be projects with edge cases and certain situations that rely on full customisation and control - but tools that lean towards being less configurable have tended to find their way into my toolbox.

For small side projects these days I'll often reach for Parcel as an out of the box bundler because it just handles anything I throw at it without much effort, and it's not surprise that Jest has become my testing framework of choice for the same reason.

Opinions are still encouraged!

Whilst I may not dedicate as much time and energy to configuring a project to within an inch of its life that doesn't mean that I don't have opinions. I still have strong opinions and preferences about how code should be formatted, which ESLint rules are good (or bad) and which test framework is my favourite. I spend a lot of time thinking about software design and what "clean code" really means. My colleagues and I regularly debate exactly how to write a piece of code and we definitely don't always agree.

I've learned that it's not about not having opinions, but instead holding onto them less strongly. Strong opinions, weakly held is a great way to frame these discussions and allow you to have opinions but be open to discussing them and having those opinions challenged. This process, coupled with the practice of Disagree and Commit has enabled me and my team at work to continually challenge our thoughts, opinions and ideas whilst not getting bogged down in debating for the sake of it, or becoming frustrated with anyone refusing to budge on their viewpoint.

Tools that make decisions help beginners get up and running

Preferring tools that make decisions is why I'll push people towards create-react-app if they want to learn React but aren't sure where to start. I've seen many people try and fail to learn React not because they aren't capable of learning React but they give themselves a huge mountain to climb and try to learn React, Webpack, Babel, and more all at the same time!

If you're teaching beginners, or working with junior developers, encourage them to focus on the job in hand and what's really important and let tools fill in the gaps.

There's nothing wrong with someone being comfortable with React and deciding that they would like to learn what create-react-app does under the hood, or spending a weekend building their own rough version of React to gain a greater understanding of the tool. But when you're just getting started it's important to focus in on what's really important. Anything else can be deferred.

Clearing your head and focusing

Another benefit I've experienced is that once you offload decisions to tools you're able to think more clearly about the problem at hand. That applies to beginners trying to learn something new, but it applies to starting a side project and working on what's really important.

I encourage you to think about this next time you load up a project at work, or at home. Where are you spending time and energy that you can offload to something else, freeing you up to do your best work?