I’m amazed at how frequently I still tell people to “Google it,” “ask Alexa,” or just “look it up.” Just today my son told us his teacher asked the class what “Yeet!” meant. Meanwhile I’m thinking, “Why didn’t she just type ‘define yeet’ into a search engine.” If it wasn’t for DuckDuckGo, UrbanDictionary, and having a clue about video games, I’m not sure I’d even be able to hold a conversation with him.

After 20+ years of having the answers to all our questions immediately at our fingertips, you’d think we’d be quicker to search than we are. This got me thinking, “Maybe I’m too quick to search. Maybe I’m searching too much and thinking too little.”

Early in my software development career, Google didn’t exist. We had Yahoo!, but the internet was young enough that there wasn’t much useful content. Blogs hadn’t yet become a thing, and so you were left to search through newsgroups, or even worse, ask a question on one. Instead, I owned a lot of books, read a lot of documentation and man pages, and basically muddled through as best as I could.

This isn’t an article romanticising the “good ol’ days”. They were good, but being a programmer today is better. A lot better. Still, I began to wonder what it would be like to take one day a week and go without all the search amenities of the modern programmer. What if I couldn’t “google” a question, search StackOverflow, or even search through issues on GitHub? Could I do it? What would I discover in the process?

What I Wanted to Learn

It wouldn’t be much of an experiment if I didn’t have some theories I wanted to test. These are the questions I wanted to answer:

How will this affect my thinking?

I believe that above every other reason, the challenge of solving problems is what draws people to programming. It’s certainly what I enjoy the most, but what if I was robbing myself of the joy of solving problems by seeking out answers from others instead of attempting to solve them myself?

This isn’t a “not invented here” issue, it’s more of a mental fitness issue. What effect, if any, will a day of fasting have on my ability to reason and problem solve? Will this weekly exercise strengthen my thinking? Will I find greater enjoyment in tackling problems which might otherwise already be solved?

How will this affect my learning?

I read a lot, and while I learn a lot from what I read, I don’t retain most of it, because I don’t always put what I learn into practice. It’s the “practice” which translates the theoretical into the practical. Unless I take the time to understand what I find from the search results I miss out on not only how to solve the problem, but possibly the steps the author used to reach his or her conclusion. Unfortunately, I usually don’t take the time; I don’t think I’m alone in that.

By taking a fast from these instant answers, I’m hoping the required extra effort will not only improve my thinking, but also force me to learn about my chosen technologies more deeply.

How much will it slow me down?

This is the big issue. I’m a firm believer that the three virtues of a programmer are laziness, impatience, and hubris. I pride myself on getting work done quickly and correctly, and slowing down because I’ve arbitrarily limited myself might drive me mad. More worrisome, it might drive my employer mad.

Will fasting from prepackaged solutions actually slow me down? Oftentimes what I search for are answers I already know, I just don’t remember the specifics off the top of my head, or I’m looking for an alternative approach. So will fasting hinder my productivity?

How will it affect decision making?

As I’ve thought about this experiment it occurred to me that sometimes I perform a search not because I don’t know the answer, but because I’m hesitant to decide on the answer and I’m looking for confirmation. Worse still – if I’m honest with myself – I’m often looking for someone else to make the call; to take responsibility.

This experiment is about uncovering and working on weaknesses. Not making decisions and not taking responsibility is a weakness. I’m going to use this experiment to highlight those times I’m vacillating between options and use it as an opportunity to strength my decision making skills.

Rules

For an experiment to be successful there needs to be some constraints. These are the rules I’ll be following:

One Day a Week

I still have a job to perform, so I’ve decided to take one day a week to fast. I’ve chosen Wednesday, because I’m able to be heads down all day and rarely experience interruptions.

If you’re playing along with the home version, I’d suggest avoiding fasting on days which are high intensity: meetings, deployments, interviews, etc.

What’s Allowed

I’m allowing myself the following resources:

  • Documentation: This includes man pages, language and framework documentation, and other technology documentation.
  • Books: Books are very similar to documentation in that they explain how to use a technology, but rarely provide the perfect answer for a problem you are looking for. They also require some amount of research to be able to form your own conclusion.
  • Source code: Is there a better way to understand your technology of choice than to see exactly how it’s built?
  • DuckDuckGo bangs: This sounds weird, I know, but I use DDG’s “bangs” to access GitHub and documentation sites.
    • Examples:
      • !hexpm timex
      • !gh phoenix
      • !gem rails
  • Non-work related searches: This experiment is about focusing on solving work related problems. Searches for places to eat, movies, stupid trivia questions, or images for a presentation are allowed.

What’s Not Allowed

The following resources are not allowed:

  • Personal blogs: Often when we search for answers we wind up on the personal blog of someone who’s solved the problem. No personal blogs are allowed during the fast. The point is to figure out problems yourself.
  • Search engines: Yes, I’m calling this a “Google Fast”, but that’s for marketing. I’m also including DuckDuckGo and the other ones as well.
  • Answer sites: StackOverflow, Quora, and the like are off limits. No easy answers!
  • Forums: Like answer sites, forums offer an easy way to get answers instead of researching and discovering them yourself. I’m including things like GitHub issues in this as well.

No Cheating!

The number one rule is No Cheating! I started my career without an infinite array of easy answers, I can survive one day today without them too.

The Results

Week 1

It figures. The first day I run my experiment, and there was literally nothing I needed to search for. There was one thing I briefly thought about looking up, but as soon as the thought occurred I remembered the fast and quickly realized I could solve the problem with a little Unix-fu.

I did search for a Paul Graham quote for a non-work related conversation. I also searched for images to use in a presentation, but both items are out of scope for the experiment.

Week 2

The second Wednesday was a complete bust. It was February and I got hit with a 2-day flu. I was out cold. There was no searching, only fever dreams and Netflix.

Week 3

Like week 1, I literally had no need to search. I don’t know if this speaks more to my competence or the simplicity of the problems I was working on.

Week 4

I cheated. I think.

I was writing tests for an API client and I needed to know how to mock out the the HTTP library. I knew Elixir could do it with behaviors, but I wasn’t finding what I was looking for in the documentation. However, I knew Elixir’s creator, José Valim, had an article detailing exactly how to do it.

I had the article bookmarked and I used it for reference. I may have been able to figure it out if I spent more time on it. I let myself get fixated on an imaginary deadline.

Week 5

I was having trouble creating records through an association using Ecto, Elixir’s data mapping library. For the life of me, I couldn’t figure out why it kept erroring on the inserted_at date. I was really tempted to do a search, but I resisted.

I don’t think I would have found the answer with a search. It turns out that I forgot to add the timestamps/1 function in the schema. I only discovered that by paying closer attention to the error message. Duh.

Week 6

Again, it’s another day without a need to search, but I’ve noticed that even on non-fasting days I’m now slower to look for the easy answers. Was it really a day I didn’t need to search, or am I building a habit of choosing alternatives?

Week 7

Today was a turning point. In the past when I would see an error message I was unfamiliar with, I’d use a key phrase from the error message to perform a search. Today, instead of searching the internet, I opted to search for the string in the library’s code. I found the answer in a test and also learned several other features on the library in the process. Had I searched as normal, I may have found a fix for my problem, but I definitely would have missed out on what I learned.

Week 8

I was working with Looker today, and even thought it’s not really coding, I continued to follow my rules for the fast. I don’t remember the specific problem I was facing, but forced myself to stick to the documentation rather than look through the forums. Wasn’t a hard call, most SaaS forums are rubbish.

Conclusions

I went into this experiment wanting to know how limiting myself from easy answers would affect my thinking, learning, productivity, and decision making. While eight days isn’t enough time to gather conclusive evidence for any experiment, it was enough to see patterns emerge.

Because I began paying attention to my search habits and which sites were restricted, I found myself less likely to follow easy links even outside of the days of fasting. More than that, I also noticed a change of attitude in myself toward sites like StackOverflow, Quora, and forums. I began to see using them as a weakness. Mind you, I’m not saying using these tools are a weakness, per se, but rather I perceived them as hindering me on my journey to improve.

Although my thinking was affected in a rather unexpected way, I can happily say that as I had hoped, the fast encouraged learning. By relying more on searching through source code I was 1) always able to find the solution to my answer; and 2) discovered new techniques and parts of Elixir I didn’t even know to ask about. This is to say nothing of what I learned merely from thinking through problems and reading the documentation more thoroughly.

Here’s the big question: did the fast hinder my productivity? I don’t think so, but that’s not a question that can easily be measured. Part of why I don’t believe it slowed me down is that both Elixir and Phoenix are relatively young technologies; more than that, however, is that they are used by a small percentage of developers. I mention this, because older, more widely used technologies have more written about them and so finding answers is easier. If there are fewer questions/answers available then it might be faster to start off with the documentation and source code rather than trying to phrase your question just so to find the perfect link.

The last question I wanted to answer was “how will the fast affect my decision making?” I can’t say I was able to tell a difference, nor did I experience a moment which stood out in that regard.

What Now?

Now that the fast is over, what happens? Will I continue limiting myself to source code and docs, or will I slide back into the path of least resistance? It’s hard to say for certain: old habits die hard. I like what I discovered about myself, and I liked the confidence that came from searching through the source to answer my questions and uncover hidden gems. Now that I see more of what my own capabilities are and the joy of uncovering knowledge, it’s difficult to imagine taking the same old path.