Jake Worth

This post is part of a series about preparing technical talks; the introduction is here.

Today, I'll be covering the third step of my process: researching and preparing slides. Every speaker I know has their own techniques for handling this part of the process; I look forward to sharing mine.


When possible, I start in an unusual place: the library. Libraries often contain a wealth of information on many technical subjects. Connecting a topical web development subject to the history of computers or general human inquiry is something Sandi Metz, a speaker I admire, does well. Even when the material isn't there waiting to be found, there's always something, as well as professionals who can help you devise a research strategy.

Technical talks, often mislabeled 'hard' talks, are what I prepare most often. They require experimentation. I use the ideas I found during brainstorming to try a lot of experiments in my text editor or REPL. What are the nuances? Which are worth explaining?

Preparing Slides

After trying PowerPoint, Keynote, and even Vim buffers navigated with ]f and [f (an idea I borrowed from Steve Klabnik; not recommended unless you're an unflappable speaker), I discovered Deckset.

Deckset is a feature-rich markdown slide builder, and I love it. If you can write markdown, Deckset is a fantastic tool. It lets me crank out ideas and proofs-of-concept in Vim. If you feel that the hardest part of preparing a talk is actually putting together the visuals and notes you need to speak confidently, then Deckset is worth trying.

As for slide design, I follow the guidelines put forth in Speaking.io and the Chicago Ruby speaking guidelines. My short version:

  1. Make your text as big as you can.
  2. Use the most basic code examples that you can.
  3. Use high contrast colors, like black text on a white background. Subtle terminal palettes like Zenburn fail on a projector.
  4. Don't show code unless you want your audience to read every line, because they will.

My talks don't feature funny GIFs, pictures, or video. Why? I don't trust myself to be funny, and don't want to invest any of the time I have to prepare on trying to satisfying that quadrant.

I structure my talks like essays, with an introduction, thesis, 3-4 supporting points, and a conclusion. A little formulaic but it works.

Something I often ponder is: could this be a blog post? One of the less effective types of presentations could be summarized as a person reading a blog post they wrote. A telltale sign of such a talk is when the speaker concludes with a link to a blog post that inspired the talk. I've given a few like this that I didn't love. A good talk isn't a recipe for accomplishing some narrow problem you solved, it is a performance only you can give.


Researching and preparing the slides is a unique process and these ideas have worked for me in the past. Experimenting and automating my process has helped me to quickly develop talks that are detailed and hopefully interesting to watch.

Next up in this series, I'll be covering the step that I think defines all great talks: practice.

Dec 29, 2017

Hi! I'm Jake Worth, a developer at Hashrocket, based in Chicago. I co-organize Vim Chicago. Read my blog, learn about my work, follow me on Twitter and Github, get in touch.