Resume-Driven Development Is Awful

Employers demand expertise in the Shiny New Thing™. We make a mad dash to learn juuust enough about Shiny New Thing™ so we can find a way to less-than-gracefully shoehorn it into our stack. Once our resume has Shiny New Thing™ added to it in some embellished manner, we start the cycle all over again at Next Employer™.

Working across the tech industry for several years has made me realize something very sobering: generally speaking, we hate our jobs. We want to do better, learn more, stay relevant. But we’re terrified that present & future employers won’t give us a second look unless our resumes are chock-full of the Shiny New Thing™-of-the-week technologies. And sadly, they’re not entirely wrong.

The cycle is maddening, and no party involved really benefits in the long term. No one likes “boring technology” anymore. But what if we did again? What if we all found a collaborative way to appreciate safe & sane tech stack options, stop chasing the next new thing for the sake of itself, and actually deliver business value over buzzwords? This talk will explore learnings & observations of this phenomenon, and postulate on what we can do as an industry to just… chill out already.

How To Start Loving Command-Line Tools

Take a moment to think about the preferred software you use to do your job. What comes to mind? Python? Nodejs? C#? Some flavor of SQL? One of the Cs? Some declarative vendor tool? tHe cLoUd?

Now: have you ever thought about how *insufferable* some of that tooling can be? Many developers are familiar with the massive potential bloat of the 'node_modules' folder, how nightmarish it can be managing your Python paths & environments, or how difficult it is to test or run any given software locally. Need to store some relational data? “Well, I better fire up a full-fat Postgres deployment!” And none of this even touches on "dependency hell". Yikes.

Many developers often forget (or in my experience, simply don’t know) that for the most common tasks on UNIX-alike systems, there’s likely already command-line software installed to do the job! The GNU Project collectively provides nearly four hundred pieces of software to solve common (and not-so-common) computational problems. Outside of the GNU ecosystem, there are plenty of small-footprint, low- or no-dependency software packages available through a package manager or direct install to solve any straggling issues you may have.

In this talk, I will demonstrate how to resolve some of the most common or frustrating tasks developers might face using a sample of these tools: local databases; data preprocessing; JSON parsing & filtering; parallelization; task orchestration; interacting with REST APIs & remote files; and more!

Dude, Who Broke My Linux Server?

Linux-based operating systems are at the very core of the magic of computing technology -- and have been that core for decades. But software developers, data engineers, cloud engineers, and so many other disciplines often have no idea how to work effectively within those operating systems. What happens when your code won't run? And how do you get things working again?

In this workshop, you'll be dropped head-first into a broken deployment machine, and you'll need to fix it, fast! This will be a fun & lightly-competitive experience for all involved. Teams/individuals will be scored automatically based on how well (and how quickly) they are able to restore their machine to its former glory.

Requirements: a laptop with internet access, and an SSH client.

Generative Programming By Humans, For Humans

The world has developed a newfound infatuation with “AI” technologies recently. Whether it be conversational output from ChatGPT, image output from DALL-E, or code itself authored by Github Copilot and the like. Many people seem to think that this is the first time software has ever been written by anyone but a human – but this couldn’t be further from the truth! Text & HTML templating engines, Interface Description Language generators, and heck, even plain-ol’ compilers & interpreters are classic examples of tools that write code for you that are in widespread use today.

This talk will take you on a journey through many methods by which you can have your code write more of your code using familiar, repeatable, and sane approaches – snake-oil not included.