Functional Tools That You Can Take Home!
An ACM Talk by Bailey Parker on 2018-03-30
Most students first learn imperatives languages (like C++, Java, or Python) and never look back. Some may be exposed to functional languages for a course, but never find other uses for them. Functional programming, though, can teach useful skills that transfer over to your day-to-day programming in school, internships, and jobs. Join us for a brief survey of the basics of Haskell spanning recursion, composition, immutability, and laziness and concluding by delving into the magic of monads. Along the way you'll pick up some tips for writing clearer, more concise, and less buggy code placed in real world contexts such a ML and CDNs.
NB: The mention of ML and CDNs is a little clickbaity! I provide some much simplified examples of how you could use functional to reason about them better, but this definitely isn't a detailed talk about either.
This talk is part of a larger vision I have to create a series of talks about practical design patterns used in the real world to great effect. Many CS curriculums focus more on computer science as a science and not an art. And while there is undeniably a science to solving problems by composing algorithms and using prior art and research to understand a problem, I contend that you should also focus on the art of CS. Treat the code you write like an essay; it should be well organized, clear, and free of grammatical errors. Code that is simple and conveys its intent clearly is a treat for the reader. And this directly translates to less friction when changes are necessary in the future. Too many programs train students to be Get It Done™ machines without giving much pause to consider the next person to read the code. Just like a writer wouldn't publish scribbled notes from a journal, you shouldn't publish code that you wouldn't want others to read. Your product should be a labor of love, a manuscript with attention to all of the details.
Metaphors aside, this manifests itself in software design patterns. I've found course offerings and academic stress to be lacking in this area. And this is only to the students' detriment. If/when they enter the commercial workforce, they'll be expected to jump into large code bases and contribute code that may still be relied on for years to come. Without experience in properly applying design patterns, they're forced to pick up this critical skill on the job while adjusting to the new environment. I content that software design should be taught as early in the process as possible to preclude this.
Students who are frustrated with CS aren't "bad" at it, they just haven't been show best practices that will make it easier to focus on learning (instead of keeping track of lots of mental state or permuting code with abandon until it "just works"). Good practices like writing simple, testable, decoupled, testable code go a long way to easing the learning process.
Patterns aren't just pedantry. They arise out of the fact that code is written by humans. And, humans aren't prefect. Design patterns allow our limited biological computing power to focus solely on the important things so that we can more easily write bug-free, beautiful code.
You should be able to view index.html in any recent web browser. The most
recent version is also available on
GitHub pages.
I fixed a few before uploading, but if you find any, please file an issue or submit a PR. Thanks!