For those unfamiliar, Artsy is a fine art marketplace. Knowing that, it follows logically to say that the form via which our partners list artworks for sale is an integral part of Artsy’s core systems. This form, known only as “The Artwork Form,” is whispered about in the halls of Arty’s New York headquarters. It is legendary. It is a colossus. It is old enough not only to predate React v16.8 hooks and context APIs, but Artsy’s use of React entirely. The first version of the Artwork Form was built in 2014 using ruby and haml, and began its refactoring into JS/JQuery/React a full 2 years later, after having expanded considerably from the original implementation. That process (at least what we’ve gleaned from our git excavation) was incremental, experimental, and passed through many hands before it landed in the lap of the current Partner Experience (PX) team.
Ufff, 2021 was a tough year. It was the second year of the Covid-19 pandemic. We tried our best to support each other through these challenging times and in that spirit, we feel like it’s important not to lose sight of our accomplishments. As we say goodbye to 2021 and hello to 2022, here are some Artsy Engineering wins from this past year.
What did you ship in 2021? We would love to hear about it and celebrate with you!
When I started at Artsy a few years ago, I’d never written a line of Ruby. I feel at home with JavaScript — it’s been my buddy since I started my career over 20 years ago. I’ve written enough tests in JavaScript that I sometimes feel like I can write them in my sleep (as long as they don’t involve async React events 😅).
Most of the code I write at Artsy is still JavaScript, but now I write some Ruby code too, and I’ve written enough RSpec tests that I’m starting to form opinions about what I think they should look like.
My most recent work has been JavaScript again. I’ve been writing Jest tests against one of our React apps. But rather than reaching for the testing patterns I’d become accustomed to over my years of JavaScripting, I’m finding that something’s missing in my Jest tests! My experiences with RSpec have me longing for two features in Jest:
We have a handful of regularly scheduled meetings in place at Artsy devoted to knowledge sharing.
But what about the unstructured ways in which we share knowledge? If structured sharing time demonstrates that a team is interested in spreading knowledge, unstructured sharing time demonstrates that spreading knowledge is the default mode for the team. Instead of the team forming habits of working in isolation or hoarding expertise, they’ve formed habits of learning from and teaching each other.
Recently, I needed to test a button that would make an analytics tracking call using react-tracking and then navigate to a new page in a callback. This presented some challenges - I wasn’t sure how to create a mocked version of react-tracking that would allow a callback to be passed.
With some help from fellow Artsy engineers Christopher Pappas and Pavlos Vinieratos, I got the tracking and testing to work. Here’s how we did it.
I recently encountered a problem where client-side data (returned from a Relay query) became out of sync after a user interaction. How can we make sure our data is consistent while maintaining a single source of truth? This post explores why a developer might want to update client-side data locally, the basics of Relay and its store, and how to delete records in the store when you’re not using a mutation.
Relay x Artsy x Me
Relay is a GraphQL client library maintained by Facebook engineers and enables rapid client-side data fetching in React applications. Artsy’s adoption of Relay coincided with our move toward using React Native for our mobile work around 2016. I joined Artsy as an engineer in November of 2020 (after transitioning to engineering from a non-technical role at the company.) When I joined, I was about a year into React development and completely new to Relay.
A common suggestion for improving pull requests (PRs) is to “make your PR small and focused”. I myself gave this suggestion in a recent article on this very blog about including context in PRs.
Like most internet advice, this can feel like the “draw the rest of the owl” meme. Even if we’re in agreement that I should make a PR smaller…how do I do it? How do I avoid a big PR when there’s a lot of cross-cutting changes to make? How do I create small, focused units of work when I’m building a large feature? How can I overcome my perfectionism and submit a PR that feels incomplete to me because the edges aren’t all polished?
I know that for many developers, especially those early in their careers, asking for help can be intimidating. I often fear wasting someone’s time or exposing myself as less skilled or smart than my team initially thought.
In my first month as a software engineer at Artsy (and barely six months into life as an engineer after transitioning from a career in communications), I was struggling through a ticket assigned to me as a “good first issue.” (The team estimated the task to be straightforward enough for someone new to the team.) After a few hours stumbling between the ticket, my code, and Google, I made very little progress.
Available now on Apple Podcasts, Spotify, and coming soon elsewhere, is Artsy Engineering Radio!
Solving problems in the world of software engineering can mean a lot of different things, and this podcast will explore what that looks like at Artsy. If you’ve followed our blog, you can expect a podcast that sounds like it. There are a ton of amazing engineers here at Artsy and we’re excited for you to hear their voices and stories. Our hope is that this podcast will make it easy for more engineers at Artsy to contribute to the public persona of Artsy Engineering.
In, a recent blog post, I discussed a fundamental difference between web and iOS deployments. Web software is deployed to servers that are under your control, while iOS software is deployed to users’ devices that you have no control over. This distinction really changes how you think about the code that you ship, because that code could be running indefinitely on devices that never get updated.
The previous post focused on this distinction through the lens of accidentally shipping (and then fixing) a bug. This focus on bugs is important, but focusing only on bugs left me unable to articulate an important, nuanced distinction between hosting code and shipping app binaries. So let’s dive in.