By now, most iOS developers have probably heard the news that Airbnb is moving away from React Native. It’s kind of weird to see the reaction online – people are acting jubilant, almost celebrating the announcement. I see some iOS developers on twitter practically rolling their eyes at Airbnb for even trying React Native. But I’m not here to talk about insularity of the iOS developer community, I’m here to talk about Airbnb’s (quite excellent) blog posts.
These blog posts are excellent because Airbnb first lays out its expectations for React Native, and then explains how well the technology met (or failed to meet) those expectations. It’s a thorough, honest discussion about how large teams evaluate technologies, as well as how those team live with the consequences of their decisions.
So I really like these posts, but I wouldn’t say that they’re complete. More on that later. First, I want to discuss Airbnb’s actual expectations. I’m going to summarize them here, but you really, really should read the posts yourself.
Here are Airbnb’s goals for React Native, which I’m interpreting as their expectations for the technology:
- Allow us to move faster as an organization.
- Maintain the quality bar set by native.
- Write product code once for mobile instead of twice.
- Improve upon the developer experience.
Okay, so how did those four things go? Well, I’ll summarize again:
- “Engineers were able to move at an unparalleled speed”, but also there were also technical and cultural hurdles that mitigated this success.
- Again, a qualified success: “we were able to accomplish a number of things that we weren’t sure were possible.” But they point to the same technical challenges and a lack of internal engineering resources that frustrated their team sometimes.
- This was a failure(*): “we wound up supporting code on three platforms instead of two”, which sounds grim. But also, “only a small percentage of our app was React Native”, meaning they had a lot of overhead for not a lot benefit. I want to discuss this expectation in more detail, but more on that in a minute.
- A “mixed bag”, which some things being way easier and faster, while other things were worse.
(*Note: my characterization of #3 as a failure has been disputed; you should go read Leland Richardson’s thread responding to me to get a really great, nuanced clarification of Airbnb’s experience sharing code between Android and iOS.)
Okay so at first glance, this seems kind of promising, right? I mean, Airbnb says that “60% [of engineers] would describe their experience as amazing”, so a majority of your team is on board. But 5% were “strongly negative”? Interesting. Very interesting. We’ll get back to this shortly, first let’s talk about that third expectation.
The big point everyone is talking about is number three: Airbnb expected to be able to reuse code across Android and iOS but were let down. They wanted that “write once, run everywhere” experience. And that’s the problem that I have with these posts and how they’re being discussed.
React Native is explicitly not “write once, run everywhere”, although this is a common misconception. The idea that you can write one app and deploy to Android and iOS is attractive, and solutions like Expo do exist to do that, but cross-platform code reuse is not the central selling point of React Native. React Native is a “learn once, write everywhere” platform; instead of building one app for multiple platforms, the main selling point of React Native is that it lets you use React.
And that’s it.
So when people frame Airbnb’s decision as a failure of React Native, I wonder about how much they really understand about what React Native is for. If you have an expectation that a technology can do something that it’s not designed for, and then you’re disappointed when it fails to deliver, it’s not really the technology’s fault, is it? And further, if “only a small percentage” of your app even uses that technology, then you don’t get as much of the benefits.
And that kind of leads me back to culture. I think these blog posts do an excellent job of describing the technical parts of their experience, but they lack a detailed description of non-technical barriers to adopting React Native.
What kind of barriers? Well, not to bring up the insularity of the iOS developer community again, but… iOS developers kind of have a reputation within the broader software industry. Charlie Cheever, writing about this topic from Expo’s perspective, has this to say:
[…] organizations where there are people who identify themselves strongly as iOS programmers and Android programmers have a really hard time being happy with React Native. iOS programmers in particular are very unhappy with it and generally regard JS as an infestation of the company’s codebase, while Android programmers have more mixed feelings.
When I see a majority of a team describe their experience as “amazing”, and a very small (but very vocal, I’m sure) minority of the team describe it as “strongly negative”, it really makes me wonder about the internal discussions about React Native at Airbnb.
Airbnb’s experience has been compared, appropriately, to a rejected organ transplant. If your company’s culture is really, really into iOS native development, then the organ will be rejected.
The reasons that Airbnb chose to move away from React Native are cultural as much as they are technical, and while I understand why they didn’t publicize these tensions, it gives the impression that Airbnb’s decision was based entirely in technical merit. And that’s just not the case.