Interview with Camille Fournier on rewriting software at GOTO Chicago 2014

Hi, it’s Mike with UGtastic. I’m here at GOTO Conf 2014. I ‘m sitting here with Camille Fournier, who gave a talk about the great rewrite, or, well, if you should, maybe it’s not so great. You know, should you have a rewrite? Well, thank you very much for taking the time to speak with me. So, your talk, what was the title of it again? It was So You Want to Rewrite That. So You Want to Rewrite That. Okay, so it sounds like a little bit of pain to get to that. So, can you share a little bit about what your experience was? Sure. So, I just led Rent the Runway through a big rewrite of our legacy sort of ugly Drupal system into a, you know, Ruby front-end, Java SOA back-end platform. Before, I want to ask, Rent the Runway. Yes. What is that, real quick? Rent the Runway is renttherunway.com. We are an e-commerce website. We do designer dress and accessory short-term rent als to women. And we are actually, you know, you may, you are not our target demographic, but probably you know someone that at least knows of us or even has used us. We have many, many customers throughout the United States. Okay, so you were saying that you had a Drupal application that you rewrote into a Java SOA architecture. And then I interrupted you. Yes. Right. So, you know, I basically joined the company when we were first starting to think about this rewrite and led the team through it. And it took us about, well, it took us about two, a little more than two years to get to all of the major consumer- facing, customer-facing pages out of Drupal. We still have administrative pages in Drupal. So, this experience is really interesting for me because it just got me interested in why do people rewrite? Right. What makes them successful? What doesn’t make a rewrite successful? I think every software engineer is going to have to rewrite something at some point in their career. Probably at some point you’re going to do one of these really big rewrites. So, how do you think about it? What are the principles you should think about in going into this process to hopefully make it more successful? Because rewrites tend to come from a place of failure, right? You’re failing in some way or you see that you’re going to fail in some way. It’s failures on the horizon. And so, you know, it’s very easy. It’s very easy for a rewrite itself to fail because you’re already in precarious, you know, in a precarious place. How do you, you know, think about it to make it successful? Okay. And so, it sounded like you did a gradual evolutionary kind of iteration away from this monolithic architecture. You didn’t just gut the system and then replace it. Yeah. Yeah. Yes, I would say that the number one thing that we did right was we did a very iterative, slow rollout. So, we would take. We would take logic that was in Drupal. We would move it into Java and put that behind API calls in Drupal. And we did this for sort of core data and then core business logic. And then once we had done sort of a good chunk, we started moving pages in Drupal out of Drupal and into Ruby. So, we built a Ruby layer in such a way that it was a very thin client. It wasn’t talking to the database in the way that our Drup al had been. It wasn’t this sort of complex, you know, ball of wax. It was simply calling. So, we built two Java services to get JSON data and, you know, maybe doing a little munging of that and showing it on the web page. So, the one thing coming from a cultural community aspect is that there’s a strong identity or tends to be a strong identity in the PHP community and a strong identity in the Java community. I’m a Java developer. I’m a PHP developer. So, when you’re saying to the PHP developers, hey, we’re going to be taking your stuff and rewriting in Java, how did you kind of. Smooth the waters with that. Yeah. Well, so. And avoid, like, insult, like, oh, PHP. Yeah, no, I mean, I think that’s a great question. And, you know, for better or worse, we just, you know, we had lost a lot of the original developers of Rent the Run way over the years, you know, had been written by an outs ource team and then in-house. And I was not, you know, in great technical shape because it had had so many owners that the team that was there writing the PHP, they were not so involved. Yeah, they weren’t emotionally invested. They weren’t as emotionally invested. They were mostly, like, exhausted because they were supporting this thing that kept breaking. So, I think in their minds, anything to make their support burden a little bit easier and, you know, take care of some of this complexity and allow them to focus more on getting features out more quickly. I think the real thing is that developers want to ship. For the most part, developers really care about shipping, especially at a company like Rent the Runway. And so, our rewrite was always sold to them as. This will let you ship faster, better, you know, with much less pain. And it did that. And I think that helped with that. And, you know, coming here to GOTO Conf, the presentation you wrote, was it based off of your, strictly off your experiences at Rent the Runway or have you felt this pain before and are you reflecting your previous experiences? So, I haven’t really felt this pain to some extent before. I had never done a major rewrite like this. I had done… A lot of actual, like, sort of refactors in place where I think with a different team would have resulted in a major rewrite that may or may not have been successful. But I actually talked to a lot of colleagues and friends outside of the company that I knew had either done rewrites , considered rewrites, and not done them. I read a lot of stories and a lot of articles by people who had, you know, successfully or, you know, not successfully done rewrites because I think that this is a common thing. You know, developers… Developers do this at some point in their career. And so, I wanted to get more perspectives other than my own as to why people do this and what makes them successful or not. And now, just to kind of wrap up, are there any resources that you read or how did you maybe look outside of your own experiences to get more information about how to approach this? Did you read, like, the refactor book or anything like that ? I mean, I read… What did I read? So, I read some blog posts. There’s some… Some blog posts, actually, it’s really interesting. About, like, six or seven or eight years ago, there was sort of a flurry of don’t rewrite that blog posts, which are really interesting and actually worth reading. I think, you know, these things come in cycles. People write about them and then we all forget. And then we all make a bunch of mistakes and then we rele arn about them. So, I read a bunch of those. I read a little bit about the, you know, Twitter has a pretty good blog post about their rewrite. And I also talked to people that were there and worked through that. I did a lot of talking to friends. So, reaching out and just asking people. Yeah, and just asking people, like, for their experiences. If you know someone who’s been a software developer for 15 or 20 years, you probably know someone that’s done this scene, that’s been around for this and has some opinions. Well, thank you very much for taking the time to speak with me. Thank you. User groups with lots to say. Interviews and more. No way. Sharing great ideas in the tech community. Fascinating conversations. A plethora of information. Find out for yourself today. At ugtastic.com.