 Hey, everyone. Welcome back to the State of the Web. You've already met my guests, Jess Lee and Ben Halpern. They were here on an episode about Dev2, their developer community. In that episode, they kicked off the developer survey. And today, they're here to share the results about the state of web development. Let's get started. [?]. [?]. [?]. [?]. [?]. ?] [?]. [?]. [?]. [?]. [?]. [?]. [?]. .]. Jess and Ben, thanks for coming back. Thanks for having us. Yeah, awesome to be here. So tell me about this developer survey. What makes it different from other surveys? Well, one of the things that really make it special is, as you know about Dev2, it's a community of developers. So some of the questions were actually able to crowdsource from the community. And that way, we got to ask some questions that the community itself is more interested in. Yeah, and I think we have a good sort of cross-domain community where people are involved with different layers of the stack. And I felt like the overall sort of people who took the survey were a pretty good reflection of developers and users of the web. And you mentioned that the community helped make the questions. And they also did some of the analysis too, right? Yeah, we shared the raw data with the community so that they could dig through it and find some insights along with us. So how many people responded to the survey? We ran the survey for about a month and had just over 1,900 responses. And what are some of the things that you found in the data set? We saw that only 10% of respondents have utilized AMP before, but that twice as many people actually use browser notifications. And we were just wondering why that was and speculated that perhaps marketing people might see more value in browser notifications versus really understanding the value of using accelerated mobile pages. I can imagine that push notifications is a more tangible thing that people can start implementing than AMP, which takes a lot of work. So 40% of our respondents believe that users should not be able to disable JavaScript on the browser, which I think is a very contentious issue, so I think the people on the other side will be pretty peeve that almost half of our members don't really think you should be able to disable JavaScript. Why is that so contentious? I think the way people want the web to be, I think there's a lot of people who are really proponents of a flatter, simpler web as opposed to some of these richer web applications, or at least that we should really start there. And if you can't even disable JavaScript, it really sort of takes away at least some of that control and some of the tradition of the web. Seems like Chrome on Android is about to do something about that. It's true. Yeah, there was some blog post recently about how Chrome on Android, if they detect that the user's on a slow enough connection and there are enough guarantees about the web page working, then they might disable JavaScript to prioritize the page rendering. Yeah, and that's kind of the other direction. So it's definitely really sort of debated to this topic in our community, absolutely. Definitely. Anything else that you found in the data set? We saw that 25% of people still support IE10, which we just thought was pretty surprising considering IE10 isn't used that much across the world. So back end tests are a lot more prevalent than front end tests. About 75% of people test on the back end and closer to 40% on the front end. So front end automated testing really still has a lot of catching up to do. Going back to the IE10 thing, I'm curious to hear your impressions about why people are still supporting it. And 25% is a lot. I would imagine maybe that's due to institutional support. Maybe schools are all on older browsers for security reasons or something. What do you think? I think we would agree with that, just institutionally. It's just trickle down still. And people, it's a habit situation. I mean, I don't really know. Yeah, in order to stop supporting a browser, you have to make a lot of choices throughout the organization, perhaps. And some things just stay as the rule until you officially move off. And sometimes it's easier just to keep doing what you've always been doing. Yeah, I mean, people, I remember a GitHub issue came into our repo a little while ago. And they're like, what's your browser support? And we're like, oh, we didn't fully have that conversation. And we're like, OK, we're not going to support this. And I think it's time for some folks to consider making some of those decisions. There was another question in the survey about package managers. And I was surprised to see that 88% say that they use a package manager. Because one of the stats that I've learned is that two-thirds of the web is on a CMS. And those developers are not directly managing their builds. Does this suggest that most of your community is not building with CMSs? Yeah, I think that is pretty representative of the community. People are a little closer to the mental layer. Yeah, I think there's a lot of communities specifically around CMSs themselves. And I think a lot of it sort of demonstrates in the respondents that I think they're a little bit sort of loop in at a different layer of the stack within our community. So that's just kind of something about this data set. Yeah, and I'd say that we probably have more developers on the platform who have contributed to WordPress open source versus being like a WordPress developer. Were there any interesting outliers in the data set? Like maybe a question that people answered much less frequently than others? Yeah, so for some reason, we had very little people share with us what their primary mobile device is. And of course, a few people chose not to answer the gender question, but not really sure why no one wanted to pick Android or iOS. Interesting. Something I'm really glad that we did was limit the number of fill-in-the-blank questions because on the three that we did, one was, how many years have you been working on software? And somebody put in 500, of course. So those were some outliers that we definitely scrubbed out. I saw somebody for their gender put Batman. So I would imagine also, by asking demographic questions and also more of the technology questions, you can intersect them in interesting ways. Like maybe you find out things about people of a certain age group and what their education is like. Did you find anything interesting there? Yeah, so one of the questions we asked was, what was your pathway to software development? Believe the answers were computer science degree, boot camp, self-taught, or more of an online program. And we found that the youngest and the oldest respondents did not use any of those. Like they chose other. But for the oldest, there were also the most number of people to say that they've done all of the above, which also sort of makes sense. Like if you started coding many, many years ago, you could have seen it all and done all these different programs because you've had time to. Or back then, it was a little more niche. And you just didn't go through any of these programs because they weren't there. Were there any other interesting correlational analysis that you did? Yeah, so speaking of the ages of some of the respondents, young people were generally more satisfied with their libraries that they used. So the JavaScript libraries they've been using and stuff, young people trended much more satisfied. And there was a pretty strong correlation there. Young people were also more satisfied with the general performance of desktop. So that's an interesting kind of angle. Mobile development was kind of flat across the different correlations. But young people definitely seemed to be more sort of satisfied and OK with the performance on desktop. Interesting. People who had higher internet speeds were also generally more satisfied with the web for end users. And actually, Ryan Paolo is a community member. And he did some analysis and pointed out that 10% of people didn't know what their internet speed was. And those 10% were also less satisfied with the web, which is interesting. And just like a quick PSA that you can, that everyone should know what their internet speed is. And you can find that out online really quickly. I think if you Google speed test, it'll actually run a speed test. Exactly. And just make sure you get what you're paying for it. And obviously, it leads to higher satisfaction. Yeah, exactly. If you become more educated, you understand and appreciate your web experience a little bit more. And there was a question about GDPR. How did people respond to that? Yeah, so that one actually had our highest standard of deviation. People were all over the map in terms of their confidence level with understanding GDPR. I wonder what could cause that? Are people unsure because it's such a new thing? Or people just don't care about GDPR? Like, oh, that's a Europe thing? Like, what could be causing that? I mean, I think it could definitely be a mixture of that. But I think one of our takeaways is that we probably should all be a little bit more informed with GDPR because it might not be exclusive to Europe in the future. Yeah, and it's not like people were universally unsure. This was actually the highest deviation. So some people genuinely thought they did know. And then a lot of people really didn't. So speaking of things where there wasn't much standard deviation, universally respondents hated ads. So we asked about ads, and people were borderline consensus towards the one end of the spectrum versus GDPR was all over the place. So you can get a sense of developer optimism with some of the sentiment-type questions, right? What were you able to find there? Yeah, so Paul Calvano from the HTTP Archive community actually did an analysis. And he created this heat map that indicated that developers generally agree that the web is improving and that it's getting easier to build for the web. Ben, anything else that you found? Yeah, I mean, we looked at some of the data. It was pretty clear that Flexbox has really taken off in terms of technology and CSS side. Pretty conclusively now, you can pretty much use Flexbox. I think there's other studies that have sort of correlated this as well. People use Flexbox a lot more than they use other sort of preprocessors, which kind of predated it. So interestingly, Flexbox has really kind of taken off as a universal technology. And then the other sort of the newer CSS stuff really hasn't at all. So Flexbox is really, you know, you're golden. You're gonna get a lot of adoption there. Some of the other stuff hasn't really taken off. Like Grid? Yeah, Grid wasn't quite there. Like Flexbox really was among the different kind of specific CSS newer technologies. Flexbox really has much stronger adoption. Yeah, and just going back to optimism question again, we asked everyone if they're a dog person or a cat person, and it was pretty clear that dog people were generally more optimistic about it. That's really surprising. I wasn't just even surprised by the optimism. Dog people tended to be more disciplined testers. They cared more about accessibility. They felt like they had better communication with their team. It genuinely seemed like you kind of want to be working with dog people. And it's the possible that we bias a survey because we are dog people, I think a little. I'm not a dog person. I'm not a dog person. But you're not a cat person either. That's true. But what do you mean by bias the survey? Well, so, you know, I don't think we really did, but we had the biggest hand in creating the survey. Obviously, we got a lot of help, and I think things are pretty good. But if I'm a dog person, I may have influenced sort of how we ask the question to my fellow dog people, make it a little easier for them. Maybe dog people just are more optimistic. Jess, are you optimistic? Sometimes. Was there anything else that surprised you or maybe in terms of responses that need more improvement? For example, like GDPR, it would be great if people were all understanding and moving the needle in that direction. But anything else that you can think of that we'll need to work on as a community? I felt like I was confirmed in thinking that testing on the front end isn't quite there and hasn't really been adopted very well. Obviously UI testing isn't the easiest thing to do compared to some of the unit tests on the back end, but we definitely have a lot in way to go with that, but the tools are really improving, so I'm pretty hopeful about that. But clearly, that is an area of our industry that really is lagging a little bit. And I think just a general reminder that not everyone has access to fast internet. People in our survey mostly did, but there are still some respondents with less than half a meg that's probably more prevalent than we realize people who have that privilege. Do you also plan to run this survey regularly and kind of like track the responses over time and see how things are trending? That'd be kind of cool. Yeah, I think we're gonna do this again. I think at any point, if you go to dev2 slash survey, we're gonna be doing something either sort of keeping on with this analysis or sort of the next thing, and definitely think we're gonna do a lot more of this in the future. Especially since we're open source now. Doing these surveys really help us have more of a pulse on what our community feels stronger in, for example, like you were talking about Flexbox, just what we know folks are more capable of and are comfortable into. Yeah, and obviously for us, this was a survey of developers, both as users and as sort of creators. And we always have to be aware that this is not necessarily the community we want all the answers for in terms of the usage and stuff like that. Our community is not really gonna be representative of everyone's needs, and it's definitely important to get data from a lot of places, although this is really useful data. I think you can bring to your boss if you need to really settle anything. Yes, that's a good closing thought. And with that, Jess and Ben, thank you so much for being here and sharing the results. You can check it out at dev2 slash survey. There's a link in the description. Thanks for watching, we'll see you next time.