 Okay, we're back here live at Velocity Conference. This is SiliconANGLE.com's theCUBE, our flagship program. We'll go out to the events, extract a signal from the noise. SiliconANGLE, an independent media company. Partnering with O'Reilly Media to bring theCUBE to all their events. And one of the most exciting things about theCUBE is we get to talk to all the alpha geeks, the entrepreneurs, the developers, the CEOs of startups and big companies, but also people who are domain experts. And our next guest is Nicole Sullivan for Twitter handle, Stubbornella, at Stubbornella. So I just tweeted on theCUBE, at theCUBE. You can get that Twitter handle there. You just gave a talk. Welcome to theCUBE. Thank you. So theCUBE is where we kind of have good casual chat. So, and the Twitter sphere is watching. So when I ask you, what do you think, what did you think about your talk and how it went over? I thought it went over really well. A bunch of people came up in the end and had lots of questions. I feel like good questions is a good judge of whether a talk worked or not. So yeah, I got hard questions. If you get mobbed at the end, and you know, I was kind of like, ah. Mobbed and I think I'm a little bit wrong. I feel like I've done my job correctly. Otherwise there's no interaction, you know. I always said, if you say the sun's going to set in the east, you're going to have someone say it's wrong and that's the internet. But one of the things we've been talking about, Nicole, I want to get your perspective up because, you know, David, and Valente who was here yesterday, he and I were commenting on the wrap up yesterday that this is not a cloud show. This is not a web show. It's not a UI show. It's kind of an intersection of all. So you've got the dev ops piece, the web ops, which Facebook, Google, and the big web scale companies have done and had to build their own stuff. You have essentially this middle kind of software optimization perspective. And then you have the UI design, all kind of interacting. So the common theme yesterday was it's a holistic systems design, thinking about the user experience with the trade-offs of infrastructure. So with that, you know, your talk is more on the CSS side, that's the front end. So, you know, what is the front end, state of the market for the front end? I mean, honestly, we all know mobile's great Facebook announcing video today. And so obviously the user experience is number one. But what's happening? What are the key trends that you're seeing around the user experience, the tools, the languages? Well, I think at first it's really funny what you said because I spoke once to a database conference, which I'm not sure why they had me. I think I was there for like, you know, we should have some diverse perspectives. But I talked to them and I said, well, regardless of how fast you make your queries, I can bring your site to a complete halt with some terrible UI code. And they're like, what? No, but it's true. So I think that in terms of performance, the UI really makes a pretty big difference. The trade-offs, I mean, the database guys are thinking, wait a minute, I'm going to serve, you know, out of cash or solid state or flash. I'm done. Hey guys, take it over, but you're saying. Well, especially for big companies, you know, so typically I do big UI refactors of larger companies or startups that have gotten to that point of success that they have to actually deal with everything they did to get there. And so when it's at that scale, you know, whether you serve 50 images or 25 or 10 makes a really big difference in terms of the overall load that you're putting on. Are you prescriptive? You walk in and saying, here's how you got to fix it. You're the doctor. You come in, just like, okay, is it surgery? I mean, the patient on the table is the UI, right? I mean, what's the conversation like? Usually there are things that they're doing really well and other things that are just not working at all. So I try to frame it in terms of taking the bits that are working well and making them play nicely together rather than having a system where different parts of the system are competing with each other. You know, I love the Zuckerberg philosophy, just break stuff, because I have rolled my eyes. That's great for, you know, the headlines. It's like, okay, does that mean no QA? Does that mean dev ops? And also, you know, code matters, right? You got to have good code. So what is the big thing that you're seeing in the performance side? Because startups can do agile, and that's been great for the startup community. I can put up on AWS, I can build it up, and then you get to that point where you're saying, okay, we got validation. We have to get things right now. Got some big funding, we're growing like crazy. But the rocket ship in the middle is in the air. How do you change the engine in mid-flight? Is that kind of a challenge? Well, it's very carefully is the answer to that. And there's some new trends in terms of the UI that are actually helping with that. So some of the flat design stuff has been really great for performance because generally speaking, simpler designs are faster to render. Some of, well, but on the other hand, some of the topography stuff has made it harder. So it feels like it's a push and pull most of the time. As well, responsive is an incredible opportunity to make things faster and better, but often goes the other way and you end up making things slower and clunkier. So I think it's all about realizing that some of those goals are great goals, but how we get there really matters. Yeah, we had the Etsy guy on yesterday and we had some other folks on. They said, you know, perception is reality. Perceived speed is an interesting dynamic and there's some art and science to that. What have you learned in that area? Because it is kind of an art. You kind of have, but there's also some, you know, some either math or analytics behind things. How do you look at that, that part of the game? Because page loading or app loading is pretty critical. So we always start by getting the data because if you don't have data, then you can't make any decisions. So I think that that's the biggest first thing. And people don't think of the UI that way. They don't think of it as something that they can quantify and something that they can measurably improve, but you actually can. So we start usually by going in and figuring out what went wrong and it's sort of like an autopsy for a site, I guess. Well, you know, what exactly has made things kind of get off the rails in terms of the UI? What have you seen the most in terms of use cases that made things go off the rails? There are a lot of things, but I think maybe the biggest is the perception that your views, the way your views are architected is going to match the way your UI is architected. And that's actually one of the best ways to make your UI incredibly inefficient. And so it's weird because you think you've got, like, you know, if you're writing Ruby in the middle layer, you're like, oh, you know, it works like this. Each view has its SAS file. Each, you know, each piece is very clearly organized, but it turns out you're, it's basically like taking your DB schema and trying to configure Apache with it. It doesn't work. I mean, you can, in this case, you can force it, but you end up with something. You can ship it, it looks good, it will work, but when you get to X number of views and traffic, it... It doesn't hold up. You know, what's to the bed, you know, it kind of crashes. And usually people get to the point where they can't really release features anymore, which is kind of a pain as well. You know, I always say, you know, the joke is, what's the rookie mistake most people make? But a lot of the developers that are doing all this front-end work for stars are really good coders, and they will do that. They'll go to what they're comfortable with. They'll say, okay, here's my database scheme, I'm going to match this, it's elegant, it's linear thinking, okay, it's done, it works. And then, you know, they're successful, and then they have to kind of go, okay, we got to fix this. So the question is, what's not the rookie mistake? What is the pro mistake that people make? Because a lot of those guys, they have to then go out and budget and then hire real UI help, real surgeons, if you will, or kind of reboot the UI. So those guys are pro developers, but then UI is a whole other ballgame, so they have to then build the discipline. A lot of startups, for instance, need to do that. They don't necessarily want to go out and fund the UI teams too many, too expensive, but they'll do it themselves and then they'll get to the point where they are, okay, fix this. Uh-huh, yeah, I think that's, it's okay that that happens, right? They're trying to be successful at something, and that's- Better than going out of business? Yes, it is, right? I mean, you could invest in making the UI layer code perfect from day one, but I don't think that makes sense. I actually think it's perfectly okay to like, pivot and throw some stuff against the wall and figure it out. So for those guys, what's your advice to them? You know, saying, hey, you know, if you want to take that approach, it's prudent to do that. It's good startup advice. Like I said, you know, your other choice is to go out of business, spend all your money and not get that B round or funding. But to those pros, saying, hey, if you're going to do that, here's how you can think about it. Is there any advice you can share with those kind of developers? I think the biggest thing is to think in terms of components and look for your patterns. All the patterns are there. It's just hard to see them. We get wrapped up in like the day to day of wait. This feature was released by that team and they don't talk to this team and you know, the sort of team dynamics and things like that and don't realize that actually these two things visually are exactly the same thing or some one component with a subtle variation. And so I think taking a step back and seeing those patterns also invest in the UI folks. Like often people have really good UI developers that just need a little bit of investment of working with someone who's a little more senior to help them kind of level up and then they do great and they can do amazing work. It's also the backend relationships too. Like, you know, having good understanding of the backend teams too. Do you see that dynamic also play well? It seems like there are these layers and they don't know how to talk to each other, no. Yeah, yeah, I'm a backend guy. We're totally fast. It's on Redis. Now what? Hire a UI guy. That's the answer most people say. But it's not that easy. No, it's a very, very hard profile to hire for. I just hired two amazing UI developers and I feel like I was so lucky that I managed to get them. Where do you live? I live in San Francisco. So that's a hard market to hire in too. It is, yeah, it is really hard. You must have had some good convincing. What can you share with the audience out there for the Velocity Conference? What is it about here today? Because, you know, it's one of those shows where it's about to bust out into the mainstream. We're already seeing some big names here from outside the community. Looking at the web performance, obviously, there's already these big names here, but like corporate guys, the enterprise guys are starting to come in. Which is a telltale sign of explosion of growth. But it's not a cloud show. Again, it's not, hey, it's not a cloud show. You know, it's not like people, you know, vendors flexing their cloud muscles or, you know, Amazon's not here. It's a different crowd. But it's a community. What would you share with the folks out there about what is Velocity? Oh, Velocity's deeply nerdy. On whatever level you're in, whether you're interested in DevOps, whether you're interested in scalability, whether you're interested in UI or performance or networking, like whatever you're into, you're going to find the deepest, nerdiest part of it here. So, yeah, I think it's great. And developers. You meet people who can just completely dig in. Yeah, yeah. The people are writing code. I was in the speaker room, I saw Theo in there writing code. People are talking in the hallways. Good hallway conversation. But again, that's the kernel community. What do you think it's going to take to move this mission up to the mainstream? Because I think that Velocity's got the right formula. They understand that the design side is really integrated with DevOps. So what DevOps did for network and ops and developers now is happening on a systems level with the UI. And you're on the UI side. So how do you see that evolving? I think it's already happened, actually. I think that ordinary UI developers are starting to talk to me about performance and they're thinking about it. And they're thinking about componentizing stuff. And they may not always know how to measure stuff. And I think that's where the folks who've been in the community a little longer can help out, is how to dig in, how to get the metrics you need, how to make decisions, because they're getting sensitized to that they should care about it. But yet, they need to know how to decide other than sort of hocus pocus rules-based stuff. We need to get people going from, I can apply these principles that you've taught me to I can think about what the next set of principles might be. And I think that's where the community really gets involved. Okay, Nicole Sullivan inside theCUBE, we got a break, we got a tight schedule because this show, we got a break early because of the breakdown here. But I want to ask you one final question and honest, what are you watching out there in the landscape of not just the velocity ecosystem of the geeks and the alpha geeks here and the nerds, but beyond as the world and society changes, we're all connected. Got the internet of things coming on. You got the iPod guy building an amazing tool for managing your utilities and so user interface is transcending just our interaction to computing devices to the home and other things. What do you see out there? What are you watching personally out in the landscape that's interesting to you? Well, honestly, I think I'm watching out to stay away from Google Glass guys. Can I admit that? All the revelations about, all the revelations about where people might be sneaking a peek into what we're up to and then having people walk around with cameras on their faces, it turns out not so cool with that. So I think that's kind of a big deal right now. Yeah. Very privacy too. Okay, this is theCUBE, we got a break, we got a tight schedule in the afternoon. This is SiliconANGLE's live coverage, Walter Wall be right back after this short break.