045: The New Theory of Teams with Sarah Mei
In this episode, Sarah Mei, founder of RailsBridge, Director of Ruby Central, and Chief Consultant of DevMynd Software, talks about the way we write software: What’s right? What’s wrong? How can we do better?
In this episode, Sarah Mei, founder of RailsBridge, Director of Ruby Central, and Chief Consultant of DevMynd Software, talks about the way we write software: What’s right? What’s wrong? How can we do better?
The conversation examines changing code and reassessing needs. i.e.: "Does it bring me joy? Should I get rid of this thing? Do I understand this code?" She also talks about what these needs mean for others on a team.
Sarah Mei: @sarahmei
- Sarah Mei: How We Make Software: A New Theory of Teams @ Brighton Ruby 2016
- The Life-Changing Magic of Tidying Up: The Japanese Art of Decluttering and Organizing by Marie Kondo
CHARLES: Welcome to the Frontside Podcast. I am Charles Lowell and with me is Robert DeLuca. We have a very special guest this week. One that I'm really excited about because the things she says and the ideas that she has - open eyes and minds all over the place, in all different types of areas that are so pertinent to the way we do our jobs. So, we'll get to it. Our guest today is Sarah Mei.
SARAH: Hi. Thanks for having me.
CHARLES: Like I said, we are super excited to have you here. Before we get started talking about some of the things that you've been thinking about recently, why don't you just give like a very brief introduction of how you got started with development, where you've been, and how has that brought you to where you're going right now?
SARAH: You know, I actually was not one of these people that got started with it real early. I came to programming in college. I was an Engineering major. I wanted to build bridges. I wanted to be a Structural Engineer. I want to build things. I had a weird schedule the first couple of quarters of college, so I ended up taking an elective earlier than most people take it. It was a programming class in Fortran that was required for the structural engineering program. I took my class and I was like, "This is really cool."
CHARLES: Wait, Fortran is what set the hook?
SARAH: Yeah, and the professor of the class was like, "Well, if you think Fortran is cool. I've got some other stuff that you might like." I mean, the language and whatever doesn't really matter. What I liked about it was the fact that I could build something. I can get that same feeling of building something that you get if you build a bridge but you can do more than like one or two in your career, like you do if you're a structural engineer. I like the constant feeling of building. That's what I liked about it.
So I ended up switching my major and graduating with the CS degree and coming out and doing a bunch of different things, mostly like starting in a large company and sort of doing smaller and smaller companies over time.
CHARLES: Yeah, there's a lot of people in the industry who are career switchers, where they started out in something else and moved into Computer Science but I actually feel that a lot of people, like myself included, I have the degree in CS too, but that was not what I set out to do at all. It totally derailed, like the course of my life in a good way. But in that way, it’s like a career switch within a career switch.
ROBERT: I'm a little odd in that aspect. I came out of high school like ready to go in software. It worries me a little bit for the later half of my life. I'm like, "Oh, am I going to do software for the entire time?"
CHARLES: Probably not.
SARAH: That might be a good thing. You'll never know.
CHARLES: Yeah, seriously, what lies ahead?
ROBERT: Who knows?
SARAH: I feel like in a lot of places that are like, for example, in public policy and in other places where we need more people that understand tech so if we can send you out into other parts of the world knowing a whole lot about programming, that can only be good.
ROBERT: Yeah, this is actually kind of funny. I was telling CHARLES about this the other days, like I'm starting to view programming more as a tool to do the things that I really want to do and less as like the thing that I'm going to be doing forever. I wanted to augment and make things that I have a passion about easier.
SARAH: Yeah, absolutely.
CHARLES: Yeah, it's like software is eating the world so what you're doing now is just learning how to chew.
ROBERT: That's a great way to put it.
SARAH: You should tweet that.
CHARLES: All right. Please continue. I'll ignore the typing sounds.
SARAH: [Laughs] Switching careers is a really interesting thing because you end up with a bunch of experience that you wouldn't have had otherwise. I'm really excited actually about the next five years as we have all these folks that switched into programming from something else who are all becoming mid to senior level because they're bringing just such amazing experience from other parts of the world.
CHARLES: Yeah, I know, right? It's like, "Where've you guys been my whole career?"
CHARLES: It's like you understand these things, just almost like it's second nature of these things that are opaque and completely inaccessible to me. So anyhow...
SARAH: That's how I got here.
CHARLES: So then, after you kind of switched in college, you went out and did you just start working in programming immediately thereafter?
SARAH: Yeah, I worked in a bunch of different product companies. I built products for a while. My first job actually out of college was at Microsoft up in Redmond and then I have worked at smaller and smaller and smaller companies. Then I spent about 10 years doing product stuff and then about 10 years ago, I switched into doing consulting mostly because I realized that I have a fairly short attention span for projects. And that working on a product, there wasn't anything wrong with me exactly but what would happen is when I was working with a product, I would get six months to a year into it and I'm starting to get antsy. I started to get bored and decided that I should just embrace that. And I switch to something where I am going to be on a new project every three to six months. I've been a lot happier since then.
ROBERT: That's interesting. I wonder if that comes with seniority in software development and knowing your way around because consulting for me is I've gotten the experience of, "Oh, wow, I'm just finally getting a hang of this person's product or this client's product or app or whatever we're building," and it's, "All right. It's time to rotate off." It's like you just get in there and understanding everything.
SARAH: There is that aspect of it for sure but even when I was much less experienced, even with my first couple of jobs, I noticed this tendency in myself to just get bored after six months on the same code base. For a long time, I thought it was because I'm not cut out for software or maybe I'm not very good at it or something. Eventually, I just realized now actually, it's just that I just need to switch projects. I'm just one of those people. That's how my brain works.
I get a lot out of switching projects because the one that I switch on to, I see an entirely different way of doing things like code bases are so different. Even if you look at a hundred different Rails apps or a hundred different Ember apps, they're all so different. So switching on to somebody else's app, I learned a ton just out of that switching process.
CHARLES: It sounds like the actual kind of studying the meta-level of the software is what really engages you and kind of understanding how the software came to be the way it is and not some other way. One of the factors that gave rise to that and kind of 'that's the problem' that really sunk its teeth in you, as opposed to individual business problem. Is that fair to say?
SARAH: It has certainly been interesting to see different business problems and to understand different parts of industries and so on. That's definitely part of it for me but what really gets me interested is the different ways that people organize their code and by how they make the decisions that they make.
ROBERT: Yeah, you get to see different problems that they've maybe put themselves into because of the way they structured something, which you wouldn't see if you wrote yourself but somebody else did and get to see, "Oh, I understand this pattern now." That's kind of been my experience out of it. I don't want to speak for you, but yeah, that's kind of how I've seen other client projects like, "Oh, this is really cool. I didn't think of a way to do this," and you get to experience many different things in many different ways.
SARAH: You get to see a lot of the tradeoffs. Like a lot of times in a single code base, what would happen is I'd make a decision or we'd make a design decision of some kind. Then I'd see how it turned out. But there's no way for me to see how it would have turned out if I did it the other way.
The nice thing about switching projects for me is just being able to see all of those tradeoffs, like the tradeoffs that you make tend to be pretty similar. You can see very similar situations where people do different things and how does it turn out for them.
ROBERT: Right, and like one of my favorite things is where you go into a project that is totally against something, like for me it was object-oriented CSS and then you go in and you actually see it in practice, and you're like, "Oh, wow. This is turning a whole new light on it. I like this in this case."
SARAH: Microservices are like that for me, where it's generally I am anti the microservice bandwagon. But then I went on one project where I was like, "Wow, they actually figured it out. This works really well. I can see why people like it," because I've seen so many work that was horribly executed. When you go on to the one where it's good and you're like, "Oh, this is why people do that. Okay."
ROBERT: Yeah, it's like that light-bulb click, "Oh, yep. There’s another side of this."
CHARLES: Once you actually see it done right, it helps you avoid every other situation where it was done wrong and you can say, "Oh, this this was the one differentiator that made it all go right." I mean, sometimes it doesn't always boil down to that. But there's these one, two, three things that we could have done. But they were just completely and totally hidden from you because you didn't have that context.
I would love to talk to you about microservices because I've certainly never seen it done right. I've heard it talked about and I've seen this beautiful world, picture-painted that looks so fantastic on the whiteboard. But I see --
SARAH: Oh, it's so beautiful, isn't it? It's like an object-oriented design diagram. I'm like, "Look at all the boxes and lines. They all line up."
CHARLES: "They're beautiful."
SARAH: "I can do this in Visio," and they're all like, the line, they are on the same shape. It was great.
CHARLES: "And when I move this one over there, it just tells me that these two are exactly the same distance apart from that other one." Ah, so satisfying.
SARAH: Yeah, and then you try and do it, is the problem.
ROBERT: Then you build it and you cross your errors and everything.
CHARLES: Which actually I think that brings us, recently -- we're talking on Twitter. I think that's actually very recently about kind of the difference between when we talk about software and the meta conversations we have around it. When we do talk about these abstract and perfect worlds of boxes and lines versus the actual code bases, which is the things that you’ve kind of been observing many, many, many since you've started consulting, and kind of the vibe between those and you know what that means. I think a lot of people aren’t even aware like I certainly, before kind of reading that, wasn't really aware that that is a very, very distinct difference, like these are two very different modes for software. One that exists and one that is kind of perfect world.
ROBERT: Kind of academia versus the real world, I guess.
SARAH: In some ways, yeah. I remember when I was in college, we had a software design class as part of our degree program. We studied how you define objects and you write a little bit of [inaudible], like we did all this stuff. When I got out and I got into the real world and I had a job, I found it very difficult to actually apply that stuff successfully, to be able to draw a diagram and then turn it into code and have it work out the way that the diagram said it was supposed to work out. I initially thought that was because I was just not experienced enough to figure it out.
But eventually, what I realized is that it doesn't work because it doesn't work. It really doesn't work to design things ahead of time and then just do them. I think there might be a certain type of person that can do that. I am not that type of person and most people aren't. I think that it takes a very unusual type of brain to be able to just draw a diagram that has already taken into account all of the things you're going to encounter once you start making it.
CHARLES: Yeah, I would even go so far as to say there's probably a brain that solved that problem many, many times, that just could skip a bunch of steps.
SARAH: Right, and they're not aware they're skipping them necessarily. Unless you have an entire team full of that type of brain, it’s probably not a good idea in general, for the software that you're building as a group.
I feel like I've been trying to talk about that concept between the difference of how we talked about software in books, in blogs, and in conference talks and then how we build the software we actually build. I feel like I've been trying to articulate that for 20 years, like since I have my [inaudible] and I was like, "This doesn't work. Why can't I make a diagram and then make it into code?"
Like two days ago, I feel like I finally found a way to articulate it that captures everything that I've been trying to communicate and it was a really strange feeling. I'm like, "Wow, I finally kind of got it." One of the reasons that I came up with that, I think, is because I haven't really been thinking about it for a couple of months. I've been off and not really thinking about software stuff for a while. Oddly enough, I've been thinking about organizing my house for the last three months. All of my free time outside of my job has been thinking about like, "I've been learning how to cook, so how can I organize my kitchen so that the things I actually use every day, I don't have to dig through a drawer every single time to find them?"
There’s actually some interesting problems there like, "How do I make sure that all of the stuff that I need is at hand that I use all the time? All stuff that I need occasionally is still around and accessible, and then things that I don't use, I should probably just get rid of." I have this problem that I think probably a lot of people have which is that I have trouble getting rid of stuff once I have it. I live in a small apartment in San Francisco and that's not a good thing to be able to unable to get rid of things because in an apartment this size, I can let it go for a week or two maybe, but like I got to be very vigilant about it because otherwise, it just overwhelms the space.
CHARLES: Yeah, there's a bunch of research that the people estimate vastly different the cost of acquisition versus the cost of loss, and they’ve [inaudible] way too much, like irrationally unbalanced like not wanting to lose something that they already have.
SARAH: Even if I bought it for a need that I don't have any more or the need has changed or shifted. I don't buy things I don't need. There are some people that have that problem, that they buy a bunch of stuff that they don't have any particular plans for it. I don't have that problem, thankfully. I've had people in my family that have that problem which I think is why I have avoided that.
But the problem I have is that once something is here, I find it very difficult to get rid of it. I look at it and I'm like, "I can think of all these reasons why I shouldn't get rid of it." Oh, that was expensive so the sunk cost fallacy of like, "Oh, I paid a lot of money for that even if it's not useful and I don't like it, I shouldn't get rid of it." Or, there'll be like a dress in my closet that I haven't worn for two years and I'm like, "Ah, maybe I should get rid of it," and I take it out and I'm like, "Oh, my God. But it looks really good on me. I like it. I should wear this. I should really wear this." So I'm going to keep it even though I haven't worn in for two years for some reason, but I should keep it anyway because it looks good. I have all these stories. I tell myself about why I can't get rid of things.
A couple of weeks ago, I read a part of a book, to be totally honest with you, called The Life-Changing Magic of Tidying Up. It's written by this woman from Japan who's a professional organizer. Her name is Marie Kondo and her method is called KonMari. Basically, what it does is when you're trying to figure out whether you should get rid of something, you don't ask yourself, "Should I keep it?" What you ask yourself is, "Does this thing bring me joy? And if it brings me joy, then I keep it. If it doesn't, then I'm going to get rid of it."
So that made it really easy, going back to the dress example. I'm like, "Does this dress give me joy?" And I thought about it, I was like, "No, the reason I don't wear it is because I went out to dinner and I had a bad experience at dinner so every time I look at that dress, it reminds me of that experience." And so it looks good and everything but I'm not going to wear it because it doesn't make me happy. So that was just like, "Okay, fine. I'm just going to give it away." And changing that question that I ask away from 'should I keep it' towards 'how does that make me feel' was a huge change for me because it's like, that's really easy to answer, where 'should I keep it' is a much harder question. There's these bunch of sort of ifs and maybes or what-ifs and what happens. I feel like that applying this KonMari question to stuff has changed the way that I calculate what stays and what goes in a very positive way.
CHARLES: Yeah, boy, I need to get this book for several family members who will go [inaudible].
SARAH: Well, you know, I've got two kids and so there's a constant flow of stuff coming into the house. Because of the amount of space I have, there has to be a constant stuff going out. So this is something I just need to be very vigilant about and this has made it so that it takes up a lot less of my time and a lot less of my brain space, which is really awesome. It feels like it's moving my house in the right direction.
I've been thinking about that sort of in various ways, on and off, for a couple of months and I haven't been thinking about software. I have this fear that like, maybe that means I'm never going to think about software again. I go through these phases where I've got like, "Oh, I'm going to come up with a bunch of new ideas," where I'm coming up with new ideas for some whatever reason. Maybe I'm making new conference talks, I'm doing stuff, and I'm thinking about software a lot. Then I go through these phases where I don't do that, like I sort of retrench and maybe... I don't know. I think about other stuff for a while. So it's been home organization for several months now. I was like this, "I'm never going to think about software again," because it's just that --
CHARLES: Career change.
ROBERT: Oh, man. This sounds so much like my life since I moved down to Austin.
SARAH: You know, I live in San Francisco and I'm not 25, I'm 40. A lot of it is like maybe I'm just too old for software now. I should just give up and live out the rest of my career doing quiet, maintenance work --
SARAH: Somewhere. I don't know. Then suddenly, this thing happened on Monday, where I was just like, "Oh, code, an organization." And boom! There it was. I realized, I was like, "I basically just had to give my brain some time off," like my conscious brain needs some time off from software and it wasn't that it had disappeared because what I came up with on Monday was really just how home organization applies to code because I realized that the feelings that I get when I'm trying to figure out what I should do with code are very similar to the feelings that I get what I'm trying to figure out whether I should get rid of a thing. I look at this piece of code and I'm like, "Should I change this? Should I get rid of this? Should I refactor it?" You know, why I can't get rid of that? We just spent two weeks refactoring it so I can't change it again.
SARAH: We just put in a story for refactoring this and we spent three days and I can't go back to the [inaudible] people and tell them, "I need to change it more." Or, "I really like this code because I wrote it with someone that I really liked."
CHARLES: So I don't want to get rid of it.
SARAH: I don’t want to get rid of it because then I would lose the memory of working with, you know.
CHARLES: I actually can say that I have experienced that.
SARAH: Yeah, there's a lot of reasons why you don't want to change code. What I was thinking about, like maybe I was asking the wrong question, in the same way that 'should I keep this' is the wrong question when you're talking about stuff. Maybe 'should I change this' is the wrong question when you're talking about code. Maybe it's sort of leading you in the same way with stuff that leads you down this conversation of reasons that don't really have anything to do with the essential quality of why the code is there or why the thing is there. We need something that helps us reassess our needs. So if our needs have changed, maybe you don't need that thing anymore because your needs have changed.
Same way with code. If your needs have changed, maybe you don't need that code anymore, at least not in the form that it's at now. I think that question for code that, "Does it bring me joy," because joy is not something that I think is concrete enough when we're talking about code. I think the question for code is do I understand this? Do I understand what it's doing? Not just understand it like a very surface level of like, "Can I figure out what this syntax means?" But understand it more like the grok level of like, I understand this at a very deep level. I understand why it's here. I understand what problem it's solving. I understand why this abstraction is necessary. I understand how it got here.
CHARLES: Yeah, how it fits into the bigger picture.
SARAH: How it fits into the bigger picture, exactly, like the application.
CHARLES: How it fits in with like our conventions that are just purely stylistic.
SARAH: How does it fit in with the other stuff that we've been doing? How does it fit in with the product needs and the features we're trying to build and the business goals and all of that stuff, all of these different levels of understanding of why this code is here and what it does?
CHARLES: Do other team members' understanding factor into that? Like, "Do I understand the way that other people understand it," so to speak?
SARAH: I think that it can. But I think the important thing is whether you personally understand it.
CHARLES: Okay, like it's a very personal decision.
SARAH: I think it is. Hopefully, what you do is you want different people looking at the same code. You don't want just one person on a piece of code that no one else ever sees, whether it's pairing or code review or whether it's something else. It need to be really clear to someone is coming in and looking at that code what it does, what it means, and why it's there?
CHARLES: Right. I guess the reason I asked the question is because a lot of times when I look at a piece of code, I try and really step outside of myself and say, "What will someone else think who has never been on this project before?" Or, "Who is on this project and they see this code, will they understand it?"
SARAH: Absolutely. It's definitely a part of it when you're on a team.
CHARLES: Yeah, so I'm just trying to figure out how that question factors into this framework.
SARAH: I think that it depends a lot on how you distribute tasks. For example, if you work in a shop where you're pair programming most of the time, so there's always two people looking at a piece of code, 'do I understand this' is a reasonable question just for the two of you to consider, both from the fact that you can pool your knowledge but also from the fact that 'are there pieces of this that you understand that I don't understand' and vice versa.
On the other hand, if you work in a shop where it's more like, "Here's the piece of code that you work on like you own this section of code." Then I think it's more important for you to be able to step outside and be like, "Okay, do I understand this? Would other people on my team understand this?" That can be a very difficult thing to assess and that's where I think it's very helpful to do things like code reviews, call people in and be like, "Hey, can I run some stuff by you. I'm trying to figure out if this is good or not," because what you want is you want a code base that is comfortable and understandable for you and for your team.
Just like the thing that makes the KonMari Method powerful for stuff is that it doesn't tell you what you're going to end up with. It doesn't tell you what level of clutter versus cleanliness is good for you. It doesn't tell you. You either end up like something in one of these simple living magazines or end up something like Quarters, the TV show. There's a bunch of places in the middle, they're all fine. Everyone's going to fall somewhere differently along that line.
So I've managed now that I've thought about this a lot to set up my kitchen in a way that is very comfortable for me, like I know where things are, I can find them really easily, things that I use are at hand. But other people come in, they're just like, "I have no idea where everything is," like it's very personal. The organizational system you end up with [inaudible] that you have is a very personal thing and that's why, if you look at something like staged houses, so you're selling your house, you hire someone to put in rugs and furniture and stuff and make it look like somebody lives there so that people can walk in and sort of imagine themselves in this space, they don't put any of that clutter into the stage. They don't put any books on the coffee table except the big picture books. They don't leave the remote controls on the couch. There’s no plunger by the toilet. There's no like --
CHARLES: There's no Legos on the floor.
ROBERT: Everything that looks good.
SARAH: Everything that makes it more personal, they leave out because it looks like somebody else's mess. You go into something like that and you're like, "This is not my mess. This is somebody else's mess. It can't possibly be my house and I'm not going to buy it.
ROBERT: Oh, do we do this for software in conference talks and posts?
SARAH: Absolutely, we do. That's sounds very similar when you get someone new onto a project, especially if they're more senior and they'll walk in and be like this, I can't live like this.
SARAH: This is somebody else's mess and clearly we need to make some changes. But that's the reason why they leave it out of the staged houses is because you want people to be able to imagine their own level of clutter and disorganization that superimposed on the skeleton. But real life is not that. Real life is somewhere between that and hoarders.
There’s a very interesting parallel there with code, which is like when we look at code, if we look at the object-oriented design textbooks, you look at conference talks, you look at blog post, sample code, it's all very staged house. It's very uncomplicated. It has no clutter in it and that's because you're supposed to be able to look at that.
CHARLES: I mean, that clutter can distract the sales process so to speak.
SARAH: Exactly, like they have an idea they're trying to get across and the clutter would distract people from the idea. But the problem there is the same with the staged house which it's very difficult to tell what it will be like once you move in. It's very difficult to take some of these ideas that you see demonstrated in these staged environments and take them and apply them to your code base which is probably closer to a hoarded house than to a staged house especially if it's a code base that existed for a while over time, that has been worked on by lots of different people. This is the problem that I've noticed with a lot like there's some really amazing books about software design that have come out in the last couple of years. Of course, Sandi Metz's book is at the top of my list.
But the thing that people have trouble with, like they love the book. They love the book. I love the book. But then they find it very difficult to apply those principles when they sit down in front of a code base that has already been worked on for six or seven years, in some cases by maybe 50 different people, who knows, over time. How do you take those principles and start applying them in a way that moves you in the right direction? That’s where people are just like, "I can't do this. I can't do this and I'm not going to do this." And it's very similar to a problem where you've got a very dirty house and you don't know where to start in order to move it towards something from the Simple Living magazines or are more like a staged house, you don't know how to start to get it in that direction and so you just kind of give up.
The powerful thing about KonMari is that it doesn't give you like, "Here's what you're going to end up with it," but it gives you a way to get started on something that gives you a very easy question to answer. It moves you in the right direction. It moves your house in the right direction without being overly prescriptive about what you'll end up with.
CHARLES: Yeah, what that direction even is.
SARAH: What you'll end up with is personal for you, anyway. I think the question about 'do I understand this code' is similarly helpful and that it moves you and your code base in the right direction without necessarily giving you a lot of prescription about how you do it or where it goes or even where it's going to end up. It just gives you a question to ask that it tells you whether or not this code needs to change and a question is, "Do I understand it?" If I don't, it should probably change, and if I do, okay, we can just kind of leave it for now.
CHARLES: So now, if you're working on a team where you have two different people, maybe different skill levels, maybe just a different temperament or different set of preferences, what do you do when the answer to that question is two different things for two different people?
SARAH: Well, sort of like when you move in with someone. This is the hard part about living with somebody else, is that you have to mutually agree upon a method of keeping your house that is agreeable to both of you. Sometimes, when they say that working through a startup is like being married to someone, there’s some elements of that because you basically have to figure out like, "Okay, we're going to live in this code together. If we're going to live in this code together, we better both be happy with it. How can we both be happy with it?"
It involves usually, some compromise, like I really hate doing the dishes but I don't mind cooking and vice versa. You have to figure out. It really bothers me when there's socks on the floor but I don't care if you leave dirty dishes in the sink or whatever it is. You just have to have these conversations about, "What is going to make the code livable for you?" You basically want to end up with a code base that's understandable where all parts of it are understandable to everyone on the team.
Now that's like an ideal. You're not going to get there. But that's kind of what you're going for. If you have two people in the code and you have disagreements about what is the right way to go, sometimes it can help to just be like, "Hey, I don't really understand this," versus, "I don't think this is the right decision and here's why I don't understand this."
Sometimes, reframing the question in that way can prompt them to communicate reasons that they have for doing this that they maybe weren't able to articulate before, for example. Just like when you move in with someone, you need to have sort of this commitment to finding a level of housekeeping that you're both happy with. When you're working on the team, you do have to have sort of a mutual commitment to having a code base that everyone can live in.
CHARLES: Right. I like that because having like, "I just don't understand this and here's the reason why," that being a completely totally valid answer because sometimes in a code base, where someone's brand new or maybe they're at a more junior level, they don't quite have the tools to understand it or there's a lot of steps that haven't yet taken. It's like understanding is not going to be accessible to them immediately.
SARAH: And maybe that means that's the wrong decision for that code base, is that right?
SARAH: Because if something is abstracting to the point that a lot of people on the team don't get it, then it's probably not the right abstraction for that code base. That abstraction might be totally appropriate in a code base in which you've got folks that are more experienced who understand why it's there, who have the scars from previous times when they didn't do it, et cetera, et cetera, and they understand why it's there.
There is sort of like intellectual understanding of like, "Yes, object-oriented design is a good thing," and then there's, I would call it almost emotional understanding of like, "Oh, yeah, there's this time that we didn't do that and that worked out badly for us."
I think that folks that don't have the sort of experiential understanding, sometimes they just need to have that. They need to get that. Sometimes, what that means is you want to let them see what happens to a certain extent. Let them see what happens when you don't do that.
CHARLES: Right. This reminds me actually, I've got three kids and the way our house is now versus the way it was seven years ago is wildly different -- the way that we live. You know, with our first child, I'm ashamed to admit it, like our strategy was just to kind of put safety locks all over everything: every cabinet, on the oven, not on the refrigerator, but just kind of 'childproof' the house so that we wouldn't have to change the way that we lived but it made the house really uncomfortable for our children. And kind of having observed that over the course of having the second and the third, there's not anything that we childproof really. We put the dangerous chemicals way up high, where only we can get them.
It's a little bit more inconvenient if we need to access the bleach but that level of discomfort is something that we live with. We've always got cups that are set out on a cabinet that sits below the counter so we've got water cups set out so that the children can get water and stuff anytime that we want, and we try, for things that they're going to need, make sure that it's accessible if you happen to be four feet shorter. That's just a condition of who you are. So it means that the actual configuration of our house, even though it's the same house, is just radically different and it is more optimized or it's optimized as a compromise for the fact that there are people living in this house now that haven't learned how to operate everything but they just need to learn that the oven is hot and you don't go there rather than slapping a lock on it.
SARAH: Your house is probably more comfortable for you as a group, right?
SARAH: And what that means is that as the 'senior' in the house, it's slightly less comfortable for you in some ways but it's worth it. It's worth being less comfortable for you in order to increase comfort across the board for everyone in the household.
CHARLES: Right, because it means that if the child needs water, I don't have to stop what I'm doing to get a cup out of the cabinet and fill it for them.
SARAH: And they feel [inaudible] over the stuff in their house. They feel like they live there, like the house is for them.
ROBERT: That builds comfort and confidence.
SARAH: Yeah, I think that's a very good analogy. Anytime you have a group of people living together, everyone makes compromises in order for the house to be set up in the way that's optimized for the group.
CHARLES: Yeah. "So man, how are we going to apply this to software? What's the next step? What are the concrete steps?" I guess it's just asking those questions, like asking, "Did I understand it?"
SARAH: It is asking those questions and it's also, if you are one of the more experienced folks on the team, it’s your job to elicit the answers to that from other people that are less experienced. They’re not going to tell you. A lot of times, sometimes, they may or may not feel comfortable saying that they don't understand something. So it's your job to really try and figure out like, "Do they get this at a level that is acceptable? Do they understand why this abstraction is here at an intellectual level or at an experiential level, at an emotional level? Do they get it?" Which is not something you can really just ask. In many cases, it's your job to --
CHARLES: To just observe it.
SARAH: To observe and to see how it works. If people are having a hard time understanding where things are in the code base, it could be because everything is so cluttered that you can't see anything or it could be that everything is so hidden that you can't see anything.
It’s sort of the staged house equivalent where everything is too abstracted, or is it the hoarded house equivalent where everything is just obscure and under piles of junk. Either way, no matter which direction you need to move towards the middle, the question is always, "Do I understand this?"
ROBERT: I like this a lot. I keep on coming to the analogy of if you put a chef in a different kitchen where everything is just totally rearranged and they don't know where their knives are, where their measuring cups are and stuff, I think that plays perfectly in a software of like you put somebody into a code base that they don't know, "All right, I'll figure it out." It's not their home. It's not what they're comfortable with or used to. Yeah, I think this is making my brain work on how I can apply this.
SARAH: Or if they're moving in like when you hire somebody and they 'move into your house', you need to be ready for things to change. And this is one of the reasons why I've been saying for many years in ways that I think maybe didn't quite connect as well as they could have, that really the team is the code and vice versa. Every time you add someone to the team or someone leaves the team, teams are not mutable. You get a completely new team. So, it's not like you can just sort of carry on like you did before.
Every time you get someone new onto the team, everything gets reimagined, every breakdown of responsibility, every decision. You look at it in a new way when you have someone new come on to the team. If they're going to stay, like in your chef example, if this person is moving in and this is going to be their kitchen and they're sharing it with other people, then what you're going to end up with is probably something in between what it is when they get there and what they had before. They’re going to bring in some ideas, you're going to keep some of your ideas and you're going to end up with something in the middle. The same thing has to happen with your code when you bring someone new onto the team.
CHARLES: I really like the way that this just focuses the discussion and I know that you've talked about this a lot before, whether it's a kitchen or a house, this idea of the code not being so much the shrink-wrapped product. It’s a structure, yes. It is definitely that but it's a structure that you, as people, inhabit. It protects you from certain things and it provides you certain things that you need to live.
When people ask us why is a continuous delivery pipeline so important in automating all these things for deploying your software it's because the idea is this is going to be a living thing that your team will actually be living in. And every member of that team will be living in from the time they start with the company or start with the project until the time that they exit and the time that they leave. It's the actual living process that you want to make comfortable and pleasant.
SARAH: And what comfortable and pleasant means will be different depending on who's on that team? It’s not something that you can have like a --
CHARLES: It's not.
SARAH: Right. This is why all of these things are like, "Here's how you design things." It always seemed to fall flat. I think it would be better titled like, "Here's how I did one thing once."
SARAH: Or, "Here's what works for me." I feel like every conference talk that is about design could be, "Here's what works for me. I did this one thing once."
CHARLES: You might want to try it.
SARAH: You could try it. It might work for you, it might not, right?
SARAH: A lot of times where conference talks fall flat and blog post and everything else was why they're more like, "This is how you do it. This is the right way to do it." You're like, "Well, it certainly works for you."
ROBERT: The one time I gave a conference talk, the night before I went through every slide and scrutinized it as much as I thought somebody out in the public would do it. And I think that might be where we go through in a 'stage our code'. It's like we're trying to make it perfect for somebody that might come through and scrutinize it or criticize. Because I know when I was going up to put those slides up, I wanted to make sure it was the best foot I could possibly put forward.
CHARLES: Right, we don't want to be wrong but I think that's where it actually, thinking of it as 'this is what worked for me' and this is an example from my house that worked. This is a way that I organize my code and my space. That'll not take a lot of pressure off of not having like, "I am right and I'm an authority at saying that this is the right way." That's a lot of pressure.
SARAH: I don't even like that. I try and frame a lot of the things that I talk about as like, "Here's the thing that works for me really well. Maybe it'll work for you too. Let me know."
ROBERT: Yeah, that's how I give it.
CHARLES: Up until really about two years ago, I felt like that was the expectation that was put on people is to say the right thing.
SARAH: That's true. And I think that there's a lot of teams where that is an unspoken requirement and that's something that we should examine. Because even within a team like 'here's a thing that works for me or here's a thing that worked on my last project' isn't very different from saying something like, "Well, industry best practice --"
SARAH: And I think that like you get to a certain level of experience and people expect you to say things like that. In my experience, the best way to do it is 'blah'. I mean, it's not actually a super useful statement because your past experience may or may not be directly applicable to the thing you're looking at right now, no matter how experienced you are. I think that it's much more friendly to have a range of experience levels to say things like, "Well, this worked for me on this project. Let's talk about whether it could work here."
CHARLES: Right, yeah.
ROBERT: I really like that.
CHARLES: I do. It’s so hard because your human nature is to try and boil things down into a simple binary.
SARAH: People would love to have a list of rules, I'll tell you that. This is a problem. This is one of the reasons why I think it's important for us to come up with these questions that you can ask that will move you in the right direction without giving you rules, that will move you in the direction of finding the rules that work for you. Because rules themselves, people really, really, really want them. But they're always misused. They’re always misunderstood and what you really need are the questions that led you to those rules in the first place. That's what people really want, although maybe that's not what they are asked.
ROBERT: Ah, the Steve Jobs approach.
SARAH: [Inaudible] to start wearing black turtlenecks. I hate turtlenecks.
ROBERT: And New Balance shoes and the jeans.
SARAH: But yeah. I think it's one of those things where people are very hungry for guidance. But we've been giving them the wrong kind of guidance. We’ve been trying to give them rules. When what we really need to do is give them questions to help them develop their own judgment.
ROBERT: Right. Like when I was coming up, I thought, in everything, there was a right way to do it and a wrong way to do it. I've been slowly, sadly figuring out that it's not all black and white and it's not all just logic. I've always treated programming as like, "Well, they wrote this and it's just logic so I should be able to understand this."
It's been a long road to come to this conclusion that kind of like what you're talking about and this has been enlightening for me. Like you are going to solve your problems your own way, your own person, and you'll think about things differently. I really like the analogy of 'this is your house and this is how you work and live in your house'.
SARAH: Right, and no one would tell you in order to be a proper human being, you have to set up your house this way.
SARAH: We feel comfortable telling people, in order to be a professional developer, you need to set up your code this way. I think that those are very similar statements and we should really examine a lot of these 'should' statements that are all over the place when you're talking about software.
Think about whether or not they're actually serving us in our mission of doing more things with tech. Like overall, my mission here is for people to be more effective with code so that we can do more interesting things with it. I live in the TV show, Silicon Valley, essentially so I'm surrounded by these companies that are solving these little tiny problems and I'm tired of it. I want us to solve bigger ones.
In order to do that, we need to get better at coding. We need to get better at managing code over time and that's what I'm trying to do.
CHARLES: Because it's not going to scale, otherwise. We’re out of time. We’re going to have to have you on the podcast again because I don't think we've got to... what? About 15% of the things that we want to talk about?
SARAH: Oh, we are overtime, aren't we?
CHARLES: Yeah. But thank you so much, Sarah, for coming on and talking with everybody. You drop real quick your Twitter handle so that if people want to have follow on discussion, they can reach out to you that way, or your other preferred means of contact.
SARAH: Yeah, Twitter is probably the best. My Twitter is @sarahmei, and that's mostly where I am.
CHARLES: All right. Well, fantastic. As always, feel free to reach out to us too. I'm @cowboyd on Twitter. And what are you, Rob?
CHARLES: All right. It’s a wrap. Thank you so much, Sarah, and we'll see you in Ether and hopefully we'll have you on the podcast again sometime.