306: If You Want To Go Far, Go Together

In this episode, Steph and Chris talk about things they've changed their minds about over the course of their careers as software developers. Steph talks about as it turns out, arm chair rests are good, feature flags and comments are also good, she's changed her mind about how teams structure the work that each person is doing at once, and believes strongly in representation in the field.

Chris is not a fan up upgrading his operating system and when he first started out, he gravitated towards learning dynamic languages, and since then, much prefers functional languages, static typing or more broadly, static analysis. He also no longer believes in the 10x engineer, and also very much believes that URLs matter on the internet. So basically, don't call them single-page applications; call them client-side applications instead!

• Arq
• Karabiner-Elements
• Kent C. Dodd’s Epic React Course
• The Art of Code Comments by Sarah Drasner
• Gary Bernhardt: Functional Core, Imperative Shell

Transcript:

CHRIS: I still have dreams that I missed an entire semester of math class, and now it's time for the final. I don't know that I'm ever going to grow out of that.

STEPH: That's wild.

CHRIS: You don't experience that? It's a mixture of I'm in elementary school, but it's a college final. Like, the physical school that I'm in is my elementary school, but it's a calculus college course that I missed. And now it's time for the final, and I won't graduate college as a result. But it's also high school at the same time. Just every part of education sort of melded together into this nightmare scenario. Do you not experience that? I thought this was normal.

STEPH: [chuckles] Not in a very long time, not since I was in college. But I'm imagining this very cute, young Chris showing up with a backpack to the calculus final like, "Oh no." [laughs]

CHRIS: Yeah, pretty much, yeah. I really thought I would grow out of it at some point. But it shows...I think it manifests when I have anxiety about something else in the world, and then I have a math terror dream.

STEPH: That's your stress sign. That's your terror dream.

CHRIS: Apparently.

STEPH: Hello and welcome to another episode of The Bike Shed, a weekly podcast from your friends at thoughtbot about developing great software. I'm Steph Viccari.

CHRIS: And I'm Chris Toomey.

STEPH: And together, we're here to share a bit of what we've learned along the way. Hey, Chris, how's your week going?

CHRIS: Oh, it's going fine. Yeah, I'll go with fine. I had to upgrade my operating system. Enough things had stopped working or seemed to be pestering me about it regularly, which normally I'm going to ignore that for as long as I can. That's sort of where I'm at in the world these days. Like, I don't want to upgrade because I don't know what's going to break and whatnot, but then things had broken already.

Text messages were no longer showing up on my computer. And it turns out that the primary way that I interact with text messages is by replying to them through my computer. I don't want to type on my phone, that's not a thing. I'm already grumpy enough about text messages, to begin with, that I will regularly respond switching to email, and then I'll go off from there. But yeah, they stopped working, it stopped connecting.

And then I got this really weird message from Apple when I tried to sign in. And I was like, I feel like I should at least try to upgrade to the new operating system, which I think has been out for a long time, and I've just been ignoring it. But then I had the added problem of I didn't have enough space on my computer to install it, which I tried once before.

So I downloaded the installer, but the installer downloader doesn't check whether or not you have enough space to do the install. So it's just like, hey, so you know how you didn't have enough space? Well, we took up the remainder of it, and now you can't do anything about it. And the installer is hidden somewhere in the computer. So at one point, it just went away, and then suddenly had a lot of space on my computer. But finally, I decided to bite the bullet.

I found a bunch of caches on my computer. So there was a cache for my backup utility, which is called Arq, A-R-Q, which was a lot of space. It was like 20 gigs or something like that. So it was like, sorry, you have no more cache. I'm pretty sure my computer's going to light on fire the next time it tries to do a backup because it has no cache to rely on, and it's got to try a lot harder, pull a lot more data down. I don't know what it does, but whatever. It's going to do that.

And then, I found the more general application caches on the computer. Spotify had like six gigs of cache. Well, what are you doing? Aren't you streaming from the internet? Stop it. That's not okay. That is not acceptable. Yarn had three gigs. I was like, what is everybody doing? And I busted all of these. I threw away everything, and my computer seems to be doing fine after the fact. So, were the caches even doing anything? I ask.

Anyway, so I upgraded, and then some stuff didn't work. And so then I had to find the versions to make stuff work. The particular one that stood out was Karabiner-Elements, which I used to make my mechanical keyboard do the right things for the function keys. That stopped working. And I tried to upgrade it to the newer version because I figured okay; they probably hopefully released a new version, but it failed in the upgrade process.

And it turns out the secret was I had to upgrade to an intermediate version. I was on 12.3, and I needed to go to 13.4. But in between, I had to go to 12.10. And if I went to 12.10, then the upgrade to 13...everything about it was everything that I hate about upgrading software. It's like, I just know it's working right now, and I feel like if I even just look at it wrong, this whole tower of software is going to fall over.

The worst thing, the thing that I have not been able to fix, is now I use iTerm as my terminal, my terminal emulator as it were. And I typically run with transparency mode on which some people look at and say, "Wow, that's a choice." And I say, "I kind of like it. I don't know; it makes me feel like a hacker or something." I don't know, whatever. [chuckles] Let me live my life. But for some reason, switching to Big Sur, the version of OS X that I'm on now, iTerm doesn't have transparency anymore. And I just haven't been bothered to fix it yet.

But, man, I got rambly. I clearly have some feelings about upgrading software.

STEPH: You have so many feelings. The fact that you kept going...People can't see me, but I'm just dying because of that whole story. [laughs]

CHRIS: I kind of felt like I had to get through it. I had to exorcise the demons, tell my tale, and then be done with it, which I think I'm at now.

STEPH: When I start laughing that hard, [laughs] I try to hide from the camera view because I want you to keep going for people to listen.

CHRIS: But what's fun is you bob and weave. You'll hide for a minute, and then you'll come back and be like, okay, I'm composed, never mind. And then you'll just fade off to the side again. So yeah, but I powered through. [laughs]

STEPH: Oh, all right, there is so much there. [laughs] Upgrading is the worst. I agree with that. That was actually something I ran into earlier this week. Well, it was a mix of where upgrading presented a problem and then upgrading something else resolved that problem. And so that was an adventure where I shared a tweet. I can link to it in the show notes as well.

But Ruby was just taking up 100%, a full core, just all the time, and I couldn't figure out why. I wasn't doing anything with Ruby. We weren't talking at the moment, but it was just turning up one of those 100% CPU or higher. And so then I did some searching. And I did find the resolution, which was to upgrade the Listen gem because there was something in the Listen gem that didn't fully support Big Sur. Is that the name of the thing that I am on?

CHRIS: That's the new one, yeah. I know because I've just upgraded to it. I have thoughts on the matter. [chuckles]

STEPH: Cool. [chuckles] Yeah, when I upgraded to Big Sur. But then someone had kindly marched in to fix it, then upgrading resolved that problem. And Ruby is back to a peaceful level as to the amount of process, the amount of CPU that it should be taking up. Transparency mode, I'm thumbs up on it. I like how you called that out, how that's a choice. And I'm with you on that choice, although I didn't realize that's broken. I guess I just hadn't...I guess I don't care deeply enough that I've tried to restore my transparency, but you're telling me to hold on.

CHRIS: We're going to get realer now in this moment. So I have a very old version of iTerm because it has a different way of going fullscreen than the default operating system level fullscreen. I really hate that it animates to fullscreen, and it doesn't quite fill the full screen. Like, it still had a border around it or something. So I have a very old version of iTerm that I've been running with forever, and I refuse to upgrade in any way as a result of I want to cling to this old version of things working. But as a result, I think I finally hit the end of the road on that. This is like years running now too.

I remember I kept it in a Dropbox folder so that each time I upgrade or get a new computer, I'm like, okay, good. I still have my old special version [chuckles] of iTerm. But I think that time is over and I got to find...I feel like there are new terminal emulators out there. It's like Alacritty and other stuff that people talk about. So maybe it's time for me to try and find something new as long as I can get that transparency because I want to feel like an uber lead hacksaw.

STEPH: You have such a brand of new-new that I'm now discovering that you are also a software hoarder, so you have both in your personality. [chuckles]

CHRIS: There was a period early on in my software career that was like, oh, I got to find all this stuff. I got to figure things out and configure it. And then I was like, wow, that's taking up a lot of my time, I should stop it. And I think since then, I haven't upgraded anything. If you go look at my .files, I don't know the last time I pushed to them, but it's been a while. I'm still doing things, of course, but not as much. I know the cost of it, and I know the cost of maintenance.

And really, this is an allegory for software overall. This isn't just about our local development environments, but entropy exists in software. Software does not exist at rest, and it will decay over time. And so the idea of we've worked with so many clients where they're like, yeah, we're on Ruby 1.8, and it's Rails 0.9. So okay, all right, well, we're going to have to deal with that, it turns out. We can't just keep ignoring that. So really, it's the same story played out but in my local hoarder cavern.

STEPH: There was a part of the saga, the story that you shared with the installer and that you don't have enough space, and it took up the rest of the space, and you can't do anything. I'm very nervous; what happened to your stuff, your space? How did that resolve? [chuckles]

CHRIS: I finally bit the bullet. And so I have a bunch of...I've tried a bunch of the different pieces of software that will visually analyze your disk space. So they crawl the whole directory starting from the very root of your computer, and it will be like, all right, applications has this much, and the library directory in your home directory has this much. Here are all of the different places that stuff might be hiding on your computer. And then you can visualize and be like, okay, that's where the most of it is.

Node modules, as an aside, we did not choose an efficient way to approach how to put code on my computer because Node modules take up a lot of space on my computer, but they're so spread out. Multiple times I've seen people share a version of rm -rf, and then it's some subshell that does find every Node modules directory underneath a code folder. So you can find every single Node module and just blow them away. That will regain you some space. But that was not the solution this time.

I've tried lots of piecemeal solutions over time. But eventually, the thing that got me there was just busting all of those caches. So I cleared the backup utility, Arq's cache. I cleared a bunch of them, Spotify Yarn, et cetera. And that cleared enough space for the installer to actually run. And then, once that was done, the installer program itself was no longer around, so I reclaimed that space. But it was this weird chicken and egg thing where I had to have enough space to complete the installation such that the installer could go away.

And now...actually, let me see what my hard drive looks like now. So somehow, according to the Macintosh hard drive info, I have 50 gigabytes of available space, which is really frustrating because there were a number of weeks where we went into a Bike Shed recording, and I was like, I have one gigabyte. I'm not safe right now because this audio is going to be more than that. And so I don't know how now I'm sitting at 50. I guess all those caches that I cleared and the installer being gone probably puts me in a good spot.

But anyway, I'm living in an upgraded, wonderful world. As an aside, Big Sur is ridiculously rounded and colorful and almost cartoonish. They're really leaning into the iOS vibes. And I'm not sure it's my personal aesthetic, but that's fine. I spend most of my time in the terminal anyway. But I think that's enough of me ranting about upgrading my operating system, which apparently I had a lot to say about. But what else is up in your world, Steph?

STEPH: I do appreciate the ranting, though. You're not often grumpy, and when you are, it's quite humorous. [laughs] I really enjoy the grumpiness. And it's often a painful process. So I appreciate all of that story.

Something that I really need to share with you and get off my chest is a couple; I don't know, x number of episodes back, you and I were talking about computer chairs. And I bragged about the fact that I have a computer chair that has no armrest, and I love it. I love my chairs like this, and it's wonderful. And I just think it's the best way to live.

And it turns out that that's bad because I happened to go see a massage therapist who's also very well-skilled in physical therapy and other areas. And they were talking to me about my desk setup. And I mentioned the fact that I get these typical headaches, and I have my chair, but there's no armrest. And they're like, "Oh, that would do it." I was like, "Why? I like my setup. What's wrong with it?" And they're like, "Well, if you don't have armrests, then your back is having to compensate and to hold up your arms and your shoulders all day. So while you're typing, you're using more muscles to then hold that. And then they eventually tighten and contract, and then that can cause headaches."

So in case, I have led anyone astray into having no armrest, they are apparently very important to not having headaches or having your back overworked to the point that you have headaches, which I'm a bit sad about. But on that front, I have ordered a new chair, and we'll see how it goes. I will have to assimilate into the world of chairs with armrests.

CHRIS: We welcome you with open armrests. [laughs] Sorry, I saw it, and then I went with it. Anyway, I'm realizing now I actually don't use the armrests on my chair per se. I actually end up putting my arms on the desk, which is probably not ideal either. I have a little wrist pad so that my wrists are brought up and so that I don't have the upward breaking of the wrist thing going on. I think that matters a lot. And then my arms are supported by the desk, but it is just right on the desk, and I wonder if that's worse. But I've never...I don't know, getting the armrests just right and then also having the wrist pad.

But I can't adjust my desk is probably the main problem. If I could bring my desk down a little bit, and if it were a thinner top, then I'd have more flexibility. The chair that I have is wonderful and has flexibility. The arms can go up and forward into the side and lumbar and this and that. And so I'm able to make the chair work to the desk. But I do wish I had more of an adjustable...ideally, like a stand-sit desk. But I haven't made that jump just yet.

STEPH: When you're ready to make that jump, I'm going to share with you where I bought my desk because I'm really happy with it. And it's also not nearly as expensive as most of the other desks that will go up and down.

CHRIS: Presumably, we can include it in the show notes as well so that we share it with everyone.

STEPH: Definitely, yeah.

CHRIS: Otherwise, that's just kind of mean. [laughs] You and I have a weird back channel that we talk about on the show, but they're not actually put in the show notes.

STEPH: We're not mean. We wouldn't do that. I love my desk. And it was from someone else. They're the ones that shared it with me, so I'm happy to pass it along because it has served me well. And yeah, I'm also not sure about how this is going to work with the chair and the armrest because I'm just worried they're going to be too wide, and they're not going to actually offer support. I have doubts. I have lots of doubts, but I'm willing to investigate. And we'll see how this goes because I would like for the headaches to stop.

CHRIS: Good luck on that front. That definitely seems like an indication of worth putting in some effort there.

STEPH: Agreed. I also have some other exciting news. Stephen Hanson at thoughtbot has organized a number of other thoughtboters to get together who are interested in really diving into leveling up, learning React, and specifically focusing on purchasing the Kent C. Dodd’s Epic React course. And it's for anyone that is comfortable writing code, whether you know React really well or if you're new to it. Everyone's welcome to join.

So we just kicked that off today where we're going to go through the course together and then meet every Friday. I think the cadence is probably three hours, three and a half hours every Friday, that then we're going to commit to working through the course together.

And I have to admit, I always nerd out a bit over how does someone build a course? Like, I'm really excited about the content as well, but I just want to know how did someone go about producing this content and then sharing it with everyone? And then what's their outline? How do they help people that are getting stuck because they can't be there in the same room? How do they record their videos?

So I'm really excited to see all the ways that Kent has crafted this workshop. And so far, there's so much content, but I'll have more to report as we really start to dive in. But I'm excited to revisit React because I haven't been in React land for at least a year and a half; it’s been a while. And so it's one of those areas that I know some bits, but a lot has also changed. And I would like to just revisit that world. So I'm really excited to dive into the course.

And so far, I really like the structure that Kent has taken with the curriculum where we're focusing first on what exactly is happening and all the effort that goes into if you wanted to actually write HTML and then layer on JavaScript on top of that. But then here's how React makes that easier for you. Here is how JSX makes it even easier on top of the React API. I really liked that. Here's some pain; feel a little bit of pain, let's get a little bit better. And then let's get even better on top of that. And that has been a really nice reminder and progression into the course.

CHRIS: I'm definitely a fan of the way you're describing it like, feel some pain, and then let's get better. But then, like, what's the hook? With any educational content, this is the sort of structure where there can be full education. But this is the thing that I feel very deeply about conference talks is my goal isn't to teach you everything if I'm giving a conference talk; it is just to get your attention just to say, "Here's the thing, here's why you might care." And starting from the problem, starting from the pain is always such a good way to do that. Because you know how this stuff is hard? What if I had an option that was easier? And then building from that totally makes sense.

I want to say that course, Kent's course was built in conjunction with the egghead team, egghead.io. And it's a distinctly branded course. But it was built on top of the framework in the platform that's there and all of that, and then some of the editing support. I don't know this for certain, but I think there was some teamwork there.

And I love just pushing forward the envelope of how we do educational content in the world of development because it is such an interesting world that has, frankly, such a need for ongoing development. The world is changing out from underneath us every two days. And therefore, having great educational content is so important. So yeah, definitely interested to hear how your experience goes both with the course and then also diving deeper into React.

Well, switching gears just a little bit, I had a topic that I wanted to dig into with you today. And so to give some context, the topic, the thing that we're going to be talking about today is what have we changed our mind about? So you and I have both done a little bit of thinking and tried to come up with some answers to this. The background, this was actually inspired by a tweet that I saw between Shawn Wang, aka "Swyx" on the internet, and Charity Majors, a recent guest here on this podcast.

And Charity is someone who is known for having strong opinions. But Shawn asked the question of what are some opinions that you've changed your mind about? And Charity actually had a wonderful list, which we'll link to her tweet thread where she shared some of her both technical and then also more personal ones, but really talking about the sort of evolution of thinking and the way someone's thoughts can change over time.

And I thought it was just such an interesting thing because, for most points in time, we experience someone's sort of snapshot of where are you at now? What do you believe to be true? But I think there's such an interesting story and sort of the arc there of what did you believe to be true that you don't anymore? What have you softened your beliefs on? What have you strengthened your beliefs on? So yeah, with that as the context, what have you changed your mind about, Steph?

STEPH: Yeah, this one really got me thinking, and I feel a little stumped on it. I have a few that I'm excited to share. But I'm very excited to hear your list to see if that also helps me reflect more on some of the things that I have changed my mind about. And I have found that there's only a couple maybe that I feel like I've really solidly changed my mind about. The others, I've either dialed up the strictness, or I've dialed it down. So the ones where I've really changed my mind about are feature flags and comments. Those are two of them. Well, there's a third one, but I'll get to that in a moment.

So starting with the first one, feature flags I was more in the camp where I very much appreciate feature flags, but I use them sparingly because then there is a tedious nature of introducing them and then having to clean them up, and then having to maintain two states of code. But now I've really seen the value of feature flags and how we can make sure that we have calm releases and ensuring that main is always in a deployable state. So feature flags is one for me. I'm very invested in having more of a robust feature flag system because I see the benefit to that.

The other one was comments. I used to be very rigid about comments are bad. We should never have comments in our code. They are just waiting to go out of date, and they're not going to be helpful. But I have since dialed down that strictness where I have certainly seen moments where comments do feel very helpful, and I can see how people use them. I still want to avoid them for the most part, but I am less strict now in regards to people who really find value in comments. I'm more open to that discussion. I want to understand what it is they find helpful about that comment, and if it is something that we can't capture with code or a test, where does that live?

CHRIS: Those are both interesting. Feature flags, for me, I think I actually was more strongly opposed in the beginning. Earlier on in my career, I saw them as added complexity, as noise. I often would encounter them left behind in a codebase. And so, I had this negative association with them. And I didn't see the value; I hadn't yet felt that pain. And over time, I've definitely shifted to where you're at where I'm like, I love feature flags. This is a critical tool in our toolset of how we actually…like you said, calm deploys, being able to always deploy main, making sure that we don't have long-running feature branches. There are so many benefits that come out of it that I'm now very strongly in favor of them. But it's interesting; I think I would say that I started in a more strongly opposed place. So that wasn't on my list, but it's an interesting one that you've brought up and probably one that I've moved more on.

Code comments, I think, actually started in my career being like, obviously, you comment your code. It's the thing that I read about and stuff. And slowly, over time, I think I've just dialed in on I don't think we should be doing that. There are, of course, going to be exceptions.

And actually, one of the things that I discovered about myself as I was trying to go through this exercise is there are very few things that I believe are black and white. If anything, that maybe is one of the things that I've leaned into over time. It's like, nothing is binary. Nothing is black and white. Everything is on a continuum or shades of gray. There are things that I believe a little more seriously. But there's almost nothing that I can be like, nope, absolutely I will not equivocate on this beyond how we interact with other humans and being reasonable, kind people. And in terms of software practices, not really. Comments, though, are one that I still am pretty strongly not going to lean into. So it's interesting that you're like, eh, I've kind of opened up to that one.

STEPH: There's a particular talk, The Art of Code Comments by Sarah Drasner, and that's the one that really shifted some of my opinions around comments, and then how we talk about them, and what benefits they can play. But I will admit, if I see a PR that has code comments, I still immediately have a negative reaction to that. And I want to have a conversation around why that comment was added and if we can remove it, and how we can remove it. But even with that negative perspective, I still find that I'm more open to that discussion versus before, where I would have been like, no, that's just unequivocally bad.

CHRIS: I do like that you always bring up that talk whenever we talk about comments. This is a great talk. And in the background, I just looked up Sarah's Twitter profile because every time you bring it up, then I mention that she has a still from the movie Labyrinth in her Twitter background, but she actually changed it. And so now that's not true anymore. It's now something from The Force Awakens. Well, it's actually a joke, but I'm still going to suggest that you watch the movie Labyrinth at some point. That's the thing that I feel actually kind of weird about. It's a weird movie.

STEPH: I'm going to take your suggestion, but not watch it. But thank you. [laughs] To share my truth today.

CHRIS: That's fair, that's fair.

STEPH: What are some of the things on your list?

CHRIS: Okay, I have a couple, some more on the technical. Let's lean into one of the technical ones. Early on, I started with dynamic languages. I think I started with Python primarily and a little bit of JavaScript. I eventually found my way to Ruby and felt very at home there. And then, I started to explore functional languages. And I started to lean into them really hard and felt that immutability and functional programming and true pure functional programming was the thing. It was the answer, and I just needed to figure out how to do it. And so I would say that is the belief that I have since changed my mind on and decided, you know what? Actually, it feels like a bit of a force fit. I have tried. And maybe for others, it is actually a really fantastic way to build software. But having worked with a number of other people in more functional contexts, I find that it is a bit of a force fit. It's a bit rough.

And in particular, of late, I've been working with Svelte as opposed to React, and React does sort of lean into the functional paradigm, especially with Hooks and all those sorts of things. And it's a little bit rough because it turns out UIs are these deeply mutable things. We're changing values or typing things in. There are actions that are changing the state over time, and having a system that just more directly models that feels very natural.

I still love functional programming for the more core of an application. So again, I reference this talk often, but Gary Bernhardt's Functional Core, Imperative Shell. Gary has really formed some of my thinkings on this. And now I've started to find the examples in the work that I'm doing of like, oh, okay, I see that pattern actually applied here. But much as I would love to use them, the functional languages I find just aren't quite landing for me. And additionally, the mutability, particularly in the front end right at the edge of the UI, is not quite as good of a fit.

STEPH: So I think that resonates with me although I do still get very excited about following more patterns that represent more immutable state just because I felt so much pain and found bugs from the fact that we have mutated state in surprising ways. I'm honestly not quite sure how I feel about it. I'm going to have to think on that one. That's a very interesting one that you've changed your mind on.

CHRIS: Yeah, similarly, my feelings are lukewarm, whereas before, they were stronger. I was like, oh, okay, I think I found something here. And then, in attempting to use it across a wide variety of applications, it just didn't quite feel right. I felt like I was swimming upstream sort of thing.

Actually, there is an interesting counterpoint. One thing that I have leaned into and definitely changed my mind on and embraced is static typing or, broadly, static analysis. But I think static typing being the most pointed version of that. Early on, like I said, I got my start in very dynamic languages in Ruby, and Python, and JavaScript. And so that dynamic duck typing runtime can be anything. We just make our systems respond to the messages, and all of that sounded great.

But it turns out I really love having a compiler that can tell me some truths about my program before it ever reaches runtime. And the idea that a typo can make it to production feels absurd at this point. And actually, as I'm working in Ruby, I'm like, man, I really got to go look at that whole Ruby typing thing we got going on. I don't know what the state of it is. I've looked at it in the past, and I need to revisit it soon. But like TypeScript, I've definitely embraced that very strongly. And I would not work without TypeScript in a JavaScript project at this point.

I've loved the work that I've done in Elm, although that also sort of blends into the functional stuff where it's like, it was a little bit noisy, though, I'll say that. But the type system and the fact that the compiler can give you so much rich information about your program, I would not trade that at this point. And I don't see myself going back on that front, which is an interesting place for me to be on of actually, I'm not that into the functional programming as the core way that I build my applications.

But I do like static typing. And I feel like functional programming and static typing actually go together incredibly well. And functional programming and, more imperative, whatever it is that I'm doing with my day-to-day life these days is a more interesting fit. But it is interesting to me to observe that sort of combination of opinions where I really like static typing, and having a compiler, and something that can tell me about my program before I get to runtime. But also saying that I don't quite want the functional programming thing, or at least not as the entire way that I modeled my application because I found it a bit difficult to work with. Because I think static typing or compilers and functional programming go really well together.

But I think generally, what I'm finding is a more middle ground dynamic optimization of a bunch of different things. And the answer is like, well, it depends which I guess if you've listened to the show before, you'll have heard those words said, so I guess it makes sense.

STEPH: Yeah. All of that makes sense to me. And I can see why you might have a favor for types or why that feels more valuable initially because that is giving us so much feedback right off the bat versus following a more functional paradigm is something that could feel like more of a force fit and doesn't provide that same immediate feedback. But it has a longer-term or a longer cycle of that reward system. So I can see why you might favor one over the other or why I myself would favor one over the other.

CHRIS: How do you feel about types?

STEPH: I'm a big fan, although I say that, but I work in Ruby. [laughs] I don't have them. But when I have worked with types, I very much enjoyed it because it makes me think more about the design of my code in a way that I don't as much with Ruby. And working with types has heavy influence than when I am working in Ruby and thinking about the design of my code. So I think working with types is a wonderful thing that, frankly, all of us should do as developers at some point because it is so influential. So I'm for types, but I'm not using types in my day-to-day.

Another thing that I have changed my mind about is how we structure the work that each person is doing. So I used to be more in the camp of everybody can work on their own very complicated piece of codebase, their own complicated feature. We can have a bunch of complicated things in the sprint, and everything will just be great; it’ll be fine. And we'll get a bunch of work done, and we'll ship it. And then we're an even more productive team.

And I very much disagree with that now where I have found where everybody is working in their own silo on a complicated feature has slowed down the progress of then being able to ship that feature. Because we often want to collaborate with someone, we need to collaborate with someone. Then the PR review process is tough if I really have no idea what you're working on, and I don't have a context that then when I look at your code, not only am I evaluating at the code level, but then I'm also trying to understand the feature and gain all of that context. And that's a heavy cost for me to have to pay to then pick all of that up and then for you to have to reintroduce me to what's happening. Or I might make the bigger mistake, and I may look at your code and just evaluate it from the code perspective but not really understand the feature, the value that's being delivered. And that doesn't feel useful.

And I have a recent example where that happened where someone was working on a very complicated feature that I didn't have any insight into. So then, when I was looking at the PR, it was easier for me to just look at the code and get feedback on that. But then it was probably a day or two later. It wasn't until then that I finally started asking, what are we building? Like, what purpose is this serving? And that opened up a much larger discussion where we realized what was being built didn't actually really deliver what we needed to deliver. So I no longer agree with the idea that everybody should be working on their own complicated features independently, and there should be some collaboration. And, you know, it's the buddy system; we all need a buddy.

CHRIS: Well, I like that one. I feel like I've shared similar ideas where it made sense. It was just the efficient thing to do, to split the work up and have everybody very independent. I also feel like earlier on in my career; I was more scared of Git conflicts and things like that or people interacting with the same parts of the code. And so in my mind, it made sense to really strongly separate like, oh, you shouldn't even be touching the controller for this. I'll handle the views, and you handle the controller; it'll be separate. And I care less about that now. And I think what you're saying of like, it's actually better if we have some shared context, and we understand what we're working on, and it's more of a collaborative process. Yeah, I like that one. I think I followed a similar arc, and I'm at a similar place now as well.

Interestingly, to go into another one of mine that I think you'll probably be most surprised by on my list is I think I used to believe in 10x engineers. I used to believe in the idea of that one developer just off in the corner fueled entirely by Mountain Dew that would just produce the perfect code. They would just solve it. Over the weekend, they would write the entire billing system, and it would be great. And I think it was predicated on the idea that the coding is the hard part, which I no longer believe. I think coding at its core is communication. It's taking this thing that we want to be true in the world and then communicating it to a computer but also ideally communicating it to our teammates, and to future versions of ourselves, such that we can revisit that code, we can maintain it over time, other people can add to or augment it.

And so the idea of this loner that can just do incredible volumes of work and have that be a good outcome that just doesn't make sense to me anymore. I've worked with incredibly talented developers, to be clear, folks that I was sort of in awe of. I've worked with people who have, I think, just truly photographic memories. They seem to remember every single bug that they've ever had and exactly where they can look it up. Or from the top of their head, they can just intuitively know, oh, this bug means this. Go look at this line of code. I'm like, how did you do that? How did you do that magic trick? And they're incredibly capable developers. But at the end of the day, the folks that I see being most impactful on a team are the folks that are able to communicate and collaborate most effectively and make the whole team more effective.

STEPH: Maybe it's the Mountain Dew; maybe that's actually the secret sauce here. That's what I'm missing from my life to take me into that status.

CHRIS: I'm now imagining Mountain Dew but in a more viscous form, like a barbecue sauce, but it's Mountain Dew flavored. That's the secret sauce because it's a very…anyway, moving on. [laughs]

STEPH: It's a terrible product. We should make it and sell it.

[laughter]

CHRIS: Career pivot, we now sell Mountain Dew sauce.

STEPH: [laughs]

CHRIS: But yeah, I do not believe in 10x engineers anymore. If anything, I believe that that is a huge warning sign if you have anyone that's behaving in something close to that space.

STEPH: Yeah, I'm super interested in that you've shared because I don't think...We've talked about 10xers, but we haven't talked about the fact that you used to think that they were more of a thing and that they existed. And now it's all I'm sorry, but it's all crap. [chuckles] That's super interesting to me. Do you remember what changed your mind? Do you remember that pivotal moment of where you were like, oh, maybe this is all bullshit?

CHRIS: I think it was just an amalgamation of experience over time. I've encountered people who fit the archetype. But if anything, I would say they're deeply problematic in teams. They're that individual who refuses to collaborate, who just goes off and heads down, writes a bunch of code, but then it doesn't integrate with the other pieces, or no one else knows how to use it, or they won't let anyone contribute to it. And yeah, I've seen that just be very, very problematic.

So the folks that most fit, I think the imagined version of this, actually end up, in my experience, leading things astray. And the folks that are actually most productive and really cause teams to improve in a drastic way behave very differently. They're much more collaborative; they’re much more engaged with the team. It's less about their individual contributions and it's more about building a system together, collaborating, communicating, engaging external stakeholders, et cetera, et cetera. It's all that stuff that matters. And so, it's very much in contrast to what the 10x engineer ethos is about. But there's no one day where suddenly this idea that I had in my head crumbled when I saw that behind the pile of Mountain Dew cans, there was nothing there. [laughs]

STEPH: It's all a mirage. [laughs] I do like what you just said around that there are very impressive people out there. And those impressive people often focus less on their individual contributions and more at a higher level around communication. And then they are the powerhouses that then is helping facilitate everybody else be their best and have high levels of individual contribution. Those are the ones that...I'm still not going to endorse a 10xer, but they are the ones who, to me, embody the idea of someone that is incredibly efficient and really good at their job.

CHRIS: There's an adage that comes to mind here that "If you want to go fast, go alone. If you want to go far, go together." And that does ring true to me. I think an individual can have their individual productivity be higher if they're working entirely on their own, if they understand every line of code because they wrote every single line of code if they know where every feature of the platform is integrated because they wrote the whole thing. But they're going to be fundamentally limited. And in order to do bigger, more complex things, fundamentally, we have to work as a team. And then the way you have to interact just fundamentally changes.

So I think it started from that, like, one person on their own I think can be individually more effective. But the minute you start to have a team, that one person acting on their own is actually dragging the team down because other people can't then work in that space, and that's a problem.

STEPH: I really like that adage that you just shared where, "If you want to go fast, go alone. If you want to go far, go together." And that touches on something else that I have really changed my mind about, and that's representation. And this is more specific to me. So when I joined engineering and became a web developer, and I joined a team, and I was the only female engineer on that team, my initial feelings were I am the only female engineer, and that is fine. We're all just a group of engineers. We're here to solve problems together. It really doesn't matter if there's anyone here on this team that's like me. It's fine if there's no one that I can see myself in that's in leadership because we're all just people, is what I was coming down to. And I've completely changed my mind and realized that that's not true.

And I've experienced this where I've worked on other engineering teams with female engineers, and it's fucking awesome, and it does make a difference. And then when I can see someone that I can see myself in, in a leadership position, that is also inspiring. So that is something that I went in where I think it was more of I was trying to shield myself from the idea that I am different from everybody else in this room, and that could be a problem. And instead, I just tried to neutralize it by saying it's not.

But I think representation is incredibly important. People are not just people. We all have very important social and racial, and cultural identities. And it's very important that we get to feel that we can express all of those identities and see people that represent those identities in spaces where we would like to go. That's a big one that I've changed my mind on.

CHRIS: Yeah, I certainly agree that representation certainly matters, and being able to bring your full authentic self to work and seeing others around you that reflect that. And frankly, having teams that are made up of people that represent the users of the software that we're building feels so critically important. And it's very interesting to hear about the arc that you've had on that where initially, you tried to downplay it, but then you found a little more truth in it. And so yeah, thank you for sharing.

STEPH: You're welcome. It feels good to say that, too, because that's something that I've admitted and realized on my own, that that is something that has changed and shifted. But it's nice to be able to share that here with you as we're going through the things that we've changed our mind about. What else is on your list?

CHRIS: Well, to round us off with one more very technical version because, of course, that's where I'm going to take us after a much deeper and more nuanced topic that you led us on, single-page applications. Broadly, I'm opposed to the name; that’s a side conversation. But, man, URLs matter on the internet. So don't call them single-page applications, but client-side applications or whatever. Broadly, the idea of a bundle of JavaScript, and so you send down an empty HTML document, and then you reference a bundle of JavaScript, which that thing boots up and it then makes a bunch of API requests to the backend, and then it starts to fill in the page.

I was convinced for a while that this is a reasonable and perhaps even necessary way to build software. We need APIs for our mobile apps anyway. So if we're doing that, then let's have that be the consistent way that we are accessing information. This is going to be fine; it's not a problem. And then eventually, we found some problems. So then we got GraphQL, and we tried to solve it that way. But overall…and I have spent a lot of time trying to make this thing work, trying to find a version of this that I'm happy with that I find the end outcome of the software to be as pleasant to work with from an end-user perspective as a server-driven application, and I can't find it.

And so, to be clear, I'm still doing client-rendered applications these days. But Inertia.js is the framework that I've leaned into that helps me bridge that gap. And the idea that the server owns routing, that the server owns statefulness, things like that, not having to think about client-side routing, not having to think about client-side state management, being able to use traditional auth mechanisms built into cookies, all of these familiar things that we've had. Leveraging the fact that the server is the more privileged in terms of the information it has access to, the more secure, the more powerful environment, all of these things feel right to me. And the nature of the application that I can build just feels more robust, more consistent, easier to evolve.

There were a lot of promises that I heard when we started building applications in these ways. And I just haven't seen an example or have not worked on an example, at least of an application that is built as a client-side bundle that boots up and does some stuff and had a good experience with that. So Inertia, as an aside, is my answer to this. And I continue to be extremely happy with that as a solution, as really a middle-ground solution. Because going all the way back to true HTML server-side rendering is limiting in other ways that I didn't like. But I find that Inertia really strikes an ideal balance in the middle there.

STEPH: I feel like I completely agree with everything you're saying. But I also feel like I have a developer secret to share where I really haven't worked on single-page applications, and I am okay with that. [laughs]

CHRIS: It's fine, skip it. Just go straight to Inertia. It's better.

STEPH: Cool, cool, cool. I am working on leveling up React, and then the plan is to go to Svelte and Inertia. So I'll just completely...I'll skip that. I'll skip that part of my career.

CHRIS: I actually want to back up just a little bit as I'm saying this because I really try to avoid being in a more negative space. And I think this space, this architecture for building applications, is complex, and there are things that will warrant it. So things like Google Maps, it makes sense to have a lot of Dynamic JavaScript and to be doing complex things on the client-side. Trello is another example of an application that that as a server-rendered thing, doesn't really make sense. And frankly, using a tool like Inertia wouldn't quite work there. That said, that is, in my mind, truly a single page within the broader application. So the Trello board page is a very, very complex stateful application, and I think modeling it as such makes sense. Google Maps, similar. But there's still the profile page, and the login page, and all of these other things.

I think routing is probably where it breaks down for me. I think client-side routing is the thing that I feel the most pain on. Because at the end of the day, the server still needs to know the answer. And if we do client-side routing, we end up with this duplication of logic across the client and the server-side. We end up with disagreements from time to time. We end up with the weird flashes of half-rendered layout, and then we go to the login page because we get an API response that is different. And so, I think that is probably the kernel of the thing that I struggle with. And, of course, it is possible to build great things using any of these technologies.

But I think my summary is I've really tried on that front, and I've just not been able to make the fidelity of application that I want using…primarily; I’d say it's client-side routing is the thing that I struggle with the most.

STEPH: Yeah, it sounds like you're saying there are very valid use cases for using a single-page app or following that structure. But we haven't really gotten there in terms of our web development expertise, where we've made that easier to maintain and easier to implement. And there's still enough pain points around it that even though it seems like a very valid idea and approach, it still feels painful enough that you actively avoid it until it feels like something that you have to then invest in at that point to then really deliver the user experience that you want to provide.

CHRIS: Yeah, I think that's an accurate summary. And I think adding on to that, I’m noticing it becoming more and more of the standard approach; this is the way we build applications, and I don't agree with that. That is probably the thing that is the kernel of what I don't believe in. I think actually server rendering is a great way to start, and then you can slowly augment or move more things into complex client-side behavior. But starting with this as the mode that we're building our applications just feels like a less stable foundation than I would want. So it's perhaps an architecture that you want to evolve to at some point as the complexity necessitates it, but I definitely wouldn't be starting there. Similar to service-oriented architecture, not going to start there. Client-side routing, I'm not going to start there.

STEPH: Ooph. I feel like I've been holding my breath this episode. I feel like this was a very interesting topic that has been challenging to reflect on what we believe and what we've changed our mind about.

CHRIS: I think it's perhaps more nuanced than a lot of our episodes where often we're saying this is what we did, and this is how we felt in the moment. And that can be very experiential and true. But this, yeah, we had to draw the line in the sand and say what do we believe? I similarly definitely feel more tension in this episode than other ones. But hopefully, it was useful. Hopefully, folks found some value in the things, and hearing our story, also, the idea that we have singular formed opinions. Hopefully, this episode has broken that idea in anyone's head. And we're all on a journey.

STEPH: I really like how this has prompted me to reflect on the things that I used to hold dear and really cherish or follow strictly to then reflect on what are things that I used to believe versus what I believe now? Because that transition often happens so seamlessly for me that I don't really stop to think about it to be like, oh, something just happened that is really changing how I approach things, how I build, how I work with teams. And I really like this reflection point to be like, oh, what did I used to believe, and what's different today? I'd like to keep this practice going and just try to track the things...I'll have to make a list of all the things I believe. That seems like an easy list. [laughs]

CHRIS: Just the easiest list to write.

STEPH: The easiest list to write. And then I'll just check in with it every so often, scratch stuff out, or update it with the things that have changed my mind about. This is the good idea, terrible idea where you go, "Stephanie, that's a terrible idea." [laughs]

CHRIS: I don't know, write it down on a list, and then look at it in six months and see if it sounds like a good idea, and then we'll be able to close the loop on the whole thing. But with that, should we wrap up?

STEPH: Let's wrap up. I've got a list to write.

CHRIS: The show notes for this episode can be found at bikeshed.fm.

STEPH: This show is produced and edited by Mandy Moore.

CHRIS: If you enjoyed listening, one really easy way to support the show is to leave us a quick rating or even a review on iTunes, as it really helps other folks find the show.

STEPH: If you have any feedback for this or any of our other episodes, you can reach us @bikeshedor reach me on Twitter @SViccari.

CHRIS: And I'm @christoomey.

STEPH: Or you can reach us at hosts@bikeshed.fm via email.

CHRIS: Thanks so much for listening to The Bike Shed, and we'll see you next week.

All: Byeeeeeeee!

Announcer: This podcast was brought to you by thoughtbot. thoughtbot is your expert design and development partner. Let's make your product and team a success.

Support The Bike Shed

Episode source