text stringlengths 0 2.35k |
|---|
And if we're talking maintenance - I mean, are we just bug-fixing, or is what we're doing trying to refactor in a direction such that you're sort of enabling further improvements? I don't have an answer; I'm throwing it out there because I feel like we have strayed from the question of what kind of work we're actually ... |
**Johnny Boursiquot:** You should have maintenance. It's in the name. You're trying to maintain it so that it can continue to provide value over time. The mistake we make is in thinking that we can keep projects green forever. That's not realistic. And I looked it up, the color between brown and green, by the way, I'm ... |
**sam boyer:** It's green that's seen some stuff. |
**Johnny Boursiquot:** Yeah, yeah, it's seen some stuff. Maybe the kids were out there and just ran over it and just pilfered it. You know, it's seen some things, but the hallmark of good grass - as long as we're talking about grass here; Gosh it shows a lot about this analogy. But if we're talking about good grass, on... |
You want that resiliency, you want your software to be resilient, and that comes through maintenance. You're not gonna get the everlasting new green. That's not what you should be after. You should be after resilience. |
**Ian Lopshire:** And here we're talking about not operational resilience, but codebase resilience. |
**Johnny Boursiquot:** Correct. |
**Ian Lopshire:** So what are the properties of that? |
**sam boyer:** I actually have some thoughts there... I was kind of reflecting on that earlier today, and I keep coming back to this idea of like bad projects being calcified, like they get rigid and hard to change, so throughout the codebase you don't know the repercussions of changing one piece, you don't know what t... |
**Ian Lopshire:** \[48:04\] I would certainly say that's one of the properties. I would certainly agree. I think that you can tease out more... I think last time we talked about failure locality as a property, as a good test. That not only do you have things that help you know when things break, but you know where they... |
I would say though that those things are hard to add after the fact. This is perhaps one of the main frustrations with brownfield codebases. It's like, "So wait, you want us to change 30,000 lines of code in order to redo an architectural abstraction that we didn't have and we're clearly surviving without? Now, what's ... |
**Kris Brandow:** Yeah. I think there does at the end of the day need to be some good justification. I think that's good, too... Like, if you're gonna rearchitect something, then at least make sure you know the value that you're gonna derive from it. It's not just because it's the new, shiny thing at the end of the day... |
**Johnny Boursiquot:** The interesting thing about this conversation is that as software veterans I think we have this intrinsic understanding of what it takes to build software. So for somebody -- like, if I put myself in the shoes of somebody who's new to this... Maybe you've been doing this for 2-3 years, you simply... |
For example, open and close principle - it's something I know we're all aware of, that is a way of basically saying hey, one aspect, one characteristic of maintainable software is classes, for example, should be open to extension, to modification. So if every time you need to add new business processes to your software... |
\[52:03\] So if you're new to this - and perhaps one of the things that we can do in the show notes is to provide some books and some articles and some references to some of these principles that have basically stood the test of time to help you understand how to build better software... And this transcends Go, it tran... |
Basically, I guess we're not telling anybody anything new here that hasn't been around for a while. I think what we're trying to say is that "Look, we're old hats at this", and we now more than ever understand the people who were writing about this stuff when we were coming into our own. Now we see more than ever where... |
So if you're new to this, there's help out there to understand this. You don't have to rely on us four telling you, giving you anecdotes about grass, and things, to wrap your head around this. \[laughter\] There's stuff you can learn out there. |
**Kris Brandow:** Yeah, I think that's a good place to end this part of the conversation. I have one question before we jump into the final segment of unpopular opinions... And that is, is Go a good language for big, messy codebases? And I feel like this can probably be its own episode... But I just wanna get a cursory... |
**sam boyer:** Yeah. Strong static analysis is the first thing that jumps to mind for me with big, messy codebases. If we're talking about being able to \[unintelligible 00:53:39.20\] It's difficult to understand the effects... It's not like static analysis is gonna cover you everywhere, but being able to just quickly ... |
**Kris Brandow:** Ian, Johnny? |
**Ian Lopshire:** I think yes as well... I feel like I always come back to the simplicity of Go thing... But even poorly-written Go and messy Go is pretty easy to untangle compared to where I came from in PHP, where the classes were autoloaded in and you kind of figured out where the files even existed... So I think it... |
**Johnny Boursiquot:** Thank God we don't have method missing... \[laughter\] You know how many times I've pulled my hair out doing Ruby and figuring that something was implemented somewhere deep up the chain and I couldn't figure out where to go to get something? Anyways, Go is -- I think one of the strong cultural as... |
I've been fortunate enough to have been doing this for a while and I've come across similar software written in different tech stacks, including Go, and I'm thankful for Go because of that, because I can understand these systems way better than I was able to understand these other systems written in other languages. No... |
\[56:05\] But yeah, I think the language itself, what it supports out of the box, and also the cultural, idiomatic approach that we overall have as a community - those of us who do Go, who have been doing Go for a while, hopefully these are things that we can import in the new generation, because these days we have way... |
**Kris Brandow:** I will add at the end, as a funny thing, Gophers also live in the dirt, so... Of course Go's great at brownfield projects. \[laughter\] They love being in that dirty grass. And with that, we move on to our final segment of unpopular opinions. |
**Jingle:** \[56:43\] to \[56:58\] |
**Kris Brandow:** Alright... Sam, you're up first. Do you have an unpopular opinion? |
**sam boyer:** Oh, boy... I'm trying to formulate this one. We'll see. The most important part of GitOps is not Git and not Ops, at least depending on how you define those things. The most important part is the transform. You have objects on disk, and you have them in Git, and you make a change to them, you push them u... |
**Kris Brandow:** Okay. Any thoughts on that, Ian, Johnny? We just leave it there? |
**Johnny Boursiquot:** No qualms there... |
**Ian Lopshire:** Aren't the transforms the most important part of any software-related thing? Like, that's all software is, is take data, transform data... \[laughter\] |
**sam boyer:** Quiet! Quiet! You're giving away the game! \[laughter\] No, that is a good point. Truly, that is what we do, is we just take some bytes in and then we transform them in a bunch of ways until things come out on the other side. So I think actually it is fair to say that Git is not the important part of Git... |
**Ian Lopshire:** Just to side-track for three seconds... I hear this term GitOps thrown a lot these days... I have no idea what it means. \[laughter\] |
**sam boyer:** It's DevOps, but with Git. We've been arguing about this, discussing this internally recently. This is why I offered it as an unpopular opinion. Write your objects in Git, and whatever your objects are, use some kind of tool, whether it's Terraform, or some Kubernetes loader type thing, which then reads ... |
**Ian Lopshire:** Gotcha. |
**sam boyer:** And the main, obvious advantage that you get out of that is by putting them in Git you have access to a pull request workflow that everyone is familiar with, you have reproducibility... You have a couple of really important properties that just come from having the infrastructure as code. So then you jus... |
**Ian Lopshire:** So a fancy new name for what a lot of people were already doing. |
**sam boyer:** Kinda... Yeah. I mean, I think it's really just as is often the case, and I'm sure that I will get plenty of disagreement, or having missed some key aspect of this. But this is why it's not tied to a particular tool. It's not the domain of -- it's really just, like, you've got Git, there's objects in it,... |
**Johnny Boursiquot:** \[59:59\] Yes, infrastructure as code, and the code live in Git. |
**sam boyer:** Pretty much. |
**Kris Brandow:** Alright. Ian, do you have an unpopular opinion? |
**Ian Lopshire:** I'm horrible at this, I really don't. My last one was like 90% popular, so... |
**sam boyer:** Was that like the biggest humble-brag ever? \[laughter\] I mean, I don't know... |
**Johnny Boursiquot:** "I'm tapped into the people!" |
**Ian Lopshire:** No, it just means I'm behind the discourse. Like, I thought this was not popular, and now everyone agrees, so... |
**Kris Brandow:** You are at the end of the spicy train... |
**Ian Lopshire:** I'm dated. |
**sam boyer:** You're giving away the game. You totally won that one. I'm trying to give it to you... Just take it. Take it and run. \[laughter\] |
**Johnny Boursiquot:** I've got one. |
**Kris Brandow:** Okay, go for it. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.