text stringlengths 0 2.35k |
|---|
**Natalie Pistunovich:** \[laughs\] |
**Angelica Hill:** Are you gonna do the same thing you did last time, when you caught us all off-guard with the many different -- |
**Natalie Pistunovich:** No, no, no... Just for those... |
**Angelica Hill:** ...many different ways to refer to it... \[laughs\] |
**Natalie Pistunovich:** It's crazy how many names are there for this... Yeah. For those who are listening now and have not listened to the previous episode, "The Art of the PR: Part 1", Angelica and I were discussing, among other things, also the many different names and concepts that represents. So Anderson, we asked... |
**Anderson Queiroz:** I do. |
**Natalie Pistunovich:** Do you review more, or do you write more? |
**Anderson Queiroz:** Now I write more. In my past job I read a lot more code. A lot of pull requests. And funny enough, I've never worked in a company that didn't have pull requests. To me, software development as a professional means pull requests and code review. That's my standard. |
**Angelica Hill:** Why did you do more reading in your past job? Is it just a very different area you were working in? |
**Anderson Queiroz:** No, my past company I joined among other things to help to lead the transition to Go. So they were pretty much a Ruby shop, and they decided to migrate to Go. They were migrating, but then needed someone with expertise in Go, to bring best practices, how to do... So I did that. |
I did a lot of workshops, and teaching and mentoring, and it involved -- a lot of the teams would come to me with pull requests, so I'd do a really extensive review, not only code \[unintelligible 00:06:49.25\] functionality, but also as an opportunity to teach Go and \[unintelligible 00:06:52.19\] the conventions, the... |
**Angelica Hill:** And then now you've moved to somewhere where Go is bigger, so therefore you don't have to play such a big kind of a reviewer role. |
**Anderson Queiroz:** Exactly. Now I am a software engineer, as a senior, but there I was a tech lead. I think when you go above senior in the tech lead and stuff you start coding less, reading more, writing more... Right? |
**Angelica Hill:** Okay... |
**Anderson Queiroz:** I mean, writing more specifications and documentations. |
**Angelica Hill:** So do you feel like there's a direct correlation between seniority and how much reading of PR versus writing you do? Like, if we were to plot that on a graph, how much you read PRs and how much you write... Could we, to generalize the industry, say that the more senior you are, the less you're gonna ... |
**Anderson Queiroz:** I think the more you are senior, you're gonna look for more things. So because you're gonna be able to evaluate -- if you're good in the language, you can evaluate the language and the conventions, you can evaluate the general software architecture. When you're in a role that as a senior you are a... |
**Natalie Pistunovich:** \[08:21\] That's an interesting question, yeah... Because sometimes in Go they have this graph of contributions that is no longer just the squares, but also what type to bring more - to you make more issues, do you review, and so on. |
**Anderson Queiroz:** Oh, yeah...! |
**Natalie Pistunovich:** So if you can look in that, then you kind of know what people's roles are, or you see how it changes over time, or something. |
**Angelica Hill:** If I showed you a load of graphs, would you be able to guess the seniority of that engineer? |
**Anderson Queiroz:** Perhaps... I think I looked quickly at mine and now it's 50%. |
**Natalie Pistunovich:** Nice. |
**Anderson Queiroz:** Pretty much. |
**Angelica Hill:** Do you think that that's the balance that most engineers would like, or do you think that there is such thing as someone who prefers to be reviewing more than writing? |
**Anderson Queiroz:** When you think as a software engineer that the patch is about writing, writing... Right? They write code... I think at least either you want or you need to pass knowledge forward. So you're gonna need to read. I really like to review code, for both reasons. One... Because, I mean, I have a passion... |
**Angelica Hill:** So you say you try to give feedback, you don't wanna overwhelm the person... How do you do that? Is it that you limit yourself to "Okay, I'm only gonna put six comments"? Is it the way in which you phrase your review? How do you make sure that you're not overwhelming? |
**Anderson Queiroz:** Yeah, I learned to experience in feedback. I put tags. So I start a PR with suggestions, or sometimes I put questions... But a question is a question, right? So it feels a bit redundant, but I put. And then I have blocker... And sometimes the suggestion is less Go convention. Or depending on the w... |
Suggestion is exactly that - I believe it can be better. If I were writing this code, I would do it different, but you don't have to take it. Sometimes I put like init. It could be just "Erase this blank line between the function call and the error handling." This is init. Or a typo... So these things. And then if I've... |
**Natalie Pistunovich:** So you say that in some situations it kind of makes sense to just speak in person instead of writing a lot, for example. |
**Anderson Queiroz:** Sometimes it's easier. I think sometimes speaking is better, sometimes writing is better... And on pull requests, writing on the pull request is a lot slower. Sometimes it's worth jumping on a Slack chat; it's already enough. Sometimes we talk. |
I think nowadays at Elastic, because we are distributed, it's a lot more through Slack, rather than really a call... But sometimes - yeah, I've jumped on a call, for small things. |
**Natalie Pistunovich:** \[12:18\] And how do you decide when it's better to do this and when it's better to not? Do you have a thumb rule, or is it all just feeling? |
**Anderson Queiroz:** If there is a lot of back-and-forth, it's easier to jump on a call. Or if it's something that I really want to understand... But most of the times I always \[unintelligible 00:12:32.13\] on the chat. Now, Elastic is a distributed company, so we write a lot more. So it's a lot more common just to w... |
I think jumping on a call is more personal, so if I'm close to someone, it's easier. It feels more natural to jump on a call. And if you don't know the person so well, or something, you kind of end up just chatting. |
**Natalie Pistunovich:** So you say there's actually three types of giving feedback. One, writing in the pull request, two, writing on Slack, and three, hopping on a call. |
**Anderson Queiroz:** Yeah. On Slack - I use it more to clarify. Because if it's specific about the code, I'd rather have it in the pull request. I kind of document it anyway, right? I try also to explain why what I'm saying... As I said, in my past company I was really with this job to teach, to mentor in Go; I would ... |
**Natalie Pistunovich:** "Why did you butcher my pull request?" |
**Anderson Queiroz:** "You are guessing. This is your opinion." And I was like, "That's true. We don't have like a metric to say that." But also it was common sense. It was like "This is too big", you know, in the company. But this guy was a new-joiner, so everything was different for him. |
So again, clarifications - I can do also in the pull request. But I think it's important to document what's happening in the code, there. |
**Natalie Pistunovich:** In the code itself, to kind of document, to make it self-explanatory. |
**Anderson Queiroz:** Exactly. |
**Angelica Hill:** So in terms of -- you talked about giving feedback, not overwhelming... Do you feel like PRs are also a good place to kind of -- especially for more junior engineers, give them like props on things they're doing well? Like, "Oh, I really like the way you did this thing." Or "Oh, this is great." Or "T... |
**Anderson Queiroz:** Yeah. I think it is. It is important, and it's something that I would like to do more as well. I mean, I never go to a pull request looking for errors or to try to diminish someone's job. But at the same time, I go to a pull request looking for errors, so they don't go to production. |
At the end of the day, you're looking for problems in issues to prevent things bad from going to production. But I think it's super-important to do the praise, to assert when someone does something nice. Sometimes someone just fixes something a bit random, but it's the same function. It's not really going away from the... |
**Natalie Pistunovich:** So we said that there's some stages that you escalate the communication through some way, and definitely you want to include more positive feedback, that is not "Correct this", and so on. Or explain yourself. What other changes would you make to the PR process based on pain points you have with... |
**Anderson Queiroz:** It's a really slow process, because you write the code and then you submit the PR. And then someone else - sometimes more than one person - they have to stop and read it. And how do you synchronize that? And then I think it's gonna depend a lot how companies do... I've been working for different p... |
\[16:04\] Now I'm definitely overwhelmed by GitHub notifications, so sometimes it just slips through, some PRs, and days later someone tags me, "Anderson, can you give a review?" Like, "Oh my God, sorry." |
**Natalie Pistunovich:** Because it's something that -- kind of a bit of a ping pong, or does it happen more with new pull request? |
**Anderson Queiroz:** It happens when you have to re-review. You do the first review, and then you have to look again. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.