Let's talk about Agile! What is it, what do we like, we do we not like?
In this episode, Steph and Chris discuss:
- Broadly, are they fans?
- What makes this practice work well?
- What makes this practice work poorly?
And also, hit specific topics and practices like Scrum, Kanban, and Extreme Programming.
- Twitter Poll re: Gotime Podcast - Is Agile's Time Over?
- The Mortifying Ordeal of Pairing All Day
- The Real Story Behind Story Points
- Agile Manifesto & Agile Manifesto -- Principles
- Extreme Programming Introduction
- Extreme Programming Explained
- Ron Jeffries - What is Extreme Programming
CHRIS: I feel like we should try a couple of different byes just so we have sort of a smorgasbord of options, and then we can pick the best one.
STEPH: With countdowns, [laughter] because I do so well with countdowns.
CHRIS: Hello and welcome to another episode of The Bike Shed, a weekly podcast from your friends at thoughtbot about developing great software. I'm Chris Toomey.
STEPH: And I'm Steph Viccari.
CHRIS: And together, we're here to share a bit of what we've learned along the way. So, Steph, I thought we would try maybe something a little bit different this week, a little bit more of a structured topic. In particular, I've been gathering little tidbits of information. I've been seeing conversations happen all around the topic of Agile, things that people like about Agile, things that people hate, mostly it's things that people hate about Agile. Lots of ire on the internet about Agile, but I think also some disagreement about what it actually means. And I think; generally, you and I are probably fans, so I want to talk about that. What parts do we like? What parts do we not like? What do we think Agile actually means or, at its best, maybe what it means? But yeah, let's start at the very top stuff. Steph, what do you think about Agile?
STEPH: I am generally a fan. I'm with you. And yeah, the internet being full of more negative remarks and ire, that sounds very true. But generally, I am very much a fan of Agile, and the very broad scope of this is how we work, and this is how we plan our work, and this is how we collaborate as a team, and then how we reflect on the work that we have completed. I can also pick apart some of the things I don't like about Agile, but in the broad umbrella definition, I'm a big fan. I've enjoyed that approach. Granted, I've also only ever used Agile. I haven't written software using a Waterfall style, at least not purposefully. And then if I have encountered a team that was using more of a Waterfall style, then we changed it quickly. I really only have known the more Agile approach to writing software.
CHRIS: I think that's largely true of me as well, where most of my work would fit somewhere under the umbrella of lowercase "a" agile, although I've tried variants of Scrum and Kanban and a bunch of other things that we'll probably chat about today. But I think in general, I find that things are most effective; things seem to move the most smoothly. And I think the software that we come out with is the best one. It's closest to those very simple ideals of Agile. And every layer of process that gets added on even though, like you, I've not done true Waterfall where it's like six months requirements gathering and then it gets handed off, and no one talks for a while. I've never done that.
STEPH: I have to interject because I actually think you have in a previous life when you were an engineer. You have done the more Waterfall. Like, you have to plan very far in advance.
CHRIS: I think this is one of those cases where people think "engineering" quote, unquote like mechanical engineering is one thing and it's actually...there is a little more structure, and there's a little more necessity of sequencing where you've got to figure out what you need to buy first because sometimes it takes a while to find the particular piece of metal that you need in the world. But it also has a lot of figuring out as you go and being like, well, we've got a bunch of stuff, and we're just going to figure it out. And also, this is something that as I was studying software while working as a mechanical engineer, I started to hear about this whole Agile thing, and I was like, huh, I wonder how I can bring more of that? Because I definitely saw cases where a more Waterfall-centric approach to engineering projects was leading to bad outcomes. It's like we decided upfront what we're going to do, and then we went away for six months, and we did it. And then we came back, and it turned out it was wrong. So that was solvable along the way. There were ways to build prototypes and things like that. So that is definitely a part of the mechanical engineering world.
Although I think there are some true constraints, but I think there are also some occasionally self-imposed constraints, but again, I see sort of the same thing in software. Anytime that we can shorten feedback loops, that's what I like. And I think that for me, that's the core of Agile. Specifically, to come to the Agile Manifesto, to start at the very top, the thing that kicked it all off is a very simple document that the first line of it is "We prioritize individuals and interactions over processes and tools." It's like, yeah, that seems like a great thing, having more regular conversations about the things that we're building rather than having those initial conversations. And everybody goes away for a while and tries to build that thing, and then they come back, and hopefully, the thing that they've produced actually solves the problem. But I think almost always there are some deviations like, oh, actually, it would have been better if it was like this or now that we're actually trying it in the field, it's fundamentally different. So in that way, I think there's actually a lot of commonality between mechanical engineering and software development.
STEPH: Okay, that makes sense. Yeah, I was thinking around the process of where you'd have to order stuff in advance versus for us; we can describe everything that we need as we need it unless we're having to procure some specific software or licensing. But otherwise, we don't have to wait on that shipment flow to then have our goods. And then, if we also mess something up, then we don't have to reorder more pieces. But I like how you started talking about that agile with lowercase "a" and then talking about the manifesto because I suspect most people are familiar with Agile, but it wouldn't hurt just to read off some of those top things about what Agile is so that way we're all on the same page together for this conversation.
So you already covered the first one that talks about individuals and interactions over processes and tools, and then the others are working software over comprehensive documentation, customer collaboration over contract negotiation, and responding to change over following a plan. And that's it; those are the aspects of Agile. And so then, circling back to what you were saying earlier where people are having more criticisms around Agile, it sounds that it's less about Agile, and it's often more about the implementation of these ideas and then how you're approaching them. Because, boy, do we have several ways to implement Agile. We have Scrum; we have Kanban; there’s Extreme Programming. Does that fall within the Agile umbrella? I think it does.
CHRIS: I believe so. And I think a lot of the things that people take issue with particularly come from Scrum and Extreme Programming. We're taken to their extremes. Yeah, it's right there in the name, so you should probably know that it's going to be a little out there. But taken to the extreme and especially where it becomes rigid and dogmatic, then it becomes a problem. But again, so we listed out now the four items that are the core of the manifesto. There is a separate part of the manifesto, which is the principles, which digs in a little bit deeper, but it's still very much in that same ethos. But I do want to highlight because there's a subtext to the Agile Manifesto that I really love, which is given there are things on the left and then things on the right when the Agile Manifesto was presented. And so it's like we like individuals and interactions, that's the thing on the left, over processes and tools. And so the subtext below it is that is while there is value in the items on the right, we value the items on the left more. And that's one of the things that I love about the Agile Manifesto is it's not this very rigid thing that says, "This is good, and that is bad," it is a statement of a preference of well, yeah, it's definitely good to have comprehensive documentation. That's a really nice thing to have, but it's incredibly difficult. And if we have to choose, we're going to choose working software. We're going to prioritize that well before we have comprehensive documentation. So I really love the juxtaposition, and the emphasis on it's not that one is good and one is bad of these two things that we're comparing but that we have a preference, and that we want to orient our work around the items on the left rather than the items on the right, which I think the items on the right are more traditional or were more traditional to the Waterfall approach.
STEPH: I like how you highlighted how those statements are presented to the reader. So then that way, as you mentioned, we still value what's on the right, but we favor more what's on the left. So one of the things that I saw recently was something that you shared with me in regards to where you're bringing up the idea of, like, hey, let's talk about Agile. And you shared with me a clip or a specific tweet that linked to a clip from the Go Time Podcast, which is a podcast that I hadn't listened to before. But I listened to that episode or at least part of it, and it's really delightful. I enjoyed listening to them very much. And they had Kris Brandow on the episode. And at the end of the episode...and they do something really fun where they ask the guests, "Do you have an unpopular opinion that you'd like to share?" And one thing that I like about their unpopular opinions is they often have polls afterwards, and they want to see was this truly an unpopular opinion? And if most people agree with you, then they actually consider it nope, he didn't win. I don't know if they use the word win if you didn't achieve the unpopular opinion. And in this instance, Kris shared the opinion that Agile is done and over with and that we should move on, which is a big thing to say. Everyone on the podcast reacted in a similar way that I would where it's like, well, how do we track things? And there are still things that we need to care about. But then also, there's a part of me that's just like, yes. I am not sure where Kris has heard it just yet when I heard that, but I'm already tuned in and very interested.
And one of the things that Kris said that also really resonated with me is he mentioned that "I've never worked on a team where Scrum specifically like sprint and story points functions well," and I absolutely agree. There are parts of Scrum, we can get to the specifics that I think are fine that I've certainly used in the past and that have worked. Story points resonate deeply. I very much agree that story points are something that I do not enjoy using, and I do not find that they really lead to building software. There's even a blog post that I published along with Matt Sumner, a former thoughtboter and guest on this show, where we talk specifically about story points and some of the concerns and issues that we have with using story points.
CHRIS: It was actually also the first episode where you came on as a guest to Bike Shed; that that was the topic that we dove into because it was so near and dear to our respective hearts.
STEPH: That's right. I forgot about that. So yeah, story points are certainly up there on my don't list. I feel like we're doing a fashion do and don't, but we're doing the Agile do and don't list. [laughs]
CHRIS: I kind of like that. Yeah, we should lean into that vibe. But yeah, continuing on with the poll there, it was interesting to see also, like you said, they tweeted out, and then there's the poll that comes after. And it was 64-ish percent of folks agreed that Agile's time is over and done with, and we need to move. Granted, it wasn't a huge sample size. It was like 85 people that took the poll but still, seeing both the statement and then also the general support from folks on the Twitter, it was interesting to see. So I do have the question of like, well, okay, if not, what else? And I share your sentiment of we should be able to ask questions and iterate. And nothing is so precious that it can't be replaced by something else that's better. So we always need to be trying to find the best ways to work. But again, I think there are still kernels of good stuff in the Agile. So I found this and I was like, oh, this is interesting. What's going on here?
STEPH: So I'd love to dive into some of the specifics around Agile to understand what are the bits and pieces that work for you and the bits and pieces that don't. So if we are taking our Agile approach and reviewing the things that do and don't work and changing that process, what are the things that you would keep, and what are the things that you would throw out?
CHRIS: Yeah, well, we can dig in, and we can bounce back and forth, I think, on this. But again, there are sort of a few different camps. So I collected together some of the lists of practices associated with some of the different approaches to Agile. So starting with Scrum, which I think perhaps is one of the most rigid, most structured, and perhaps most ire-deserving of the approaches to Agile, one of the first things is sprints or iterations, so the idea of starting...before you begin the work, you sit down, you define how much work you think you're going to take on. There's often an estimation process. Actually, we'll say that because that's maybe a separate idea, but even just broadly the idea of sprints and iterations, which often involve the idea of committing to a certain body of work. And that commitment is always handwavy and loose. No, no, no, we won't hold you to it, but then it's a constraint that's placed on the team. It's an expectation that's set, but it's wildly difficult to estimate software, as we all know. So sprints and iterations, personally, I am not a fan of. I really like a more continuous flow where we're constantly reprioritizing the work to be done. We're constantly measuring against what we built, what we think we need to get out there. How can we get something out in front of users as quickly as possible? But I've not found a ton of utility in the sprint or iteration workflow. But what do you think of that one?
STEPH: Yeah, I'm generally not a fan of sprints, and it has taken me a while to get there. And I feel like I can admit that openly because it is something that I feel like when I first started doing software development, sprints were life. It was how you planned everything. It was how you committed to work. It's how you measured your work. It's how you then looked back to see what you could and couldn't accomplish in two weeks’ time or maybe a week's time, depending on how long your sprint is. But over time, I have realized that I don't like the mentality of sprinting, and that may just be a nitpick on my part, but that is something that I don't enjoy because we write better software when we have breaks. And with the sprint methodology, there's really never that break unless you're going to plan that into your sprint.
And then there's the idea of the upfront commitment, as you'd mentioned, it's one of those, don't worry, we're not going to hold you to this, but can we all commit to this work? And it's one of those you just feel compelled to say, "Yes," to the person who's asking because then you feel like a jerk if you push back and you say, "Well, actually, I don't know if I can, so I'm going to commit to way less." And then that's the approach that I started taking of, well, I don't know. So I'm going to always commit to a little bit because I'd rather overachieve and then deliver more than come in under because I could work really hard, but I've over-committed and then still feel like I didn't reach my goals, and that's a rough feeling. So I found that I was already lowering my commitment there. So then, it felt more appropriate to be in line with that sort of continuous workflow instead of trying to commit to all these features or all these tickets that needed to get done. I think those are the two areas for sprint where it doesn't align with me and where it can work for teams. But I feel like there's always that underlining unhappiness that a lot of us just don't want to talk about because we don't know what else to do other than to keep sprinting.
CHRIS: Yeah, I think you said something about the specific, like nitpicking the word sprint, but I do think that's actually meaningful. It's The Bike Shed, after all; if we're not going to Bike Shed about some words, what are we doing here? But I do think that we're using that word...it's obviously the wrong word; this thing's a marathon. You can't have 26 2-week sprints back to back throughout a year. That's not going to work. That's not how humans work. But any amount that we let that thinking into our head, I think, is problematic. If I'm understanding correctly, it sounds like you've come to a place of comfort around committing to a smaller body of work and then ideally overdelivering. But in my experience, many developers, perhaps even most developers, don't feel comfortable. It's so difficult to say, "Yeah, I know that the login form should take a day. That's what I feel in my heart. But let's be honest, every other time we've done a form, it's taken a week. So I'm going to say a week." It's so hard to do that. And so I think continuously, we end up in a mode where we are failing to meet the collective commitment that we made, and that's demoralizing. That's going to constantly just be a drag on the team, even if they're fake, made-up deadlines that we're constantly setting, that we're constantly not hitting. Just doing that over and over, I think, is really detrimental to the morale of the team, to the cohesions, and the feelings of are we actually doing this work? So perhaps pedantic, but I definitely share all of that.
STEPH: I do want to highlight, as I mentioned earlier, I'm feeling more comfortable that I can under commit and then I can overdeliver, and that is hard. That is something that still in the moment, even today, is very hard for me to do. And it's like how you said, in my heart, I feel like this should take a day, and the heart lies. But on top of that, it's often it's also my ego that's driving me all the time. And with that, it feels like a competitive environment to me where someone's saying, "Hey, can you get this done?" And in the moment, that brings out my more competitive side where I want to say, "Yes, I can get all this done, and I can deliver all the things." When, in truth, that's often not how it's going to work out.
There is one thing I do like about sprints that I want to reflect on, or perhaps it's actually two. And one of them is that we are getting together every so often, and we're agreeing on the important work to be done. And I really like that planning process that is typically coupled with a sprint. So you get together, you review the work, you address any concerns or raise any concerns. And then you could say, "Yes, we all agree this feels like important work." And essentially, we're buying into the work that's getting done, and I really like that process. And then, as an extension of that, I really like how we often then pick themes. So as we are agreeing to the work, we're often grouping together work that makes sense where it's either the most cross-functional or collaborative. We're already going to be in that space together. We're aware of what everybody is working on. And those are the aspects that I really do like about sprint and some of the other styles, that more continuous workflow of where we're always pulling from a backlog. It feels more of a grab bag in terms of I don't really know what I'm going to get next. I don't know how this work has been reviewed or vetted. I haven't really gotten to talk to anybody, perhaps. I'm making some broad statements here. But I haven't really gotten to talk to anybody from the product side to understand this change. And I also don't really know what the rest of the team is working on, so I feel more disconnected from them.
CHRIS: Yeah, I definitely share that, the planning or the meeting where we discuss the work that's coming up and shape it a little bit; I love that. Although it's interesting within the context of Scrum, I think like truly to the letter Scrum; my understanding is there are very discrete meetings, and they each have a distinct purpose. And so there's the sprint planning meeting, there's a backlog grooming, there's a sprint review and the sprint retrospective. And each of those are these four distinct meetings that are happening once every two weeks or so or whatever your sprint cadence happens to be. And the splitting of those becomes interesting. And some of the practices in there, I think, are...I think you and I share not being interested in doing them or not finding them to be super valuable. But I think broadly having some version of hey, let's sit down and talk about the work before we have to do the work, definitely a fan of that. For me, it often can be let's collapse four of those meetings into one sort of thing and maybe have it more regularly or something to that effect. But actually, we'll touch on the rest of those. But if you're good with bouncing from sprint/iteration, I think we've covered that topic well. Let's move on to one that I think we can do pretty quickly because I'm pretty sure I know how we feel, but sprint planning/planning poker/estimation. How do you feel about this one, Steph?
STEPH: We grouped a couple of things in there. There's sprint planning, and then there's sprint poker, and those are different to me.
CHRIS: Yeah. So let's go specific to the planning poker as the most pointed version of it but also generally estimation and sizing of stories.
STEPH: Nope. Throw it out. I don't know how to play poker. Let's just get rid of it. [laughs] I was never a good poker player.
CHRIS: Playing poker can be fun, but planning poker...Well, so actually, to ask a slightly different question, I think in the past we've talked about keeping aspects of it, definitely not keeping the let's figure it out, let's hash it out. Let's get down to an exact point value, and then we know we can have 34 story points a week, and that's what we're going to do. But the version of using planning poker, using this numerical communication tool to see if we're aligned, that one I think we've talked about liking that. I have enjoyed that, but under the strict guidelines that we throw the numbers out. The numbers are only a communication tool. They get thrown out after the fact. We do not commit to a set amount of work or anything like that. We just use it to say, "I think it's an eight. I think it's a one. Oh, we should talk," just for that. That's when it's useful.
STEPH: I agree. Yeah, in my previous answer I was being flippant about it, but I do agree very much where I don't like the specificity of where you're trying to plan exactly what numbers are these. But I do find it very helpful for the reasons that you just said where the team agrees with the estimation around how long they expect something to take. Because then that is really great where you have someone who's never touched the codebase, and they're like, "I think it's a five or whatever system we're using here." It's an elephant...whatever scale you're using. And then someone else is like, "Well, I think it's a doughnut size." I'm making up silly stuff because it's more fun for me. And then those two people can talk and reconcile. So I do like discussing the estimation of work for that purpose but then not actually writing it down or maybe going with t-shirt sizes, something that's more simple, and then doesn't have anything with points, really. Anything with points can then be gamified and also brings out people's more competitive side. So, if you can make it something that's more fun, maybe around t-shirt sizes or a bunch of cute animals, various sizes, whatever works for your team. I'm trying to think of other fun measurements now [laughs] that we could use instead of t-shirt sizes.
CHRIS: There are the sizes of bottles of wine as you go past. So there's a regular bottle of wine, and then there's a magnum. And then it gets to weird names like a Nebuchadnezzar and other things. These are big performative champagne bottles. So I think we should use that kind of sizing because I think they also have a geometric progression type thing, not quite Fibonacci but something like that. So I'm going to make that push for Nebuchadnezzar as being my go-to [chuckles] sizing in story points.
STEPH: I have never heard of that, and I love it. That's great.
CHRIS: Okay. We'll find a relevant link to the wine bottle sizing, and we'll put that into the show notes. We will also, of course, include a link to your wonderful blog post. What's the story with story points that you wrote with Matt Sumner? Because I think that really does dial into this topic really well. And again, coming back to that core idea around Agile, while we see value in the item on the...which side is it? While we see potential value in story points, I have worked with countless teams who desperately wanted to make this thing work. So it would be great if we could quantify the work and then numerically understand the work that we had ahead of us and sequence things and talk about deadlines and whatnot. Man, that would be amazing. I would really love to do that. So with every other developer and every manager of a team of developers in the world, I have not seen it done. I am still looking for that day. When that day shows up, then I think this will be a wonderful practice. But unfortunately, my experience has been that this doesn't work, and trying to do it causes more harm than good.
STEPH: I agree that I certainly understand the reason that people want story points to work because it's very nice to then say, "We can calculate, and we can measure, and then we can have delivery dates." And that's really nice from a management perspective. But that does blend in nicely to the next topic, which I think fits nicely underneath the Agile umbrella, our daily syncs. Because that does bring us closer to that goal of where we can't give real valid updates on how something is going and provide a more real estimate as to when we think something is going to get delivered. That doesn't have the same effect of where we think we're able to plan and then promise delivery dates a week in advance because we're getting those updates in real-time, but they're going to be more reliable. And that is, we're so much more than where we try to over commit to work or if we try to say how much time something is going to take. And that is so much more valuable to have that reliable update and estimate versus trying to trick ourselves into thinking that we know when something is going to get delivered.
CHRIS: Yeah, I think the daily sync or sometimes called the daily Scrum, or standup, or otherwise morning meeting often in the morning, this is one that I see lots of folks really hate, and I'm personally a big fan of. This is one that I would definitely hold onto. But I think you have to be very, very purposeful with how you structure it. It really should be as short as possible. And there's one particular thing that I see very regularly in teams, which is almost a performative version of what I did yesterday. It's trying to demonstrate to the team that yes, I, in fact, did work yesterday. I was a valuable team member. Please don't let me go from the team. And I think that's the sort of thing that we should try and just get rid of. There are definitely times where what you did yesterday is relevant to the team, or you worked on something, and now you have a bunch of questions, and bringing that to the team is useful. But that version of everyone needs to prove that they did work yesterday or...it's the sort of thing like if anyone says that sort of thing, then everyone else is like if you don't say what you did yesterday, then it sounds like you did nothing because everyone else is saying what they did.
So you have to, I think, get a team buy-in to do this, say, "We're not going to talk about sort of bullet-list what we did yesterday. That's not going to get us anywhere as a team." But what's useful are those little magical moments of connection where I say, "Yep, I'm working on this. I'm going to implement it in this way." And someone's like, "Wait, wait, that way? Oh, we shouldn't implement it that way." And then ideally, what happens there is okay; let’s connect after this meeting. You've now made this connection, but you don't need to hold up the rest of the meeting for that. You can just say, "Cool, this connection has been made. That's an incredibly valuable little point in time, but now let's continue on with the flow of the meeting," so that it keeps that rapid pace. And so times where you're blocked, times where you have questions, times where you're just describing what you think you're going to be working on. So if anyone's like, "Oh wait, no, we needed to stop that work because we actually made a decision yesterday that impacts whether or not we actually wanted to build that feature at all." If you can head off incorrect work at the pass, there's so much potential value in that meeting that it is interruptive. And it does take up some time, but I find that it is so, so worth it if you're able to really keep it focused, keep it concise, and keep that end goal of those little connections. When those happen, they're so valuable. So I think it's really worth the input.
STEPH: I'm still smiling from where you said performative of what I did yesterday because that is something that took me a while to understand, one of the things that I did not like about the daily sync or daily meeting whenever your team gets together to talk about the work that's being done. And it was finally when I realized we're just going through a list of who has the longest list of the things that they accomplished yesterday. And again, it felt like it was bringing out more of that competitive mode in folks to talk about what they did, and it didn't feel very useful. Every now and then, maybe there was one thing that was interesting that someone did. But most of the time, it was always more helpful to hear what the person was working on that day for all the reasons that you just highlighted.
There is one practical concern that I have with these types of meetings or with these types of events. And it's where you'd mentioned where if we can keep it concise…and someone brings something up, and it starts to devolve into a conversation right there. So then whoever was up next is now waiting while that conversation is happening. And that part gets awkward because then there's usually one person who is then willing or no one frankly is willing to then say, "Hey, so sorry to interrupt, but let's actually table this discussion and let everybody else go, and then we'll come back to this." And if you have people on the team that have been there for a long time with that culture, then that will just work because everyone will keep each other in check. But otherwise, if you're starting that new process, or if you start to notice there's always that one person who's doing that awkward thing of trying to then set that culture of this is how we do our daily chat, and these are the things that then we wait for later, it's really hard. And I say that because I have often been that person that's in that space where then I encourage people to table a conversation. And it always just feels awkward to interrupt someone and ask them to please wait until everybody else has gone.
CHRIS: I share your hesitations around that, but it is very important. And it's that sort of ideally someone in a more senior position will model that behavior and model it in a positive, friendly way. Where I have done that often it's in the form of a question, so it's, "Actually, do you think maybe we could take this offline?" or something like that. Not a command, not taking over or shutting people down because it is somewhat interjective, and you're sort of correcting course. And so, being as friendly and empathetic in that moment as possible, but that's a hard note to strike. And again, if it's something that only one person is like the taskmaster, the Hermione Granger of the team who's trying to keep everyone focused and doing their homework sort of thing, nobody wants to be that. Well, Hermione did, but otherwise, nobody wants to be.
STEPH: I love all the Harry Potter-themed references that have been coming through in the last couple of episodes. And I agree it is something that's hard to help teams course-correct, but it's important, and it's very much something worth doing. I just recognized that I think that's why these roles get implemented, why there's this concept of a Scrum Master, and then why we designate these tasks to specific people because then you have someone who can do it. And then when they do interject, it feels more appropriate because that is their role, and that's one of the things that they're supposed to do versus putting it more on the social pressure of whoever is comfortable speaking up to then course-correct. So I do understand where that implementation of Agile has then tried to create those roles, which I've been on teams that have a Scrum Master. And my experience is it's often been a very positive experience because the person that is in that role is often very kind and caring about that team. And so they are a wonderful person to work with, but it's also one of those...I've also been on teams without them, and things have been fine. So I have mixed feelings about that one. It's one of those; it feels like an extra heavy process, but I've also been on teams, and it worked.
CHRIS: It's interesting the way you frame it, of the utility of that role. Like, having a role where we've now all bought into the idea that this person may take these actions say, "Hey, can we take that conversation offline?" and rather than one individual choosing to do that. I like that framing. I share what you're saying about the rest of the baggage that comes along with having this formal position, and often, that person is otherwise removed from the work. That can often be an aspect of Scrum. I think that gets complicated. But now I'm wondering can we make a software solution to do this? Because, of course, that's where my head goes. Can we have a standup bot that is listening and is like, "Hmm, it seems like you two people have been talking for the past two minutes. I'm just going to interject like my little bot self that I am and ask maybe take this conversation offline," in the way that we've sort of automated a lot of code formatting things, and that's been really wonderful, so that's not a part of PR review. Can we do the same for standup? I don't know.
STEPH: I think all the award ceremonies have these where they start to play the music, and that's your cue to move off stage.
CHRIS: Oh, I like it.
STEPH: I think that's it. [laughs] So you cue the music whenever someone has been going for quite some time. On a slightly separate note but still related to this, some conversations that have been bubbling up around me have been related specifically to this idea around stepping in to say, "Hey, I'll take on that thing that you need a volunteer for," or "Hey, I will help the team stay on track," will often fall on people with a specific personality and then they will often be the one that continues to do that. And so they will end up taking on additional work or taking on additional roles just because they may be in a more empathetic spot where they feel that's the kind, helpful thing to do. And so, we've been looking for more ways to make sure that those tasks are being distributed evenly across the team. So we're not just waiting on someone to say, "Who would volunteer for this?" And then typically being the same handful of people that are always speaking up and then volunteering for it.
And then trying to shift to more of a purposeful approach of having a queue of people and then cycling through that queue, and then if someone can't do it at a certain time, then we move on and then we just put them back in the queue. But this way, we don't have people that are typically just always taking on these responsibilities. And that's something that is a new consideration for me but one that I have found really helpful to be aware of and notice on your team who's the one that's always volunteering for these roles and checking in with them to see if they're comfortable with this, or if they're feeling compelled to volunteer for stuff because they may feel more inclined to speak up versus others are okay with staying quiet. But circling back to some of the Agile discussions earlier, you'd mentioned a handful of meetings and that you have some feelings about those meetings. What are those meetings that you have feelings about?
CHRIS: Yes, the meetings. So again, this is somewhat contextual to Scrum, but the structure of Scrum has a handful of meetings that sort of define the sprint. So you have some at the beginning, the middle, and the end. So there's sprint planning, there's backlog grooming, there's sprint review, which typically includes a demo for stakeholders, and then there's sprint retrospective. And these, as far as I understand it, are four distinct meetings and are intended to be kept distinct so that their purpose stays purified in each of those meetings. And I think my feelings would be that again; I don't really find a ton of value in the sprint structure or in the two-week cadence or things like that. And so I think it can make sense in those contexts to be like, we need to make sure we have space for these things. But in a more continuous context, I think the backlog grooming or, more generally, let's talk about the work that's coming up. Let's make sure that we're all unified in how we're thinking about that work, what we think matters, what's prioritized. I think that is an incredibly valuable meeting. I think sprint review and specifically demo for stakeholders I'm really intrigued by that one. I don't know that I feel like that needs to be a distinct meeting. And in fact, more and more these days, almost every feature I deliver has either screenshots or a screen recording of what that workflow looks like. So we're continuously demonstrating to the stakeholders what does this look like now that it's a real thing? What does an end-user see? What's that experience like? And in retrospect, I think we'll probably spend a minute on that one. I like retros; some people hate retros. Yeah, let's loop back to that. But of those, what are your thoughts? What do you like? What do you not like about those meetings?
STEPH: I think grooming is a very helpful meeting that can help a product manager and a technical team have discussions about the upcoming work. I don't necessarily think it needs to be the whole team. I think it can be a couple of engineers from the team; maybe those people rotate, maybe it's the team lead. And they get together with the product manager, and they essentially answer any technical questions about upcoming work. So then it can be refined. So then, as we get closer to that planning session, whatever we want to call it, then it feels more in a ready state for folks to react to and then have opinions on. So I do like grooming, but I wouldn't necessarily advocate that the whole team needs to be present for those.
For a demo, I'm with you; it really depends. I've worked on projects where the stakeholders are less close to GitHub and Slack and areas that we could demo some of the work that's being done, and maybe they weren't poking around on staging as much. So it was really helpful to then have a more formal demo to then show them the work that's being done. And then I've also worked on plenty of teams where a demo was something that we used as a fun internal event where we have all these different teams, and we get together. And then we get to show off all the great work that we have done across all the different products. So then us, as fellow teammates, can then celebrate what the other teams are working on.
Retros, you know I love retros. I think retros are a microcosm of your team's culture and process. And if your team is struggling to have a productive retro, your team is struggling. Because I think that is representative of your team's ability to get together, and reflect, share concerns, celebrate wins, agree on what's important, and run measured experiments. And if you're not having a retro, then I think you're not going to know how your team's doing until it's too late, and it's going to be harder to course-correct.
CHRIS: #HottakeswithSteph. I like it. I like the intensity that you came in with there, but I know retro is near and dear to your heart. So I'm unsurprised that that is the line that you've drawn. I definitely share all of those feelings, particularly around retro, because I think much like the daily sync, I've seen many people who are just like, "This is a bad meeting. It's useless. Nothing ever happens. I don't like it." And I'm often surprised by that because I've found so much value in it. Retro similarly is this magical meeting that can just regularly change the course of how we're working as a team. But I also have come into plenty of teams where it definitely did not have that shape, where it was basically a place that everyone sits down, and somewhat downtrodden restates their list of grievances, their airing of grievances, and then nothing changes. And much like the sprint iteration thing where you're constantly missing the commitments, and that's just going to wear a team down. I think if you constantly have retro and nothing changes and it's that same list of concerns, then that is going to be bad, but that, like you said, is not the reason not to do it. [chuckles] Oh, we just keep saying the same things in retro, so I don't think it's even that valuable. I would say that maybe we should change the things. But I've definitely been on plenty of teams where retro was just so valuable. And it's definitely one where I feel like having a facilitator, having someone who is in that particular seat trying to guide the conversation without necessarily being in the conversation, can be incredibly valuable.
There are also structures that I've seen work particularly well. We have a video on Upcase that we can link to. That's a format that I've found; it’s a very lightweight format, but it basically involves getting everyone's input on a positive note, on a more critical note, and then revisiting and sort of sorting and waiting, and then digging into topics that need a little bit more focus. But I think a lot of different formats can work as long as retro is a way for people to sincerely meet up, safely talk about the things that they are feeling about the work, and then ideally, some change comes about as a result of that. You mentioned having measured experiments, and I love that as a framing or like something that retro can do for us.
STEPH: I really do think that retros are so important because they're the health check of the team. As you'd mentioned, if people are having a very negative retro experience, which I understand, I've had very negative retro experiences as well, and I've walked away feeling like that was not a productive use of my time. But then that is our warning. That is our signal that's saying, "Something is not right, and something's not great, and we're not working together as we really want to be working together." And this retro is just that reminder that is right in our face, that is making this so uncomfortable and feel like a waste of our time because it is informing us that something needs to be improved upon. And we can feel like retros are not productive when we feel powerless to make that change. And that again is then another discussion to have with the team, to have with management, leadership, to talk about how do we get the power to then make the changes that we need to then have productive, happy retros? Because that's going to be a reflection that you have a happy, productive team.
CHRIS: Love it, love the framing, love the symmetry there between team happiness and retro happiness. So to summarize, I think we've gone through most of Scrum now. So just to...correct me if I'm wrong on any of these, but I believe sprints and iterations, nah, we'll leave it. Planning poker, definitely not. That doesn't seem good, although maybe just to bring up conversations, but not as an artifact that we save in any way. And then otherwise, daily sync, we're fans. Retro, definitely fans. Sprint review, backlog grooming, some version of those, a lightweight version of a bunch of the meetings seems may be good, but a couple of things definitely are going to leave on the cutting-room floor. Does that sound about right to you for Scrum specifically? We've got other topics to cover.
STEPH: Yep. All of that list sounds really good.
CHRIS: All right. So we've now found our refined version of Scrum, re-Scrum as we'll call it. But now there's a couple of other pieces...So Scrum is very focused on the ceremonies and the team activities, but there's another facet of the Agile umbrella, which is Extreme Programming, which that's a book. I believe Extreme Programming Explained is the name of the book. And there are various different links that we'll include to point at those. But there are two particular practices that stand out that I have heard some people love, some people do not. So we'll go into both of them. The first is pair programming. What do you think, Steph? Do you like pair programming?
STEPH: I do. I'm a huge fan. [laughs] Yes, I very much like pair programming, although it still has its limitations. I definitely want time on my own, and I can get exhausted from pair programming. It is a very vulnerable experience, too, where you have to share with someone: this is what I know, this is how I work, this is how I think. And I think that is incredibly challenging. I find that I am typically more productive when I'm pairing with someone or when I have the opportunity to pair with someone at least every couple of days.
CHRIS: Yep. I'm definitely a huge fan of pairing. Although I think specifically to Extreme Programming, I think the idea is 100% pairing. I think you already spoke to this, but pairing is exhausting. And the idea of 100% pairing is I can't really even imagine that; even 50% pairing feels like an incredibly high bar to hold for any extended period of time. There's a recent article that was going around the mortifying ordeal of pairing all day, which spoke of one person's experiences getting deeply burnt out just going through that process. And so, as valuable as pairing is, it's definitely a tool to be used not all the time. That feels like a lot.
STEPH: That's a lot of Stephanie singing because I tend to sing a lot whenever I'm stuck or thinking through things. So that's a lot of singing that I don't know if the world wants.
CHRIS: I mean, based on all of the various Bike Shed intros that involve you singing, I think the world wants it. That's maybe one person's take. But definitely, something that you said of there's a vulnerability to it. And so many pairing sessions I've either been the one saying this or someone else that I was pairing with has said this to me, but they're like, "I swear I know how to type, just now that someone's looking, my hands don't work." It's like you're in a dream, and your legs don't work. You're like, I know how to run, I swear. But for some reason, my legs are made of jelly right now. Or you can't remember a particular method, or there's just something that happens, and so getting over that hump, getting comfortable with it, I think it is a skill and something to become accustomed to. And so, again, being conscious of that when you start doing it is super important.
STEPH: I don't know if this is true because I only have access to people’s thoughts when I'm pairing with them, and then they're sharing their thoughts with me. But I do feel like people tend to beat themselves up more when they have someone watching because then you feel the need to say, "Oh, I normally can type, but because someone's watching..." which is so true; that definitely happens. But those moments are some of those really great moments to then reflect on the fact that just because someone's watching us doesn't mean that then we suddenly need to beat ourselves up. And I don't know how philosophical that I want to get with this, but I feel like there are so many opportunities while pair programming to then encourage other people around us to be kind to themselves. That is one of the things that I have really benefited from pair programming is learning to be more kind to myself. And even if I don't know exactly what's happening or what I'm doing and I may not be as confident with someone else, I can still be positive and kind. Just because you're in a vulnerable space doesn't mean that you then need to be unkind to yourself.
CHRIS: Yeah. I definitely agree with the idea of being kind to yourself also, where you can, be kind to someone else who you're pairing with, especially if they're finding that they're like, "Ah, suddenly my hands don't quite work." But I have pretty uniformly seen that a pairing session may start out that way. And then as everybody kind of just relaxes into it, suddenly you'll see someone just kind of flying around their editor. And you're like, wait, what just happened there? That was so fast. I don't even know. And so there's just this comfort level that sometimes it takes a little bit of time to ease into. But yeah, so pair programming, broadly yes. 100%, oh, that's going to be a no, no, thank you, not that. All right, so one other practice that comes from Extreme Programming, which is Test-Driven Development AKA TDD. What do you think about that one, Steph?
STEPH: I feel like you're giving me lay-up questions here. For anyone that's familiar with us, [laughs] I feel like this is an easy one. Test-Driven Development is a thing. It's a thing that I enjoy. I don't always write tests first, though, so I don't always follow TDD, but I am definitely a fan of tests. So, I guess in that light, it’s not so much that I adhere always to TDD. I don't feel the need that I have to write tests first, but I have found that with practice, that often helps me write code where I have tests then help me write out the logic for my code. So generally, yes, thumbs up on TDD, but I'm also not terribly strict about it where if you want to write some code first, write some code first.
CHRIS: Yeah, I think I'm definitely in the mode where I like testing. I like Test-Driven Development. I can't always pull it off, frankly. It's hard. It is hard to know how to write a test in advance of the implementation that you're going to write such that the test will correctly constrain the system that you're about to write. That takes a couple of levels of knowledge that if I'm writing a Rail’s controller action form sequence, I can probably TDD that because I've done it so many times. But if I'm doing something that's a little bit more new, novel, less familiar to me, then likely I won't be able to pull it off. TDD is like a fancy move that I don't always have available to me. But I consider that whenever I'm in that mode like that's not oh, it's fine to just write the thing before the test. Like, I want to be able to do TDD 100% of the time. I'm just not a good enough developer, frankly. And I don't know that I ever will be because I always want to be working a little bit past the edge of my comfort. So it's a delicate line of when I will not use TDD, but wherever I can, wherever I do have that level of knowledge of the system and the frameworks and whatnot built up, I find it is a vastly more effective way to work. It's not that I feel cool when I do it. It's like I feel much more effective. It helps me stay focused and on task and get the thing done. So it's very utilitarian in that way but also not something I can always pull off.
STEPH: So, circling back to when we first started chatting, you were asking about Agile and then my thoughts about it. And having this conversation with you, I'm realizing, or I think I was already aware, but it's helping me re-solidify I'm very much a fan of Agile. There are specific implementations of Agile that I don't find enjoyable, and I don't find helpful to writing software, and I don't find helpful from the project management side either. But broadly speaking, I'm still very much a fan of the approach that we use generally for Agile, where we want to work in small deliverable increments, and then we also want to have the ability to change any moment what is the most important thing to work on? To me, that is the heart of following the Agile process. And I don't think that's going anywhere. Like, I don't think Agile's going to disappear. But I wouldn't be surprised if we see another implementation of an Agile variety of the things that you and I just shared and the things that we like. And so, I feel like most teams that I work with follow Agile within their own unique bespoke version. And we don't have to give it names because everybody's going to have their own custom version where they decide which process works for them and which one doesn't work for them. And that's what retros are for so then you can figure out which process works for you.
CHRIS: Once more, Steph on the record about her love of retro. I think the core of Agile, the Manifesto, those core ideas about small iterations, delivering value, staying close to stakeholders, all of that feels deeply true to me. And I would be really surprised if a year from now or two years from now I was doing something that was wildly different from that. But then each of the layers of practices on top of that to varying degrees I like or don't like. And I wouldn't be surprised if aspects of that were swapped out down the road. But that core, that idea of this is how we think about building software. I like that thing; that seems like a good thing. So I'm going to hold on to Agile for a little bit longer personally.
STEPH: Same. I still see Agile in my future. On that note, shall we wrap up?
CHRIS: Let's wrap up.
STEPH: Show notes for this episode can be found at bikeshed.fm.
CHRIS: This show is produced and edited by Mandy Moore.
STEPH: If you enjoyed listening, one really easy way to support the show is to leave us a quick rating or a review in iTunes as it helps other people find the show.
CHRIS: If you have any feedback for this or any of our other episodes, you can reach us @_bikeshed on Twitter. And I'm @christoomey.
STEPH: And I’m @SViccari.
CHRIS: Or you can email us at firstname.lastname@example.org.
STEPH: Thanks so much for listening to The Bike Shed, and we'll see you next week.
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