5 ms·
Hi! Early Agile person here, but I hate what the term Agile has become. [1] I think there are a lot of good reasons to grumble about Agile, especially when mana
by wpietri 10y ago
Hi! Early Agile person here, but I hate what the term Agile has become. [1] I think there are a lot of good reasons to grumble about Agile, especially when management use it as a stick to beat developers. But I don't think the problem described here is one of them.
I think expecting business to know what they want is a mistake. When my dad was making software in the 60s, 70s, and 80s, we were mainly automating existing business process, so we could draft behind all the hard work done working out the right paper forms and reports. But modern software comes from a process of experimentation and discovery, competition and evolution. As the people who understand the technology best, we should be as part of that process from beginning to end.
This analogy is misleading because software gives us power that we don't have in the real world. If we start with port-a-potties and want real bathrooms, we just have to change the description of a bathroom and the computer will do the rest. In software, if you are disciplined about automated testing, refactoring, and keeping up with technical debt, an evolving architecture is totally doable.
I think where people fall down is that they don't believe that managers will give them enough time to keep things clean as they go. And I agree, they won't. You have to take the time. But if you do it from day one you can do it gradually and incrementally. You can gradually redirect management pressure away from developers and toward making better product choices.
So please let's take control of our working conditions back. We're professionals, and we can say, "No, I'm still writing tests. That's a best practice, and it would be negligent for me to leave that out." But we can do that while delivering incrementally, so that no week goes by without the business getting some new benefit in production.
[1] http://agilefocus.com/2011/02/21/agiles-second-chasm-and-how-we-fell-in/ http://agilefocus.com/2011/02/21/agiles-second-chasm-and-how...
- existencebox 10y agoLet me give two stories. I am in full agreement with your statement, however want to give supporting evidence as part of a "I'm not sure how to push back well even after all these years without compromising a lot of political capital". 1. Just this week, I watched a manager wait until after I was on vacation, and then publicly chew out my employee for slipping a deadline. I would agree he should have pushed harder, BUT he slipped it due to our peers being very rigorous on his CR (he is a new employee) to bring him up to standards. (I would add we would have had even MORE buffer had the person in question not continued to change the spec 2 months into a 3 month project, which can happen, but should be acknowledged as a cost) . The manager in question found this (the deadline slipping) regardless unacceptable and gave him grief in front of the whole team about it, and this is the second time I've watched the manager place these meetings when I'm going to be gone because otherwise I come down like a bag of bricks on someone going after my employees. But to why I mention that last sentence regarding my nature, NO AMOUNT of this seems to motivate positive change in the long run, _especially_ since I'm "managing up" and having to do this as an engineer myself, any big "push" I make costs a lot of effort and karma, and doesn't seem to stick long term despite my best efforts. 2. More specifically, I had a conversation you summarized above (taking TIME to dev well) just a few weeks ago with my manager. Over the last few years we've repeated a pattern, I take longer than he'd like, he gives me shit for it, I tell him it'll pay dividends long term, he grumbles, but then it does. (bug rates, iteration speed under frameworks I write, reuse due to good modularity/code cleanliness, these are all benefits you see rendered on a very short time scale, even just keeping with the most broad basics.) Now, I will give him credit for saying "you have a point and I'll leave it to your judgement" but that took me _two years_ of very incendiary 1:1's of my trying to put my money where my mouth is against someone who has been at this a lot longer than me and with some very strong opinions about that. I frankly don't see a "shortcut" after that experience, although the conclusion was more positive. So let me tie up this ramble with addressing your conclusion: I'd LOVE to take control of my working conditions back. But when the same people (PMs) who lead to my just throwing my hands up at even TRYING to get work done at work due to constant context switching/interruptions _SCHEDULE MEETINGS_ asking us why we're so slow (and then spend the whole meeting complaining about how WE pull them into meetings they don't need), it seems like there are just massive gaps of understanding or lack of shared incentive between the various powers that be. I don't know how to take control of my working condition back, but I've damn well been trying, and would be curious to hear any anecdotal suggestions from you in working this through more "political" environments.
- ebiester 10y agoI personally don't tolerate abuse and would GTFO, but I understand that's not always an option. :(
- wpietri 10y agoHopefully others will jump in with answers more relevant to your situation. I work mainly in a startup and small-company context because the customer and business realities are more prominent than the political situation. In my mind's eye, I see companies as opaque spheres of different sizes. The bigger they get, the farther the average person is from the external reality, and so the less that reality matters. I also think this is tied in with current American managerialist culture. The notion of controlling work by setting arbitrary hurdles and deadlines is so pervasive that many people can't imagine an alternative. This very easily turns into pure fantasy, even on the billion-dollar scale. [1] I expect that there are plenty of companies that would literally go out of business before they would change their ways. So in your shoes I'd take a hard look at your company and ask what its true values are. If they really value appearance over results, politics over substance, dominance over cooperation, then I don't think that's something you can change on your own. I'd personally try to find someplace better, and if I couldn't then I'd mainly play to survive, not to win. But if you think the company has its heart mostly in the right place, then in your shoes I'd start with weekly team retrospectives. [2] [3] Each week, you'll make note of problems and possible solutions. Each week you try to make things incrementally better. And gradually you'll build up solid data to say things like, "For the last quarter, the most common thing in team X's weekly top 3 productivity drains has been insufficient contiguous time to focus. If you want us to go faster, please help us to limit meetings and interruptions to the mornings for the next quarter as an experiment." If possible, I also strongly recommend finding ways to ship early and often. The longer managers go without experiencing actual progress, the crazier they get. If you release new business value at least weekly, then they'll focus much more on picking what's next then on yelling at people. At my last startup, even with a small team, we were releasing a few times a day. That meant that the CEO wanted something urgently, he could always get it soon just by telling us what thing currently in progress was less urgent. He always felt in control. It also meant that he was kept busy looking at new results and thinking about what came next; he didn't have time to fuss much with our work. Related to that is building cross-functional teams. To me, a team is a group of people that win and lose together, a group that can play the game basically on their own. So I'll try to put a few developers physically together with a designer and a product manager and whoever else is vital. If PMs think of themselves as their own team, then it's hard to counteract those bad incentives. But if the PM only wins when the project and the rest of the team wins, they'll be much more collaborative. And I'd also encourage you to make good use of every disaster that comes your way. The best luck I've had changing culture and process has always come when there's some mega-fuckup that's fresh in people's minds. That's especially true when you already have a successful small-scale example going. E.g., I remember one retrospective after a team spent 9 months building a big upgrade that was a flop. One of their peer teams was at that point releasing every few days and quickly responding to business data: killing bad ideas quickly, doubling down on the good ones. The pain of failure plus the envy of colleagues doing well was a powerful combination. Sorry for the ramble in reply, but basically my answer is: as problems present themselves, solve them in ways that move you toward greater agility. By which I mean more frequent releases, faster feedback loops, closer collaboration, less work in process, more responsive to business and customer needs. To it continuously on a small scale and regularly at the large. If you can make a 1% weekly improvement, things will be 68% better by the end of the year, and 181% better at two years. Does that help? If not, feel free to reply here or email me. Glad to help as I can. [1] For example, the dumbfounding mess in Mississippi, which we know about because of one stubbornly honest engineer: http://www.nytimes.com/2016/07/05/science/kemper-coal-mississippi.html?_r=0 http://www.nytimes.com/2016/07/05/science/kemper-coal-missis... [2] A good book is Larsen and Derby's Agile Retrospectives: https://www.amazon.com/Agile-Retrospectives-Making-Teams-Great/dp/0977616649 https://www.amazon.com/Agile-Retrospectives-Making-Teams-Gre... [3] Also handy is the Retromat: http://plans-for-retrospectives.com/ http://plans-for-retrospectives.com/
- zippergz 10y agoI think expecting business to know what they want is a mistake. I'm not much of an agile true believer, but as a former developer who is now "management" (shudder), I have never-ending frustration when teams expect full, perfect, set-in-stone requirements from day 1. When you're developing a new product in a new field, that's not how it works. We have to build something that minimally works, get it in front of real customers, and then make adjustments based on their reaction. Lather, rinse, repeat. This isn't "management can't make up their minds." This is "management are not clairvoyant." To those who hate requirements changes, I'd love to know how you are able to determine with 100% certainty the exact product the market wants before you build it. If you can do that consistently, you should be a billionaire. If you do understand that we don't know for sure what needs to be built from the beginning, you have two options. First, I can give you my current best guess at what we need to build, and we can work together to figure out what the right starting point (mvp) is, with the understanding that the rest may change later. Or I can figure out the mvp myself, and only tell you that, preventing the situation in which I told you about something that later has to change, but also robbing you of the opportunity to see the big picture and contribute to the plan. Which do you prefer?
- wpietri 10y agoExactly. That said, I think "give us the perfect requirements up front" is a reasonable defensive reaction. Because the first thing a lot of developers get asked is "When can you have this done?" In most places, if the (made up, arbitrary) deadline is not hit, it's not managers who say, "Damn, our bad; we should have been clearer about what we were asking for." It's almost always an inquisition into either a) why the estimates were wrong, or b) what's wrong with the developers that they aren't delivering on time. I now have a bag of tricks for keeping this kind of trouble to a minimum. But I think managers and execs have to stop asking, "When, to the exact hour, will this (hazy and ever-changing) thing be done?" It's not a sensible question, and if we pile blame culture on top of that, it's no wonder most projects quickly go off the rails.
- FollowSteph3 10y agoI think the main issue most devs struggle with is not the changing requirements but that the timelines are not adjusted accordingly and have to be swallowed by the dev with overtime. And in most cases when there is some adjustments it's not nearly enough to cover the actual costs. If that was somehow resolved I think the vast majority of devs would be fine with changing requirements. Problem is I don't think there is an easy way because most non-dev don't truly understand the costs involved. Fr example can u imagine telling a plumber no the bathroom is now going to be 30 feet to the left. The plumber says ok but that will cost u a week of work. How many people would say no, u have 1 day to do it and u will be held accountable for my estimate and not ur estimate because I don't trust u when u say a week. Now assume the walls are closed and it's only after u open them u learn there are other issues. Doesn't matter, one day. This is what most de s deal with. I assume ur not like this but this is how most devs have to work. It's frustrating. If u would be given the week then u would be fine, but instead ur told the timeline and/or cost. U may even be told no u can't use those tools they r too expensive, but u need to complete the task in the faster then even people who can use the tools. My whole point is it's often not the change requests that are the issue, it's what it entails, in that the schedule and costs are not adjusted fairly.