TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Ask HN: Ticket driven development: are you sick of it and how'd you change it?

5 点作者 blizkreeg超过 3 年前
Software development in most startup teams today has essentially become ticket-driven development. Tickets are created, assigned, and are magically expected to be delivered on in 1 or 2-wk cycles. Then, more new tickets are created, assigned, and the cycle repeats.<p>The PM&#x2F;EM writes a bunch of poorly detailed tickets (often with no detail) and expects devs to deliver on them. Is this the situation in your team too? If there is detail, it&#x27;s usually a few hastily written lines in the JIRA ticket description with the rest left to the developer to figure out, fill in the blanks on, or ask questions around.<p>1. have you experienced this? and if so, have you seen that this leads to poorer quality deliverables or does it not have any impact?<p>2. what would you fix about this process (systems, people, tools) to make it better?

3 条评论

PaulHoule超过 3 年前
Where I am today I often get tickets that are too vague to start on.<p>I think about it a bit, spend ten minutes talking to the person who wrote the ticket, turn the ticket into a good ticket, and proceed.<p>The effort to do that is maybe 5% of the work to complete the ticket so I can&#x27;t complain.<p>Even though we do sprints what we are doing is really kanban where the main focus is finishing what we start and not having excess WIP. If a ticket takes a sprint and a half to finish, there is always something from the backlog to add to the sprint.<p>There are many ways to use a ticket system. One of the times I was happiest was when I was able to quickly break down my work into 10-20 minute subtasks and make super-accurate estimates. That was dependent on having a framework in which the tasks were predictable and also quickly punching tasks into a spreadsheet. If the work involved a lot of struggling with the runtime, build system, algorithms, database, etc. that kind of predictability is impossible.<p>There is always some disconnect between &quot;what we say we do&quot; and &quot;what we really do&quot; in a ticket system. Those agreements and disagreements can drive people crazy (fine-grained tasks help me focus but will drive many people up the wall) but it&#x27;s also possible to reach an accommodation where work gets done without a lot of space.<p>(e.g. it drove me nuts to write a daily standup summary of yesterday and today in the morning with the rest of my team. in the morning I&#x27;d find it hard to remember exactly what i did yesterday and what I was planning to do last afternoon and it would slow me down getting to work. Now I write my summary when I leave with the headlines &quot;what i did&quot; and &quot;what i plan to do&quot; that recognizes that those things don&#x27;t always agree. Those notes help me get started in the morning and nobody complains I am 180° out of phase.)
nick0001超过 3 年前
Establish 1 year product road map. Break down that year into quarters and so on.<p>Hire someone to help you offload tasks so you can focus on long term growth or to help you with the long term vision.<p>If resources are limited, at least take a week or 2 to plan this road map.<p>(I am in a similar situation)
postno123456789超过 3 年前
Yes. Estimation has about a 50&#x2F;50 chance of being correct. When it&#x27;s wrong management is unhappy and thinks we&#x27;re doing it wrong.<p>Change? Use the cards as a guide only.