Scheduling, getting things done, motivation for trailblazing project

March 13, 2014 7:36 AM

I am working for myself on a software project based on an idea of mine. I have no clear estimates of how long things would take, because most steps are new concepts, and I don't know how to do them off the bat. I have been trying out Scrum, but it seems the same item gets recycled week after week because it is not done yet. I have the problem of getting easily distracted, so this apparent lack of progress based on these metrics is making me lose confidence and procrastinate. I need a better way to organize myself and maintain productivity. I use Trello for managing the project and it's great, but it does not help with productivity, time estimates, WHEN to do things and how long to allocate.
posted by spacefire to Work & Money (4 answers total)

If you've never done something before, it's basically impossible to estimate how much time it'll take. This is an ongoing massive problem software development that nobody has really cracked yet. Sure, if you've built a module that does X then building one that's kind of X-ish might take a roughly similar amount of time, but if it's totally new to you? You really, truly, cannot estimate how much time it'll take. I'm a professional software engineer, and I'm also working on a side project of my own, and I don't pretend for an minute to have any clue how long any part of it will take, because it's all totally new to me - even the parts that a professional working on this kind of thing would be able to provide very reasonable estimates for ("oh, yeah, two days for that.") Everything I do I have to learn first; learning takes time.

Short version: Focus on blocking out the time to work, but don't kill yourself trying to figure out how long something will take to do. It's more or less impossible when everything is brand-new to the team (ie, you.)
posted by Tomorrowful at 7:54 AM on March 13, 2014

Perhaps you should go smaller with your sprints? If your item is taking weeks and weeks, it's too big. Each week should have a goal you're 90% sure you can reach. Maybe that means this week you create a mock class that takes input and doesn't do anything with it. And maybe next week it just does one thing with the input, but not everything you want it to do by the end of the project.

You may not know how long everything will take, but I bet you can come up with a next step that is reasonable and can be completed in a week.
posted by jsturgill at 9:25 AM on March 13, 2014

Scrum is apparently intended for a development team. If you are working solo, I think you need to look for another guide.

Having done a lot of "trail blazing" type things, here are a few thoughts:

1) Work on a basis of milestones, not deadlines. Figure out what your milestones are. Most milestones will not need to be tied too closely to a time frame.

2) Put things in order, working backwards. A Gantt chart might help you with this. Some things naturally need to be done before others. It sometimes helps to figure out the end point and work back to the start to figure out the order of events you need.

3) If you are working alone, you need to find a process that works for you as an individual. A lot of this is about human issues like hunger, sleep, etc. Some people work better in the morning. Some are better in the evening. Etc. Figuring out what you need to do to get yourself in optimal working mode is one of the best things you can do and it is largely outside the scope of any software development models. It is more the purview of the "human resources department," so to speak.
posted by Michele in California at 11:14 AM on March 13, 2014

What works best for me for motivation is my version of Getting Things Done. I keep a list of to-do items (trying to make them the smallest step possible) plus I have tickets to work through that my team makes when doing QA. At the beginning of the day, I write the next three things down on a separate piece of paper and I work on those three things in a round-robin kind of way. So I work on the first for about 30 minutes or until I get to a point where I'm kind of burned out or stuck. Then I move to the next item. Rinse and Repeat until 2 of the items are marked off the list. Then you can add 2 more items.

This helps me with my distractibility because generally the 3 items are different enough to keep me interested. And because it narrows my focus. If, while working on one of the items, I realize that another step needs to happen then I put the next step at the end of my list and write the "stuck" item on it after that one. My "rules" for this are very loose. Basically, if I find myself spinning my wheels on an item then I try to think of the very next thing that needs to happen with it and just add that to the long to-do list to revisit later.

If I'm working tickets then when I'm ready to move on to the next item, I write the next step in the comments so that I remember where I was when I come back to it.

If the item on the list is a research item then I spend about 30 minutes on it before moving on so that I don't waste a whole afternoon in a wormhole.

Along with this and some time for reviewing the project each day and week, I can get quite a bit done during my work sessions. But it also helps to have caffeine and headphones with Rage Against the Machine blasting the whole time. good luck!
posted by dawkins_7 at 8:58 AM on March 16, 2014