Todd Wolfson

Software Engineer

November 07, 2015

Staying active in open source is a problem I see every now and then. It's a problem that I have experienced in the form of burn out.

This article is to serve as a brain dump for what I have learned over time.

If you aren't passionate now, then you won't be passionate later

If you are thinking about building a project because you feel obligated (not because you want to), then it will likely become a nuisance rather than something enjoyable.

For example, I maintain a project called spritesmith which helps easily create CSS sprites. Then, one day I decided to build a sibling project named fontsmith for web fonts. Unfortunately, it was never something I wound up using and eventually, after enough issues and frustration, I decided to deprecate it.

https://github.com/Ensighten/grunt-spritesmith

https://github.com/twolfson/grunt-fontsmith

Keep things focused

The Unix Philosophy of "Do one thing and do it well" has helped an enormous amount with keeping feature size down and projects focused.

Anything that you build, you will need to maintain later. As a result, adding in unnecessary one-off features will lead to a feeling of constantly being overwhelmed later on.

For example, a common request in spritesmith is to add in using 0 over 0px and combine common CSS properties (e.g. background-image/width/height).However, as I explain in the issues, this is starting down the road of CSS minification. As a result, we left that complexity out of spritesmith and requested people use a CSS minification tool after spritesmith.

https://github.com/twolfson/spritesheet-templates/issues/11

https://github.com/twolfson/spritesheet-templates/pull/43

Be prepared for bugs/features

Don't lie to yourself, open source will take time. People will report bugs and request features so make sure that you will have time for it.

Maintaining open source isn't exactly a part time job but it's definitely not a vacation either. Bug reports will pop up at your least convenient moment and feature requests will pop up right before you were going to work on that brand new project.

Learn to handle these issues quickly and efficiently via tools like tests, friendly debugging errors, and expansive yet concise documentation (e.g. lots of examples).

Provide great customer support

By open sourcing a project, you are implicitly agreeing to become its customer support and project manager. As a result, try to obey the golden rule and provide the best customer support experience.

Here are some rules of thumb I follow:

  • Reply within 48 hours with an ETA of when I can look into the issue
    • I typically give myself 5 days padding
      • For example, on a Friday I will say "Thanks for the report! I will take a look at this the end of next weekend"
      • On a Monday, I will say "by the end of this weekend".
    • If I am unable to handle the issue in a timely fashion, then I communicate that.
      • For example, "I am on vacation for the next 2 weeks. I will give you an ETA on when I can triage this when I get back."
      • We don't give a full ETA since something more important could pop up within the next 2 weeks
  • Frequently communicate
    • If you are about to look into an issue, let the user know. For example, "I am taking a look at this issue now"
    • If you resolve the issue relies on something else, let the user know. For example, "We have found the source of the issue and reported it to <Link to repository>. <Link to issue>"
    • If you are unable to fully resolve the issue now, leave notes for your future self and the user.
      • For example "I have successfully reproduced the issue. I'm currently out of time for now though. I will take a look at implementing a patch by the end of this weekend. Here are notes for my future self: <Bullets outlining how to reproduce>"

Empathize with users

Unfortunately, being quick to reply and resolve issues is only half of the battle. We also want to make sure both parties are happy about the experience. As a result, be sure to empathize with end users.

Never close issues on impulse. If someone is asking for a feature you consider unncessary, don't immediately close the issue with an unhelpful response. This helps nobody. Instead, take the time continue the thread and ask them "Can you explain your use case further?" Sometimes you will be surprised and be able to relate.

For example, in spritesmith one day I got an issue about disabling sorting. My initial reaction was "Why would you want that?" but as it turned out, the user was using it for animation sprites. As a result, we built a feature to disable sorting of sprites.

https://github.com/twolfson/layout/issues/5

If someone is asking about a duplicate issue, take the time to write out an example for their use case and link them to the existing issue. They will be grateful, you might catch an edge case in their setup, and you have yet another example to point to when someone is having trouble.

Don't overwork yourself

If you feel like you are doing too much, then you probably are. Always give yourself time to relax and de-stress from other parts of your life. This is typically why I will give padded time estimates.

Also, don't forget that you aren't an island. Feel free to invite people who submit good PRs as collaborators and mark issues with "help-wanted".

Don't forget why you started

Sometimes when you are in the thick of everything, it's hard to remember why you started doing something.

Every now and then, I need to remind myself of why I do open source in the first place.

http://twolfson.com/2013-04-22-why-i-open-source

Sources:

Related articles

Why I open source

An explanation of my morals/ethics and why I devote my free time to open source.

Top articles

Lessons of a startup engineer

Lessons from being a 3x first engineer, former Uber engineer, and working at even more startups

Develop faster

Removing the tedium from creating, developing, and publishing repos.