Open Source and enjoyment

A few years ago I burned out on Open Source projects, to the point that it was close to impacting other parts of my life. I have only myself to blame as I didn’t manage to distance myself at all well. Distancing yourself is not easy to do when your fun hobby for relaxation is the thing that is working against you.

There were a variety of incredibly minor issues that piled up Jenga-style. Then to carry the weak analogy forward, it toppled over in a split second like Jenga in a rubber dingy on the rolling seas with a fumbling drunken friend wearing a blindfold.

I’ve decided to write down a few of the things that were happening, both as an act of catharsis and to help me spot the problems in future. Perhaps it will help others who are experiencing the same thing but haven’t quite spotted the catalysts yet.

Dealing with needy users

Some users seem to feel a disproportionate amount of “you work for me”, seemingly more so than happens with actual paying customers on other projects which strikes me as strange. They’ll bug you endlessly in emails, in the issue tracker, on IRC and in real life if you end up at the same conference. You need to stop this before it takes hold. It helps if you can be polite, but if not just stop it anyway. It is for your welfare, so just do it.

I actually found myself burning the wrong people towards the peak of that imaginary Jenga tower. You need to make a concerted effort to figure out which people are delivering the seemingly endless grief or perhaps just relentless bikeshedding, and stop them. It is all too easy to accidentally become annoyed at well meaning useful contributors when you’ve been trawling through a heap of useless meta-discussion about a change nobody is making.

On one library project I worked on there was a user who didn’t read the documentation, couldn’t use the language enough to write tests, yet still endlessly sucked up support time by opening awful issues and sending pointless emails wishing to discuss changes they might make. I kind of hesitate to say awful issues, but - for example - if you can’t write a two line assertion in a language you’re not in the target audience for a library in the first place.

Those people aren’t even adding value in pointing out deficiencies in documentation, because they’re not a target user. It would be like me writing to JCB telling them they must document why they use round knobs on the digger I saw someone else using. So try spot it early, and put a stop to it.

I suspect a big part of the issue here is cost. I’m sure if there was even a single penny cost to opening issues or sending a mail it would do wonders for communication in general. I suspect that would remain true even if you reimbursed that penny immediately.

The answer

I figured out how to work with this entirely by accident. Politely offer answers that move them away from you. It is by far the easiest solution.

In one instance I suggested alternative packages to an expensive user, with the hope they’d quietly move on. Which is evil really, as what you’re really doing is dreaming that they’ll move on to bother a developer who isn’t you.

On another occasion I stole an idea from a friend and said “Sorry, I can’t figure out how to implement this. Open a pull request, and I’ll merge it!” to another. It worked well, because it stopped the direction-less discussion and was never going to be followed up with a patch later.

Dealing with unreasonable requests

I used to get quite annoyed when I received an impolite or unreasonable request on an Open Source project, nowadays I tend to just respond with pre-canned answers for most of them. If you can cut the time it takes to respond to them, you’ll obviously spend far less time thinking about them.

The typical one that springs to mind is a user complaining that you’ve not licensed your GPL code in a way which allows them to use it without contributing anything in return. I choose reciprocating licences where I can, because frankly that may be the only value you provide me as a user. I get that some people need to use their Open Source projects as a way to improve their CV or build a presence, but for many of us it is just for fun and reciprocation.

Another example that I recall, but wish I didn’t, was a user telling me to recompress a tarball on PyPI because their infrastructure couldn’t support bzip2. Seems reasonable enough, but it came wrapped in a tirade of abuse. I ignored the abuse, uploaded another tarball and then received another abusive email in return the next day. That was a long time ago, far before the burnout started to creep in so I just addressed the actual issue and moved on.

The answer

I had forgotten about that whole previous story until at Europython a friend was telling me about someone who was banned from a number of events, hell-banned on various mailing lists and on forced lockout on Stack Overflow for poisonous behaviour. I started to tell my story as an example of strange abusive behaviour, and it turned out to be the same guy.

The point is these people can be everywhere so just don’t let them get to you. I realise that is both obvious and feels hard to do, but it is quite easy in the virtual world. If you have the hell-ban option available just use it; perhaps there is an ignore option in your issue tracker, add a “send to trash” rule in your email filter. Even in person it can be quite easy to do, just politely acknowledge their issue and move away instead of engaging them.

The solutions

I chose a solution I hate, for all intents and purposes I don’t release new Open Source projects any longer. I still work on a lot of projects, and the code is normally available somewhere for strangers. However, for the most part I’ve made a point of not hosting or releasing it in the common places. This works for me, but it saddens me deeply.

I have a friend who does something interesting to combat the same problems I’ve mentioned here. He hooks the issue tracker to only show issues reported by people in his second-degree FOAF circle or people whose user profile is an active Open Source contributor. It works, and it cuts out most of the problem users, but it is nasty. I feel it is worse than my “solution” because it still gives the impression of support to the users you won’t support, but also gives the impression of bad support through heaps of unanswered issues to the users who would receive good support.

The greatest trick the devil ever played was convincing users they had skin in the game.

I may be misappropriating a quote from a great film there, but it is important. A lot of the actively poisonous people who contact you aren’t providing any real value to you, and they’re probably never going to. If you can remove them somehow you’re probably not losing anything, but you’ll be gaining a whole lot personally.