justanotherdot •

Ryan James Spencer

Newsletter

Nothing of Value Will be Lost

on November 10 2019, 6:49PM

Drop your backlog. Burn all of your tickets. Eject your issues into the sun. Nothing of value will be lost. Teams and maintainers alike cling to reminders of work as if they are the same as the result of the work itself. Backlog grooming sessions pass and only thin slabs of the gelatinous mass disappear into the abyss all the while delaying developers from focusing on actual work.

Design and debate need to occur for a project to progress and when those things happen it's good to record the results. As these records accumulate they age because parts take priority over others. Ok, maybe you don't want to drop everything but you definitely want to drop items older than a certain age. I'm fond of choosing a natural period of time where you, the human, can easily enumerate key points that have happened. Longer periods of time produce smaller, less detailed lists. Periods of time that are too small might experience churn on the issue tracker as items disappear and return repeatedly.

This process might sound crazed. How dare we close valid issues tied to real people on an open source project or abandon fixes and feature work that could drive up revenue and delight users purely because of age? Finding what to work on is not the hard part, despite what you may think. Prioritising, hashing out ideas, and setting goals has value but ninety percent of everything is crap and issues sitting in the dark, ignored and unloved, are alike.

Those using todo lists will know the value of scrapping them at the end of a day or week. Copying a few things over from prior days or periods of time can be beneficial but usually the gain is marginal. Adopt a process that reflects the fact that things change rapidly. Work that needs doing is from problems and pain points that are being frequently encountered. It's work that's at the tip of the tongue. This is the reason we care about 99th percentiles and avoid one-off optimisations and bug fixes. What if the bug fix is data being spuriously deleted? I can guarantee that issue won't stagnate and if it does I think there are deeper issues that need handling. In the same way Kent Beck told people on Stack Overflow that he gets "paid for code that works, not for tests", one also won't (shouldn't?) get paid writing or pruning issues.

Neglect the fool's gold of issue trackers. Your brain isn't a storage device; enable your mind to process what it ought to be processing by using these glorified todo lists to offload information.