This has probably occurred to me, and likely to everyone else out there – programmers will spend orders of magnitude more time to automate a tedious, boring and error prone process than it would actually take to perform that process weekly for the next four years, even though it likely would take place for only a few more weeks.
This comes up in light of converting a relatively simple static site I maintain for a kickball league to use a templating engine(more on that later) so that I wouldn’t have to a) change all six pages to change the headers/footers/menu etc, and b) wouldn’t have to use an excel spreadsheet and type and copy and paste scores from place to place every week to update the results. In my case, I have an issue where I’m very interested in learning new things, but always need some tangible project to perform otherwise I lose interest – so I can chalk it up to learning something new, even if I have saved negative time in the course of the implementation over doing things the old fashioned way.