Design principles

  • Idempotent.

    The program can be run against it’s output and should not make any changes to it. This property is checked by integration testing.

    This allows users to integrate the program into their workflow and let yaml4rst do the repetitive tasks.

  • Keep it simple.

    The Python implementation is thought to be close to the way a human would solve the problem. It basically just parses each line of input into fold sections and checks if something needs to be changed. At the end, the sections are formated back to lines and written to the output.

  • Committed to excellence.

    This program is reasonably small and it’s scope can be clearly defined so that it is possible even for one person to try to follow most of the best practices in Python programming which includes:

    • Extensive unit and integration testing
    • 100 % unit test code coverage (including branch coverage)
    • Python version test matrix
    • Python linting tools like pylint
    • Documentation build testing
    • No redundancy

    Furthermore, most of the best practices listed above are automatically checked (where possible) on each commit and before PRs get accepted.

  • Why not do it manually? Hoped to be helpful.

    “How long can you work on making a routine task more efficient before you’re spending more time than you save?” (Ref: xkcd 1205)

    “I spend a lot of time on this task. I should write a program automating it!” (Ref: xkcd 1319)

    Based on the analysis of git-hours it took ypid 118 hours for the initial public commit. This equals 14.75 days (assuming 8 hours a day). The time for ongoing maintenance of the program is not included and can be estimated based on the public commits.

    From that it is clear that for ypid, it took much longer to automate the task then it would have been to do it manually. But that is only part of the story because ypid is not the only person doing this kind of maintenance. So it is believed that this effort will still pay of in saving time for everyone doing this and lead to more productivity and higher quality. (Ref: Collective)

  • No formal parsing/lexing.

    Existing parsers for YAML where found to be unsuited for the use in a statical analyser/reformatter like this program.

    Also, ypid is more familiar with this hacky approach then he would like to confess (ref: opening_hours.js).

  • Don’t reimplement error checking from Ansible or RST parsers.

    The input is not extensively checked. Use Ansible and RST before and/or after the reformatting through this program.