{
  "path": "/ode-to-plain-text-files.html",
  "site": "at://did:plc:x67qh7v3fd7znbdhauc45ng3/site.standard.publication/3mjcd2t6afe25",
  "$type": "site.standard.document",
  "title": "I <3 plain text files",
  "updatedAt": "2016-09-09T00:00:00.000Z",
  "publishedAt": "2016-09-09T00:00:00.000Z",
  "textContent": "I really enjoy working with plain text files. More software should work with plain text files. It obviously depends on what kind/amount of information you want to store, but I think for a lot of use cases text files are a very good solution.\n\nVersion control everything\n\nText files can easily be put into a git repository to be versioned, compared, tracked, and shared.\n\nI have not seen a database system that easily allows you all of that.\n\nApp configs need not be complex\n\nInstead of putting configuration values into huge XML files (XML != 'plain') that cannot be edited by regular humans, it makes sense to choose a readable (approaching 'plain') text format, like YAML or TOML. JSON is okay, but not as nice as TOML.\n\nAutocomplete with JSON Schema\n\nVSCode supports autocomplete based on JSON Schema files. Doing the same for similar formats should be doable.\n\nSanity check, linters\n\nWhen human and machine want to read, edit, and understand the same file, it usually pays off to let the machine a bit of sanity checking. Linters and formatters are your friend.\n\n- - - -\n\nSee also\n\nAndre not only wrote about this a year before I did, but his arguments are also much better: Plain Text Reigns Supreme.",
  "canonicalUrl": "https://deterministic.space//ode-to-plain-text-files.html"
}