\n
\n How to create an automated web browser:<\/p>\n
\ud83d\udd39 Study Electron.js\ud83d\udd39 Study React.js\ud83d\udd39 Make the two play nice together\ud83d\udd39 Abandon react-redux and write your own\ud83d\udd39 Create an SDK, and make it self-documenting\ud83d\udd39 Make sure all events are native (isTrusted flag is true)\ud83d\udd39 Create an IDE from scratch\ud83d\udd39 Make the SDK and IDE play nice together\ud83d\udd39 Create a modules system\ud83d\udd39 Create a public API for the modules\ud83d\udd39 Create an I\/O system\ud83d\udd39 Create a command-line interface<\/p>\n
And even though you obsessed over UX all this time, the app still feels uninviting, for lack of a better word.<\/p>\n
\ud83d\udc49 Then one weekend, add the ability to record automations with point-and-click. \ud83e\ude84 Boom! Magic.<\/p>\n
This is some version of the Pareto Principle in action. Even though a minority of actions produce the majority of the consequences, you really can’t skip the work.<\/p>\n
I could not have added the recording feature without first building everything else.<\/p>\n
And moving forward, this pattern will repeat. Most of the effort that goes into uindow will go unnoticed. But from time to time, users will say “wow” to magic that can only happen because of the invisible work.<\/p>\n
Uindow is a free, and source-available automated web browser. You can check it out on GitHub.<\/p>\n