Some of the filesystem notification APIs have limits on the number of files and the number of watcher instances that can exist. In a case where a software project wants to watch different sets of files and perform different actions based on the change they could use multiple instances of pocket but that would consume more of the available watch resources and could impose a limit. If pocket implemented multiple watchers / command runners in user space it would be able to scale to more actions without consuming more of the native API resources.
Some of the filesystem notification APIs have limits on the number of files and the number of watcher instances that can exist. In a case where a software project wants to watch different sets of files and perform different actions based on the change they could use multiple instances of pocket but that would consume more of the available watch resources and could impose a limit. If pocket implemented multiple watchers / command runners in user space it would be able to scale to more actions without consuming more of the native API resources.