Specifications document for a Podcast rebroadcasting tool:
The Problems:
- You have podcasts from shows spanning years, that have changed server multiple times over the course of their life. As a result, feed-centric podcast apps (Apple’s Podcasts.app, Doughnut etc.) will display those episodes as separate shows, so your archive gets split up and unmanageable.
- The Podcatching app you use on your desktop computer has been obsolesced, forcing you to change to a new app, and the import / export options are for the feeds you’re subscribed to, not the back-catalogue of files you’ve kept on disk. Your new app’s podcast library episodes are now disconnected from your back catalogue.
- You are an obsessive hoarder, you probably collect(ed) comic books, and you want your Podcast library to be organised the way you want your podcast library organised, because the point of computers is to do what you want to do, the way you want it done, and podcast / podcatch apps are moving towards keeping episodes in locked file structures, with obfuscated filenames etc.
- You like the way podcasting used to work, where you downloaded to a library on your computer, then synced episodes to your mobile device to listen. You don’t want to have to sync your large podcast files via the cloud to move them from your computer, to your mobile device on the same desk. You don’t want to rent a cloud service that might disappear next month to store your library on someone else’s computer. You don’t want to subscribe directly on your mobile device, or stream shows. You want that oldskool iPod-era Podcasting.
- You want to be able to edit audio files, change their metadata, their dates etc, and this will break their membership of podcast feeds where the canonical details are controlled by the feed publisher.
- You would like to be able to roll your own Podcast feed, to create the equivalent of a Podcast playlist of shows from different sources that is independent of any one player app.
The Solution:
- A script that automatically generates (and updates) a local Podcast RSS feed based on the contents of an arbitrary folder. These feeds can then be subscribed to in a Podcast player app on the same machine, or in a player app on a different device on your local network, which can then download the files to device without needing to use an iTunes-like syncing process.
The Architecture:
- Publisher’s RSS Feed
- Local Podcatcher
- Local Filesystem Archive (manually configured, or with user’s choice of automation)
- (Re)Cast RSS Feedmaker <- the bit we make
- Local Web Server (MAMP, MAMP Pro etc) <- (Re)Cast Server we make this bit as well. Integrated into the feed maker, or a separate piece of software.
- Local and Local Network podcast player (streaming or download to device if going off-network)
- Local Web Server (MAMP, MAMP Pro etc) <- (Re)Cast Server we make this bit as well. Integrated into the feed maker, or a separate piece of software.
- (Re)Cast RSS Feedmaker <- the bit we make
- Local Filesystem Archive (manually configured, or with user’s choice of automation)
- Local Podcatcher
The User Experience:
- A typical use case might be that a user has a large back catalogue of shows they’ve been downloading with iTunes, or other download-centric podcast tools.
- They would manually set up a folder structure for their Podcast Archive, and move their back catalogue into these folders.
- For new downloads, they could have a dedicated Podcatcher app. If their Podcatcher builds its own folder structure, they could replace those folders with symlinks to the matching folders in their Podcast Archive, or have automation that automatically moves downloaded files from the Podcatcher’s folders, to the Podcast Archive folders. Part of that automated process could involve stripping problematic metadata, for example.
- The user would create a Folder Action / Shortcut / Keyboard Maestro / Hazel trigger (whatever tool they choose to use) for each folder in the Podcast Archive, to fire the script to (re)generate the RSS feed based on the folder’s contents when a new file is added.
- A feed would be created for each folder in the Podcast Archive, and those feeds would then be subscribed to in the user-facing Podcast app. The user could change their user-facing Podcast app without having to export / import OPML etc, they can just re-subscribe to their local feeds in the new app, and their whole back catalogue remains – no need to re-download etc.
Technical User Experience:
- The best-case scenario is the task can be accomplished with a shell command, piping output to input between standard included command-line tools, so that all we have to do is provide the command which can be pasted manually into a terminal navigated to the appropriate directory. That way an automation tool merely needs to apply the text string as a command in the appropriate places.
- Failing that, setting up and managing the script should be editing a text file.
- The user shouldn’t need to modify their shell’s command path, or install extra software – There’s Automator and Shortcuts built in on the Mac, which can both run shell scripts, as well as third party tools like Keyboard Maestro and Hazel.
- As long as the scripts can be stored in a specific place that we can specify, that provides unambiguous implementation. If they can be stored in each podcast’s directory, perhaps it requires no editing because it automatically applies to its containing folder (best possible use case, because we just drop the script into the folder, and it’s done), or we require a dedicated /Scripts folder in the Podcast Archive where the user configures scripts with the appropriate relative paths to where the files are located.
- We only need to provide a commented edit the path here direction.
- The script would default to file creation or file modified dates for feed order, and users could change the order by using a file metadata modifying tool to change that.
- Allowing folders filled with aliases / symlinks / shortcuts to work as if the original files were in them would allow users to build Playlists, without needing a Podcast app that has Playlist support (Apple Podcasts.app notably removed Podcast Playlists).
The Big Picture:
- A main goal of this exercise, is to make people’s Podcast experience and workflow filesystem centric, rather than feed, app, publisher, or metadata centric.
- The idea should be to abstract the acquisition of shows, from the management of a content library, from the listening to those shows, from the syncing those shows to another device.
- The goal of this abstraction, ultimately, is to protect against Enshittification by ensuring that any one part of the process can be replaced and any choice of tool can be used to effect changes within the process.
- By concentrating on just doing the generation of the RSS feeds, we provide a utility that can be used in as simple, or ambitious a setup as the user desires, to the extent that they can even swap between the automation systems used to fire the scripts, if one is deprecated (Automator, for example).