I've (very) slowly working on a knowledge base and planning features. All focused on myself/FOSS. A big desire i recognized early on is being able to store immutable references to the source material. Good for offline usage, but also nice as a "Reader". I would probably ingest direct to some reasonably bare HTML or Markdown format.
Alternatively, i wonder if adhering to Gemini/etc would be a good way to achieve this?
Ie rather than store data in some adhoc format, you just write a "Traditional Web -> Gemini" converter, and then just build the Reader ontop of that. Clearly i'm just speaking of Gemini as a storage format, ignoring transport mechanisms or protocols. Ie the HTML aspect of Gemini. I am unclear if Gemini is just literally a subset HTML - but even if it is, it adds a nice constraint on what HTML this Reader would produce.
Seems neat and general purpose. As long as Clients can markup Gemini content with enough semantic visual meaning as to not detract from the experience in my eyes. Ie i'd want to be able to view Codeblocks with some simple syntax highlighting.. not sure if Gemini prevents that somehow.
Now i need to look into Web -> Gemini scrapers, got me interested