Think of potential advantages of using IPFS folders for pages
For example we could do <cid>/page-1 , <cid>/page-2, etc.
We could actually set up the entire subs like this subplebbit/hot/page-1, subplebbit/hot/page-2, etc
We could also do it for post replies pages subplebbit/postUpdates/<post-cid>/replies/top/page-1, subplebbit/postUpdates/<post-cid>/replies/top/page-2, etc
I'm not sure what would be the advantage of doing this, but it's something interesting to think about.
Another interesting thing to think about could be to have a single cid/file for pages, and only download the first 1mb, then to scroll to the next page, you download the second 1mb of the file, etc, kind of like how bittorrent lets you download partial files. I'm also not sure what would be the advantage of doing this, but something to think about.
another idea could be to use ndjson (newline delimited json), which is something kubo rpc uses sometimes, and make each new line a post, and then people could stream pages by fetching bytes 0-1000, 1001-2000, etc, they wouldnt need to fetch a full page, and then fetch nextCid, etc.
not sure it has any advantage to our current design, but it's something to think about for later
I'm not sure what would be the advantage of doing this, but it's something interesting to think about.
wouldn't it be possible to FUSE mount a unixs-fs folder from a local machine? Then you could at least read from and possibly write posts to the local file system. https://github.com/ipfs/kubo/blob/master/docs/fuse.md