Reconciling pagination with offline mode

As #15 demonstrates, pagination becomes a non-trivial problem in the presence of offline mode and dynamic content. This page collects research and thoughts on solving this problem.

I have a sneaking suspicion that this can be framed as a classic CS problem with a well-known solution, but I haven't found a way to do that yet.

Resources

• https://medium.freecodecamp.org/how-to-implement-cacheable-pagination-of-frequently-changing-content-c8ddc8269e81
• https://stackoverflow.com/q/41961255/504611
• https://softwareengineering.stackexchange.com/q/321345/13821

Constraints & capabilities

An ideal solution would have the following characteristics:

• MUST NOT require modifications on the server side
• MUST be correct
  • Posts MUST NOT be skipped in the pagination process
  • Posts created/edited/deleted offline MUST be handled correctly
  • Potential conflicts caused by simultaneous editing MUST be handled reasonably
  • Any given post appearing in the list MUST NOT be stale after a refresh
• SHOULD minimize network usage; gzip compression can be assumed
  • SHOULD scale to large blogs of 10k+ posts - a refresh should take no more than a few hundred KBs

Here is what we get from the Ghost server:

• Fetch all posts
• Fetch a page of posts, defined by a tuple of (page num, limit)
• Order posts using arbitrary fields before pagination
• Filter the fields included in the response. This is useful because we can cache post contents separately from the list of posts, and for the latter fetch only post ids + updated_at timestamps (to determine ordering and freshness of content respectively), with a GET request to the following URL: /ghost/api/v0.1/posts/?fields=id,updated_at&...