After debugging some issues on IRC with Jan he invited me to propose an article for the CouchDB blog about the work I'm doing for Drupal and CouchDB.
For those that haven't seen, the TL;DR is: We are exposing CouchDB compatible endpoints from Drupal (the PHP based CMS, you knew that, right?), so that we can replicate content between Drupal sites using the replication.io protocol. It'd be awesome to write a blog post giving a wide arching overview of what we are doing. We've had to: - Alter all content in Drupal to be revisionable and enforce that - Implement revision hashes - Prevent revisions from being deleted (and add _deleted flag) - Store parent revisions and generate a revision tree - Normalize and denormalize all content to and from json - Setup all endpoints - Implement a replicator in php so we can do full Drupal to Drupal replication without CouchDB One link I shared with Jan was https://travis-ci.org/relaxedws/drupal-relaxed/jobs/116109952 which shows a test we run adding documents to couchdb, then using out PHP replicator to replicate these documents from couchdb, to drupal, then to another drupal site, then back to couchdb again. We can a number of additional documents through the process, but these are Drupal uses which we are treating as documents and also replicating.
