Before Dolt, to share a SQL database with a fellow data practitioner, you both needed to share the same view of the data. Only one write could happen at a time. Making a copy implied creating a point in time backup and restoring on a separate running server. Once that copy was made, the two databases could change independently. There was no tractable way to compare the two copies of the database to see what changed. Moreover, there was no easy way to merge the two copies back together. In source code parlance, the copy was a hard fork of the database.