[tjalfi]

I know more than I'd like to about this topic; I spent most of a decade supporting legal document management systems.

Large law firms use document management systems[0] to store their documents. It's a really primitive VCS that integrates with Microsoft Office. The user who checks out a document usually has exclusive access to it until it's checked back in. Other people can check out a copy of the same document, but it won't usually contain any changes the other person made. There's additional work required and many users don't know how to do this. It's common for changes to be lost or a partner to have the help desk unlock documents. I'm guessing that's what happened in this case.

[0] https://en.wikipedia.org/wiki/Document_management_system

[justinclift]

That approach sounds like pessimistic locking: https://en.wikipedia.org/wiki/Lock_(computer_science)#Databa...

Optimistic locking might work better, as long as the underlying data can be automatically diffed and merged reliably well.

Git generally does well using this approach, though it's pretty easy to diff text.