What is a "read-only follower"? Is this a common term when handling databases? Is it different than a slave?
I will answer in the context of PostgreSQL which is the system I know the best. PostgreSQL allows you to create "copies" of the database using replication. A database which is replicating is always read-only. The database from which it is replicating can be read-write (a master) or read-only. In other words, replicas can replicate from replicas in a tree structure.
Any PostgreSQL server which is replicating is referred to in the documentation as a "standby" or "slave". The terms are used interchangeably.
The replication can be synchronous, in which case the "master" and "slave" have exactly the same set of committed transactions at all times. Or, it can be asynchronous, in which case the slave might possibly lag behind the master. Only a master can have a synchronous replica. It used to be that you could only have one synchronous standby, but with recent PostgreSQL you can have as many as you like.
A standby is called "warm" if it is not serving client requests. It's "hot" if it is. Of course, standbys can only serve read traffic.
To answer your question, "read-only follower" is probably a "hot standby". At least, a Heroku follower is generally there to serve read traffic.
https://www.postgresql.org/docs/9.6/static/high-availability...
https://devcenter.heroku.com/articles/heroku-postgres-follow...
"leader/follower" or "primary/replica" are much more neutral terms that won't prompt negative emotions in many people.
People who choose one of the latter two options do so either because they feel it is more accurate, or because they wish to avoid the negative connotations of "master/slave", or a mix of both.
You might not care, but there are a lot of people that do. Taking steps like this improves the comfort level others while affecting yours none. Why is that not worth it?