The fundamental problem is that words are an approximation, a low resolution model of the actual thing. The actual thing is the code, the schema, the configuration, etc... but because we need to get on with our lives we use low resolution models.

Then the problem is, how to make sure that when I say "user", the people I talk to understand that I mean the "User domain object" and not "a human using the system" or the "user_table".

There really is no way around the need to have a calibration making sure that all have a more or less similar mapping word => actual thing.

How much calibration is needed, how to go about it not wasting time and not offending people is tricky.