You don't have to sell me on the value of a good naming convention!
As a past data administrator, I've had many discussions on appropriate names for attributes. It seems that every place I go, they have words that mean different things to different users!
That is why we need to shield the users from the naming conventions we choose for database objects. Better to provide the users with a meta-data dictionary and thesaurus. Hopefully, you can integrate this into their adhoc reporting solutions.
From a DA point of view, I've always found two areas that were particularly sensitive to naming: cascaded foreign keys and administrative attributes.
Since many CASE tool can automatically cascade the foreign keys, you have to worry aboutthe children attribute names versus their parent's!
Administrative attributes are usually common to many entities (last updated on, last updated by, etc.). We, of course, have to worry about whether or not the names should be unique (for example,by using an entity-specific prefix) or not.
Another debate is whether or not an entity's attributes should all have a common prefix when implemented physically.
Finally, you want to balance information redundancy with developer-friendliness. For example, why should I call a foreign key constraint FK_child_parent_column when I can find all that information in the DBMS data dictionary? But it's a lot easier that figuring out which of the umphteen SYS000nnn constraints is the one the DBA is interested in.
Let me know what other issues you've had to tackle in regards to data/database naming conventions.