Is a brick class, for a point, required to be unique for that device?

full disclosure: I'm new to Brick/SparkQL

What I'm trying to do is resolve whether a unique device name is required to ensure compliance with BRICK. If the brick class is unique I would think it would negate the need for a unique device name.

Some IoT devices doesn't provide devices names. In that case, I'm assuming I'd have to contrive a unique device name, if that's required.

Simple example: An air temp sensor (ats-1) in a room (r1).
-> Using classes ats-1 can be designated "r1:ats". Note that the location or containment hierarchy makes unique the "leaf" node, which is a class.
-> Using device name, it can be designated "r1:ats:ats-1". In this case, I'm using a unique device name.

More complex: 2 air temp sensors, ats-1,2 in a room (r1). "r1:ats" won't work because the location/containment hierarchy no longer makes the class (ats) unique. But the following does...
-> class designation:
r1:hvac_zone1:air_intake:ats
r1:hvac_zone1:air_output:ats
-> The equivalent device name designations would be
r1:hvac_zone1:air_intake:ats:ats-1
r1:hvac_zone1:air_output:ats:ats-2

Both ways "work". The question is, is the class designation a valid way to obtain uniqueness or is a device name required by BRICK schema?

Can anyone give me any insight into this validation problem?

NOTE: Yes, I know this isn't valid syntax. But this isn't a syntax question.