Convetion for table name number

[khl02007]

@edeno @CBroz1 @samuelbray32

We had some discussions previously about whether table names should be singular or plural. Maybe everyone knew this, but I learned today that it should be determined by what single row represents (first comment here). For example, it makes sense to use SpikeSortingRecording because single row in the table represents a recording (or a part of one anyway). On the other hand, it also makes sense to use SpikeSorterParameters because single row in the table represents, well, parameters (multiple entities that together defined a configuration for running a spike sorting algorithm). This seems eminently reasonable to me and I suggest that we stick to this rule.

Now only if we can settle Parameters vs Params...

[CBroz1]

I'm on board with 'Table name represents one key'. Conventions are building the car while you're driving it, so it's hard to feel like 'yes, this is a convention' when there are pre-existing exceptions.

I think our current best outlining of conventions is in the contribution doc. I wrote it as descriptive, but I think it could be more prescriptive to align future pipelines