Database Software

We have a facts table with some of the columns identifying dimensions, and value column with measurements data. But there is one more column - parameter_id, which identifies what we are measuring. For example, we may measure weight or length, so measurements are not comparable with each other, like in classic OLAP examples where sales are in dollars and can be compared. Should we in this case create a dimension for parameter and have a single measurement, or create a set of views where every view selects a subset of measurements where particular parameter is measured, and create a measurement for every parameter?

Structure of data:

date object_id parameter_id value

Here object_id is an id of an object whose parameter is measured (for example, we are measuring students in a group, then object_id identifies a student being measured - all examples are hypothetical, actual data is from another domain) - definitely a dimension, date is date when measurement has been done - definitely a dimension, and parameter_id is what has been measured (several values per date-object_id combination). For example, let's assume parameter_id=1 means weight, parameter_id=2 means length:

23-sep-2006 1 1 73.4
23-sep-2006 1 2 178

Object 1 has been measured, its weight is 73.4 and length is 178. Should we create 2 views: select * from facts where parameter_id=1 (and 2 for second view), define 2 parameters - weight and length, and map them to their views, or make parameter_id a dimension? In the future we may have more parameters, that's why such structure of facts table has been chosen.