We declare actual observations by using the DATA: command.
In our simple example there are but two pieces of data
(observations of 
and 
) which we declare
to [B/D] as follows:
BD>data:D1(1) = 5.4 
BD>data:D2(1) = 9.8
Here, the notation D1(1) simply means the first observation on D1. Had we more than one observation, then we would be defining D1(2), D1(3), and so on. You can think of the data as being stored in a large matrix where the columns represent different quantities, and the rows represent different cases. We generally refer to the columns as data vectors. (The quantities need not have the same number of observations, and observations can be ``missing'', but you might need to be careful when mixing such data vectors.) We can examine our data specifications by issuing the command
BD>look:(d)
giving the output shown in figure 9, listing all the elements for which data has been specified, together with their data values.
