we agree with your proposal of including our discussion as an appendix to the first draft of the benchmark. Below you can find the latex source with small corrections.
Let us discuss here some aspects about the problem of identifying actual
entities in the real world within the proposed database instance.
First of all, it was said that our two employees are identified by ED and DI, respectively. Are those a sort of surrogates or not?
\noindent If they are, how can we use them in formulating queries and in
results of queries? (They do not appear in the proposed taxonomy).
\noindent For instance, in the given examples, employee names are allowed to change over time and it is assumed that the system is responsible to mantain the links between the tuples referring to the same real-world entity. However there
is no mean to distinguish in the query result between a person who has changed
name from two different persons. If we consider the query Q~2.1.1 ``{\em Find the names of employees that have been in a department
named Toy for a shorter period than has DI}\ ", how can we establish that
Ed and Edward are two different names of the same person, rather than being the names of
two different persons?
\noindent A similar problem occurs if the user wants to know something about the whole story of an employee formulating a query knowing only his name at some point in time. It would be reasonable that the system will retrieve all the tuples with the specified employee name and the tuples with the previous/successive names of the same employee. But in this way we do not allow the system to retrieve the tuples which refer to the name the user has explicitly specified, {\em
only}.
We think that it would be useful to support both kinds of queries:
\begin{itemize}
\item those where one refers to all the tuples referring the same
entity even if he specifies only an attribute value valid at some point in
time (e.g. {\em Find the salary of the employee ED});
\item those where one refers {\em only} to the tuples that
precisely satisfy the selection predicate (e.g {\em Find the salary of
the employee whose {\bf name} is (or was) ``Ed"}).
\end{itemize}
The problem is that we need a way to distinguish between the identifier
ED (whatever it is) and the string ``Ed" representing the name of an employee at a given
time, both in query formulation and query results.