(ref.doc)bs 110593

[email protected] (Tucker Taft @ Intermetrics Inc, Cambridge MA) writes

 > I will certainly agree on the fundamental point that C++ and Eiffel use
 > the "class" construct to define both a module and a type.

In the case of C++, I wouldn't. In C++, a class is a type. The ``a class is
a module'' statement is something that seems to be imposed on C++ by people
discussing it from ``the outside'' and I find little support for that view
in my writings on the design of C++. On the contrary, I have repeatedly over
the years pointed out that C++ does not have a proper module concept
and that simulating it with source files and classes are workarounds. 

 > It is interesting to
 > see the debates about adding a "namespace" concept to C++.  Clearly there
 > is some recognition that neither classes nor files are the ideal
 > "high-level" organizational structure in some cases.

as there has been at least since 1981 or so.