Review: Scientific and Engineering C++

A good terminological suggestion p 304:

A template class results from fixing the parameters of a class template.

12.7 Function-Structure Categories, p 354
This paragraph deals with a specific use of a pattern named elsewhere name injection (terminology maybe conflicting with a use in the ISO/ANSI standard) or derived-to-base feedback.

The mechanism we introduced in the previous section can be viewed as more than a trick to control global template expansion.

12.9 Notes and Comments, p 357
Two notes about the meaning of type in C++ and in other languages

12.3 A type can also be defined as the set of all objects that have a certain behavior. Such a view -not taken by C++- leads to a language without programmer specification of typelike classes.
12.4 Our description of interface or use types in terms of a partial specification of behavior matches the substitutability of types principle in [Wegner & Zdonik - ECOOP 88] and in [Nierstrasz - OOPSLA 93]. The create/use dichotomy evident in C++ and other languages is being formalized in new theories of type, for example [Liskov & Wing - OOPSLA 93]. (These works attempt to systematize types in a wider context than static, sequential object-oriented programming.)
Marc's note: [Nierstrasz] is not a "wider context"... It is a different, largely incompatible context (the contrary was probably not meant, but I stress it anyway). The "attempts to systematize types" have IMHO failed largely: the concept of type as in "static, sequential object-oriented" programming is still largely incompatible with this in use elsewhere, e.g. in OOD!
Table of contents
Marc Girod
Last modified: Thu Mar 5 13:57:30 EET 1998