We believe that with respect to the manner in which control problems are formulated, a standard is both feasible and desirable. The matrix notation of Matlab-like languages is so natural that we do not see any need for another notation in the future. Although Matlab's division operators "/" and "\" for right and left matrix division are certainly not standard operators in the classical mathematical sense, they are very convenient and easy to get used to. In Impact, a Matlab derivate, we suggested additional operators besides "," and ";" for a third dimension, thus Impact effectively operates on complex tensors in place of complex matrices. Multivariable systems can be expressed in terms of polynomial matrices. The "^" operator separates polynomial coefficients, while the "|" operator separates polynomial roots.
Also with respect to the embedded procedural language, a standard can be achieved. CTRL-C, another recent Matlab dialect, is very powerful in this respect. It basically extends the Pascal programming style, operating conveniently on matrix data structures. Very useful, for instance, is the extension of the Pascal-like for statement. Impact employs an Ada-style instead of a Pascal-style. It actually hardly matters which style is adopted in a forthcoming standard, but a standard must be found in order to allow smooth exchange of the extensive available soft-coded macro libraries between different CACSD software systems.
Also with respect to the user communication interface, de facto pseudo-standards have already been established. Window interfaces look more and more similar to the Macintosh interface. The Swiss mouse is a very convenient, flexible, and fast input device. To our displeasure though, there exist mice with one, two, and three buttons. Any standard would be equally acceptable, but a standard must be found. Once the fingers are used to one system, it is very hard to adjust to another.
Another interface, which is rarely even noticed by the casual CACSD software user, is the interface to a data base where results of computations, as well as programming modules, notebook files, etc. may be stored. To promote the state-of-the-art of CACSD software further, it is imperative that a data base interface standard be defined as soon as possible. Lacking such a standard, most current CACSD software developers simply rely on the file handling mechanism (directory structure) of the embedding operating environment. This results in a jungle of small and smallest data and program files scattered over different subdirectories which makes it very hard to retrieve data and programs.
With respect to the actual functions offered, we shall probably not see a standard quickly. The current diversification into different application areas and design methodologies is most likely to be around for some time, and we actually welcome this, as too early a standard can freeze the lines and hamper the introduction of innovative new concepts.
Interested in reading the full paper? (1 page, 74,884 bytes, pdf)