These mixed approaches must be relevant for analysis of appropriate (i.e. mixed) classes of systems. Therefore, to disclose possibilities and possible disadvantages of such integral view on systems modelling, we need to consider several examples of mixed systems. Are they exist in the real world What about its properties and features Is such way of generalization in systems modelling just artless composing of existing approaches without principal innovations Or we can expect more efficiency of such generalizations
First of all, we need to note that the gap between the hard and soft approaches is more impenetrable then the gap between these ones and the failures approach. This is caused by obvious differences between hard and soft classes of systems. Indeed, hard systems are useful for problems that can justifiably be quantified. Suitable approaches involve numerical simulations and often use the techniques of operations research. Such approaches cannot easily take into account unquantifiable variables, e.g. opinions, and may treat people as being passive, rather than having complex motivations. On the contrary, soft system models are usable for systems that cannot easily be quantified. Suitable approaches can be most effective for understanding motivations, viewpoints and personal opinions, and interactions and addressing qualitative as well as quantitative dimensions of problem situations (Bode & Holstein, 2003). Thereupon Peter Checkland says: "Our experience in developing soft systems methodology is that the world is highly complex and mysterious, and far more complex than any of our ways of notating it And what I'm talking about here is the fundamental distinction between hard and soft systems thinking. The hard systems thinker takes system as in everyday language to be a label word for complexity in the world. The soft systems thinker is perfectly happy to make a conscious use of the hard systems ideas." ("Systems Practice", 2004).
However, usage of a soft systems approach to systems modelling and problem solving does not necessarily mean that hard systems approaches are no longer of value to the systemic analysts. At systems development and applications layers we can see lesser opposition then at abstract layer. For instance, while hardware engineering typically deals with just hardware and software engineering with just software, the systems engineer is responsible for seeing that the software properly operates on the hardware, and that the system composed of the two entities is capable of properly interacting with its external and often non-quantifiable environment, especially the user, while performing its intended function (Bode & Holstein, 2003). At the application layer there are no neat boundaries between hardware and software systems engineering. All the relatively neat boundaries between hardware and software systems may be violated, depending on the people and program. For example, firmware embedded into a microcontroller or a GPU is often assigned to the hardware engineer, but it is essentially a software development (as in OpenGL or DirectX applications) which is based upon interconnections with the operating system and end user.
Such cases can be considered as a shift from the well-quantifiable systems to the non-quantifiable systems which is caused by implication of unmanageable factors or unpredictable human-based subsystems. It