Recently introduced a little of the concept of model-driven mechanism (MDM) on the blog, including its important form, multiple MDM. These topics, however, does not seem to be concerned. It is not theory but the basic principles about constructing a system. After they are initially gotten in the process to try to solve the practical problems of enterprise applications, it is found that it has a broad universality or generalization. In understanding, this universality is first due to the universality of the concept of model; on the other hand, MDM answer one of two fundamental questions: how models work (another fundamental question is what are models), and many of signs suggest that MDM may be just the fundamental answer to the question, it is the basis of more different answers.
The question about how modes work, can generally be summed up as model working mechanisms (MWM). As far as I see, it appeared there are not quite much investigations on this issue. As the uses of the term modeling, it is more on that get and build a model, and it is often ignored that, how does an entity serve as a model in a system. In addition, it can be seen that, in computer and the applications, much researches to models and modeling are related to constructing and using a language (DSL direction perhaps takes this to the extreme) and also the relevant modeling process and tools. MDE is involved in the MWM, but almost all concentrated on model transformation. Just as we have pointed out, it is indeed a use of the MDM.
There are in fact much about models and modeling in computer and the applications, and it is quite fragmented. In my opinion, MDM is something the same behind them; in other words, the understanding of MDM can help us discover how to combine the efforts dispersed in these different areas, such as
- Database. The architecture ideas to database application. The matching of OO and RDB. There are many sayings about RDB is dead (or at least obsolete) , may quite a few misunderstandings, but of course, the architectural thinking about database since E. F. Codd is really stagnant for a long times, yet needs to be a breakthrough.
- The data modeling, information modeling developed on the database (esp. relational model and RDB), it is closely related to above issues.
- Often very ambiguous conceptual modeling. It involves a lot of things, may worth to list alone.
- Ontologies and the application. It should very useful, if it out of the altar of artificial intelligence (AI) and knowledge representation.
- Enterprise modeling (EM) and enterprise architecture (EA). They are representative of the left-side approaches, and we suggested that MDApp should be the inevitable way to take it to be grounding .
- Workflow and BPM. BPM is actually a good example to MDApps, and MDM should be the key to solve many of controversies in it development.
- A more specific example such MVC (model, view, and control). Using MDM can bring such the architectures clearer understanding.
- Even to the basic programming architecture or paradigm, such as object-oriented, process-oriented and so on, MDM could also bring some fundamental thinking.
- And so on.
In fact, it can be strongly felt, MDM is a juncture, a door; some what important things behind it, is about the missing left-hand side theory which pointed out by Brian C. Smith (1985) . The in-depth research may bring a wealth of amazing results.
 Yu, Tong-Ying. “Model-Driven Applications: Using Model-Driven Mechanism to Bridge the Gap between Business and IT.” In Advances and Applications in Model-Driven Software Engineering. Díaz, V.G. et al. eds. IGI Global, August 2013 (in press, see Recent Work on Model-Driven Applications (MDApp).
 “we have no theory of this right-hand side relationship.” (i.e., between model and the embedding real world, see this post), from Smith, B. C. “The limits of correctness.” ACM SIGCAS Computers and Society 14, no. 1 (1985): 18-26.