Model-Driven Application Platforms
A model-driven application is a software application that the functions or behaviors are based on, or in control of, some evolutionary applied models of the target things to the application. The applied models are served as a part of the application system in which it can be changed at runtime. The target things are what the application deals with, such as the objects and affairs in business for a business application. Follows the definition of application in TOGAF, a model-driven business application could be described as an IT system that supports business functions and services running on the models of the (things in) business.
The ideal of the architecture for a model-driven application was first put forward by Tong-Ying Yu on the Enterprise Engineering Forum in 1999, which have been studied and spread through some internet media for a long time. It had influence on the field of enterprise application development in China; there were successful cases of commercial development of enterprise/business applications in the architectural style of a model-driven application. Gartner Group carried out some studies into the subject in 2008; they defined the model-driven packaged applications as "enterprise applications that have explicit metadata-driven models of the supported processes, data and relationships, and that generate runtime components through metadata models, either dynamically interpreted or compiled, rather than hardcoded." The model-driven application architecture is one of few technology trends to driven the next generation of application modernization, that claimed by some industrial researchers in 2012.
F.A.Q about Model-Driven Application Platforms
What is Model-driven development?
Model-driven development (MDD) is a format to write and implement software quickly, effectively and at minimum cost. The methodology is also known as model-driven software development (MDSD), model-driven engineering (MDE) and model-driven architecture (MDA).
The MDD approach focuses on the construction of a software model. The model is a diagram that specifies how the software system should work before the code is generated. Once the software is created, it can be tested using model-based testing (MBT) and then deployed.
What are the benefits of model-driven development?
The MDD approach provides advantages in productivity over other development methods because the model simplifies the engineering process. It represents the intended behaviors or actions of a software product before coding begins.
The individuals and teams that work on the software construct models collaboratively. Communication between developers and a product manager, for example, provides clear definitions of what the software is and how it works. Tests, rebuilds and redeployments can be faster when developing multiple applications with MDD than with traditional development.
What are the core concepts of model-driven development?
Model-driven development is more in-depth than just having a model of the software in development, which makes it different from model-based development. Abstraction and automation are key concepts of MDD. Abstraction means to organize complex software systems. In MDD, complex software gets abstracted, which then extracts the easy-to-define code.
Once developers transform the abstraction, a working version of the software model gets automated. This automation stage uses a domain-specific language (DSL), such as HTML, and scripting languages, like ColdFusion, which can integrate other programming languages and services -- .NET, C++, FTP and more -- for use in websites. DSL is a language specialized to an application domain. A model is written in a DSL language and is utilized for transformation in coding language from the model to working software.
Agile software development methods are often paired with MDD. The Agile development approach enables short sprints where the project scope can change. Agile model-driven development (AMDD) establishes short development iterations, while changes can be redesigned and shown on the model. In AMDD, design efforts are split between modeling in both sprints and coding.