Towards fast metamodel evolution in LiquidML

Abstract

The software industry is applying Model-driven development approaches due to a core set of benefits, such as raising the level of abstraction and reducing coding errors. However, their underlying modeling languages tend to be quite static, making their evolution hard, specifically when the corresponding metamodel does not support primitives and/or functionalities required in specific business domains. This paper presents an extension to the LiquidML language to support fast metamodel evolution by allowing experts to abstract new language concepts from primitives while supporting automatic tool evolution and zero application downtime. To probe our claims, we evaluate the evolutionary capabilities of existing modeling languages and LiquidML in a real world language extension.