Getting the semantics UML models used in a project, their consistency relations, integration and refinement (by transformations), precisely defined would be the first step. These definitions should be considered in relation to the verification and validation techniques and tools you use. We have quite a number of publication in relation to the rCOS method of component and OO systems development. -- Rushed

Thank you Prof. Zhiming.

 Hello, take a look at this paper using UML and finite automata

https://www.researchgate.net/publication/280302139_Anlisis_de_consistencia_de_Casos_de_Uso_con_simuladores_de_autmatas_finitos

or using patterns form the business model

https://www.researchgate.net/publication/280302061_Implementacin_de_Patrones_en_la_Validacin_de_Modelos_Conceptuales?ev=prf_pub

Conference Paper Análisis de consistencia de Casos de Uso con simuladores de ...

Conference Paper Implementación de Patrones en la Validación de Modelos Conceptuales.

Hi Omar,

In my opinion it would be benefitial for you to look at systematic reviews papers in the field.

Please consider as a starting point:

Carlos A. González, Jordi Cabot, Formal verification of static software models in MDE: A systematic review, Information and Software Technology 56 (2014) 821–838,

Cabot et al., On the verification of UML/OCL class diagrams using constraint programming, The Journal of Systems and Software 93 (2014) 1–23.

Maybe following topic would be ineteresting for you to expand look:

Mazin Saeed et al., Empirical validating the cognitive effectiveness of a new feature diagrams visual syntax, Information and Software Technology, Volume 71, March 2016, Pages 1–26.

Best regards,

Tomek

Contrary to popular misconception, the core subset of UML has a formal semantics definition that has been standardized by the OMG. It is called the "Semantics of a Foundational Subset for Executable UML Models" (http://www.omg.org/spec/FUML/1.1/PDF).

This uses an operational approach to define the semantics of basic things in UML such as classes, objects, associations, links, etc. It is itself based on a subset of itself (called "base UML or "bUML" for short), that is defined using a formal first-order logic language standardized by NIST.

This core semantics specification, colloquially referred to as "fUML", was subsequently used to define the precise semantics of composite structures (ports, parts, connectors, etc.), which is another OMG standard (http://www.omg.org/spec/PSCS/1.0/PDF). Finally, at this time, work is progressing on using this same approach to define the precise semantics of UML state machines.

So, for all those that claim that "UML is a notation without semantics" (I've heard this fallacy repeated endlessly), they are just plain wrong. It not only has a semantics, but those semantics are defined in a matheamtically formal way.

Given that these are all adopted standards by the same institution that owns the UML standard, all work on formal verification efforts of UML models should use the semantics defined therein for their models of UML. Otherwise, they will be dealing with a proprietary and idiosyncratic formalization of UML.

I like Bran's answer. It is not quite fair to criticise UML has no semantics. There are tow points I would like to make:

1. there are a number of definitions of semantics (including the one that Bran mentioned), and each of them may have its own features (or different aspects of abstraction).

2. Also it is not all an disadvantage that UML does not have restricted semantics. Different users in different application areas may like define their semantics.

We present here a homemade tool to work with UML requirements

https://www.researchgate.net/publication/284633757_Herramienta_para_administracin_y_validacin_de_requerimientos_de_sistemas

Article Herramienta para administración y validación de requerimient...