I wrote about this in 1996, published in IBM Systems Journal.  The object-orientation craze was in full swing, hence the title.  But that is not really the point, but rather paying detailed attention to what people say in the business.  I'm sending links to the published article, and an unpublished paper that supplements it.

https://app.box.com/shared/lc98a6tnca

https://app.box.com/shared/79q8tcgk3g

DEar Dough I was not able to get your articles. Please provide me another link

Dear Mussa -- I'm not sure what the problem is, but let me try attaching the files to this note.

UML is just a language which can be used to describe both business processes and systems architectures. 

One approach would be to look for a modeling paradigm that could bridge the divide:

Alternatively, one can consider how to align the respective paradigms:

https://caminao.wordpress.com/about/community/modeling-paradigm/

https://caminao.wordpress.com/2014/10/04/empathy2abstraction/https://caminao.wordpress.com/2014/09/15/alignment/

First of all, let's have a look at the terms you used. So SBVR is an OMG standard aimed at, well, standardizing how people express business vocabulary and business rules in structured natural language. The idea is that such vocabularies and lists of rules could be understood both by non-IT people and computers (i.e. specialized CASE or BPM tools). Such business vocabulary (a.k.a. glossary) and rules models can be part of the outcomes of the following processes:

- business analysis of a certain domain (no software is being developed),

- certain software system requirements specification (innitial phase of system development),

- system design (detailed description of software so that it can be implemented by the programmers or alike).

I think SBVR is best suited for business analysis and requirements specification, although if we speak only of vocabulary, it can be used throughout the information system life cycle. However, one must not forget that vocabulary [and business rules] would only be a part of the models you create.

Now, UML is another OMG standard aimed at providing a graphical language that could be used to specify requirements and/or design systems. It can be extended using profiles and has extremely wide applicability. UML as such is neither a business, nor a system model. It's just a tool. You can create information systems models - both for requirements specification and design - using UML. Adapted UML can be used during business analysis (e.g. you can define business use cases), but it is not it's primary purpose.

The important bit here is that UML does not include vocabulary or business rules capturing capabilities. Business rules can be expressed in OCL (another OMG standard), but it's extremely far from structured natural language and cannot be understood by non-IT stakeholders. Therefore, SBVR supplements UML allowing (at least in theory) analysts to create vocabularies and rules that can be integrated with UML models. This is possible because both UML and SBVR use the same meta-meta model - MOF. Besides, both UML and SBVR are part of the same model driven architecture (MDA) paradigm.

Actually, me and my colleagues are working on integrating SBVR, UML, and BPMN2. It's not easy, but doable.

Using meta-models to combine standards induces rigidity and overheads. With regard to SBVR the alternative is to use UML stereotypes :

https://caminao.wordpress.com/about/sample-problems/eu-rent-car/ 

Remy, that's exactly what we did (used UML stereotypes to implement SBVR metamodel). In fact, as you know, many CASE tools are built on UML metamodel and all additional languages and notations are implemented on this metamodel. We use MagicDraw, btw.

There is a problem of terminology here: whereas meta-models are used to translate from one language to another, stereotypes are used within a language and therefore have nothing to do with meta-models.

https://caminao.wordpress.com/2012/09/17/uml-us/