UAF or SYSML?

Presenters: Laura Hart, Research Engineer Senior Manager, Lockheed Martin; Gene Shreve, Senior Systems Engineer, Integration Innovation, Inc.
As MBSE practices mature in organizations, the question of "should we use UAF or SySML?" often arises. This question indicates that many engineers don't understand the relationship between these two modeling languages and how they can and should be used together to answer different questions from various stakeholders. This paper will provide a brief overview of each language, its intended purposes and then provide a working example of how the two languages complement each other, aiding the decision process at every stakeholder level.
Become a member: www.omg.org

[@ms1923]

For the segment starting at 12:00 on "Enterprise vs Systems Architecture": Considering the definition of a "system" from ISO 15288 or the INCOSE Systems Engineering Handbook, one might ask: Isn't an enterprise also a system? Another perspective to consider: In reality, the term "system" is a conceptual construct or sometimes just a suffix in a product name used in everyday language. The concept of a system doesn't exist as a tangible entity; rather, it describes certain behaviors and relationships. Some assemblies of objects exhibit "system-" behavior within the processes they are involved in. In these scenarios, they take on what we can call a "system role." A "system role" is a role that emerges within a “composition of elements” where, in at least one arrangement of these elements and within the specific process they engage in, the composition exhibits behaviors that the individual elements alone do not. Therefore, an organization, a collection of engineered artifacts, or parts/elements of nature can take on a "system role" in certain processes. However, in essence, these compositions are simply what they are-assemblies of components with potential behaviors, not "systems" in and of themselves.

[@ObjectMgmtGroup]

It is a misconception to think of a system as a thing. System is really more like a way of perceiving some thing, similar to the way a collection of things can be perceived as a “set” using the methods of set theory. The way we perceive a system is based on systems theory which is based on systems philosophy. www.systemsphilosophy.org/about-systems-philosophy In common speech, most people think of a system as a thing but that it is an illusion if you properly consider systems concepts from systems science and philosophy. Systems engineering has for a long time adopted the “common speech” about systems which has impeded progress in the world of SE practice. A group of INCOSE Fellows conducted a study a few years back to investigate this topic and proposed the new definition of system for INCOSE. www.incose.org/about-systems-engineering/system-and-se-definitions An enterprise is a human venture. In common speech, especially in the IT world, an enterprise is equated with a large organization that IT supports. They talk about “enterprise systems” which really confuses things; what they mean are IT systems and software that is used across the whole enterprise, eg travel accounting, payroll. So, an enterprise can be perceived “as a system” but that does not make the enterprise a system itself. When you look at an enterprise as a system you are forced to look at it through a systems lens. But if you instead look at an enterprise through a sociological or political or economic lens then you can see things that the systems viewpoint cannot perceive.

[@ms1923]

​@@ObjectMgmtGroup I believe we share a common understanding of the term "system" and recognize its frequent (mis)usage in certain communities. (By the way, I've been a member of INCOSE since 1987.) From an ontological perspective, anything that exhibits system behavior (at times, e.g., an engineered system while operating) must be included in our entity taxonomy. As a matter of worldview, I choose to exclude concepts, sets, and possible worlds from my hypothetical ontology. This leads me to define "system elements compositions" as a collection of material or immaterial entities that, at times, exhibit system behavior and assume a type of system role (e.g., operating system role, enabling system role, etc.). Moreover, my initial comment was about my objection to the de facto differentiation between Enterprise Architecture and System Architecture, as seen in ISO/IEC/IEEE 42010 Software, Systems, and Enterprise - Architecture Description. If these terms refer to entities that can exhibit system behavior, why not use "Enterprise Architecture" or "Engineered System Architecture" as specializations of "System Architecture"?

[@jaHorsman92]

@@ms1923 Interestingly if you analyse ontological use of language of entities you find often a singular entity never has the system word added to describe it. For example a human is not a system or referred as one in biology. Instead a human has systems, i.e. a collection of parts that deliver something of value to the whole, or a human is part of systems. Both cases it implies that most humans intuitively understand, even if they do not recognise it, that a system is more than one real or conceptual entities that are causally integrated and have some boundary. I also agree with your statement about EA and Engineered Systems both being specialisations of system architecture.

[@luigiturco8492]

Interesting. I see often choosing UAF over SysML for SoS architecture, and by consequence overlooking the definition of common interface standards

[@llvienna]

Yes and yes. Continuum Depends what you are doing.