Objects for different roles ​

In relation to any fragment of the world (background), different agents perform multiple roles (and sometimes one agent can perform several roles simultaneously). In different roles, the agents' thinking will highlight objects from the same piece of the world differently: either completely differently, in other boundaries, or in the same boundaries, but paying attention to other characteristics, even giving them different names. Worldviews for different roles will differ, but in the world, they will occupy the same place.

Now imagine yourself in the same kitchen, but in a different role, not as a cook. For example, in the role of an interior designer. Or in the role of a parent.

The interior designer highlights other objects because they are going to perform completely different kinds of work. The interior designer will highlight chandeliers and light fixtures, not light sources. Switches and cupboard handles will be important to them. The placement of the stove will be an important criterion for them, but whether it is connected to gas will not matter to them: the designer's task is solved if the stove is in place and harmonizes with the interior. However, to the cook, an unconnected stove does not exist as an object at all.

For a parent of a young child, the latches on cupboards and the fridge are important, but it would not occur to a cook to provide such safety measures unless they step out of the cook's role.

However, both the interior designer and the safety-conscious parent care that it is indeed a kitchen, a place intended for cooking, a place where a cook wants to cut vegetables near a hot stove. When designing the kitchen and ensuring the safety of a child in the kitchen, you have to consider the actions of the cook, i.e., you need to know their practices and objects, but not all of them, only those related to your new role (the interior designer should not take into account the interior designer's passing practices in design decisions).

When in your own role, you must understand:

• Your colleagues in other roles often (but not always) highlight objects differently.
• Your colleagues in other roles often (but not always) name objects differently, even if they are the same objects.
• Sometimes they use the same names for their objects as you do for yours.
• Even if the objects of your colleagues and your objects coincide or have something in common, their requirements for the state of their objects may still differ from yours, and their actions to change their objects may lead to changes in yours, and your actions may lead to changes in their objects.

Therefore, the most important aspect of working with any model, starting from its creation or choosing an ontology for it, is to know about those with whom you will interact regarding the objects of your model (about other roles and their practices) and to timely consider their worldview. When working with an ontology (especially when creating it, acting as an ontologist or ontological engineer) it is necessary that it is shared by some community; otherwise, what's the point of it. The wider this community, the better. If it includes representatives of several roles, such an ontology and the models that use it become much more useful.

It is important to choose culturally conditioned names, i.e., terms from some accepted systems of terminologies (we call such systems "ontics" to emphasize that they are not yet fully-fledged ontologies).

The way representatives of a specific role describe their subject area is called a viewpoint/method of describing the subject area of that role (this term is introduced by the International Standard on Systems and Software Engineering Architecture Description ISO/IEC/IEEE 42010). Without knowing the specifics of the method of description for a role, you risk at some point missing an important object for that role. What descriptions are and how the methods of descriptions (description templates) are defined for different roles, how they are harmonized - we will study further.

Ninety percent of an ontologist's work consists of identifying where important objects for someone (for some roles) are missing, finding these objects in the world, gathering people (or studying computer agents) playing roles, showing interest in these objects, gathering documents describing them, agreeing on objects, their boundaries, and names. Therefore, it is important to accurately determine all important roles in the company or project, find and use their viewpoints/description methods, and then go through lists of objects several times, looking for omissions, intersections, and correspondences of objects (as ontologists say - build mapping). If a role (and its description method) was not identified at the beginning of the work, there is a great risk of switching to the viewpoint of a similar role or an associatively related role that, despite the similarity, may have fundamentally different practices, and for which the object may be completely unidentifiable.

A simple example is creating an information model of a family. After a careful analysis of the worldview for different roles, the same object for you in the role of the father of the family will be identified as "mother-in-law Anna Petrovna," for your wife as "mom Anya," for your son - as "grandma Anya." In the project to help raise the grandson, the name "grandma Anya" most adequately reflects her role for all involved, but when buying a vacation trip with her and the grandson, she will be identified not as "grandma Anya," but as "A.P. Ivanova, passport series ... number ...".