Why I don't need a Bounded Context 
DDD practicioners proudly call themselves “domain linguists”. There is much emphasis on the words we use to describe a domain: language is essential for understanding the subject, defining problems and solutions and communicating about it. But what about our own domain: DDD? We have our jargon with fancy words like “ubiqitous language”, “bounded context”, “value object”, “aggregate root” or “event sourcing”. Is their meaning precise and consistent? Are there any redundant concepts? Do they always serve the goal of communication well? Or can those basic tools be improved? In part 1 of this series I investigate the concept of "bounded context".
Definition of Bounded ContextIn his DomainDriven Design Reference (current latest, from March 2015, page vi) Eric Evans defines a bounded context as: "a description of a boundary (typically a subsystem, or the work of a particular team) within which a particular model is defined and applicable". This is an odd definition. The adjective "bounded" gives the impression it is a kind of context: a bounded one, as opposed to an unbounded one. This "definition" however states a bounded context is a description: a description of a boundary. I'd not call that a "bounded context", but a "context boundary". The context was mainly called "bounded" to emphasise that it has boundaries. A context is a modelWhen you describe (selected aspects of) a domain, you are modeling. The choice of what aspects you select is part of the modeling process. To describe a domain you need a language. Using language is modeling: the words you use are always an interpretation of what you describe. That language, used in modeling, is what is called the "ubiquitous language" in DDD. It is the language of the model, an intrinsic part of the model. If you have a domain and a model of that domain, then what is that context in which the model is applicable? It cannot be the model itself, for that would be a circular definition. So it must be something wider than the model for which it is a context. I always saw a "bounded context" as a part of the domain. But when you are describing that context and its boundaries, you are describing the domain... and so you are modeling. You cannot describe the domain without modeling. The context is a model too. The boundaries you describe can be boundaries that exist in the domain, but even if that is the case, the boundaries of a context are always an interpretation, a design choice: you describe the boundaries of a model. Occam's razorSo we have a (domain) model and an outer model, called the (bounded) context. That outer model is used to define the boundaries of the inner model. Let's apply Occam's razor to clean this up a bit: if I define the boundaries of my domain model, I don't need a (bounded) context. I have a model and that model has boundaries. That's it. I can clearly communicate that to all stakeholders using the model. There is no advantage in using a "bounded context" besides the domain model.
Ubiquitous languageHow ubiquitous is a "ubiquitous language"? It is only applicable within the boundaries of a model. That is why I use the word "model language" instead of the fancy expression. Something that can only be used within limits is a bit the opposite of "ubiquitous". It emphasises that the language is used by both the technical and nontechnical modelers. But the use of that word is not very beneficial for the communication in my experience. We only have to convince the whole team to use a common model with a common model language. This language is only defined and applicable within the (domain) model. Our model is the context for our model language.
What's in a name?You might say that the exact words don't matter, as long as we can satisfactory use our current DDD language to communicate about modeling domains, as we are doing for over a decade now. What's in a name, anyway? But because language is the carriage of our thoughts, according to the linguistic relativity theorem even influences our thoughts, we might improve our analysis, ideas and communication by improving our language. Especially when communicating with noninsiders in the DDD jargon.
(last edited: August 29, 2016) Also posted on DDD/CQRS mailinglist. Discussions and feedback are possible there: https://groups.google.com/forum/#!topic/dddcqrs/X3v_b7lxOak
