Domain orientation or subject orientation?

Martyn Jones, New Jersey

Narrator: Here, in this piece, we wander down Differential Avenue to look at what people consider domain and subject orientation. This story is about the good, the embarrassingly lousy hyperbole and then the ugliest provocative nonsense.

Dud: I was listening to this bloke ask the data mesh guru about the difference between domain orientation and subject orientation, and she said, “Well, with the subject orientation, you get all these people in a room to come up with one massive entity-relationship model of the businesses’ data, and with domain orientation you build everything iteratively.” Therefore, you either boil the business data ocean or take an iterative approach.

Pete: Blimey! Is that what she said? What a bloody joke! There you go, Dudley, that epitomises the wisdom of a callous youth, one spent hacking Java, the knowledge of feckless opportunism, and a total misreading of vital information in the world of all things data.

Dud: Like, err, Norman Wisdom.

Pete: Yes, that is it. Exceptionally good, Dud. I can see that you have been swatting up on the subject. Anglo sense, Saxon insight and Norman wisdom… it’s all in there.

Dud: In your view, what are the critical differences between domain and subject orientation?

Pete: Blimey, Dud, that’s quite a sea-change in your line of investigative abstraction.

Dud: Thanks, Pete.

Pete: Look, Dud, it’s all about warm and fuzzy terms. Like things that tickle the fancy, get the attention and generate interest. Dud, let’s say that domain orientation in data and analytics covers a multitude of interpretations, misrepresentations and sins. It can, for example, refer to data and information found within a particular business domain or field of investigation. But this can happen on several levels of abstraction. We can discuss domains such as healthcare, finance, education and digitalisation. However, we can also look at domains within healthcare, for example. Purchasing, real estate management, research, manufacturing and IT could also be seen as data domains in their own right.

Dud: And subject orientation?

Pete: Well, partly to reiterate, my view of subject orientation is entirely personal. I take my lead on this from Bill Inmon, so it’s an accessible mainstream interpretation.

Dud: Go on, Pete.

Pete: Subject orientation with data typically involves structuring and managing data and information models around in-focus entities and subjects of interest, intending to produce artefacts and perspectives that are closer to the real world and, as such, are much more intuitive and friendly for real people who interact with and use these subject-oriented artefacts and models. Subject orientation of data provides many features and benefits, including intuitiveness, user-centricity, simplified data management, consistency, scalability and flexibility, performance, and easier integration, as well as helping to enhance collaboration, modelling and semantic interoperability.

Dud: It also helps with object-oriented data modelling, right, Pete?

Pete: That too, Dud! That, too. But I would add this. Both domain and subject orientation can be used as models to provide data to feed analytics. The difference is that domain orientation is like fried food, whereas subject orientation is like a healthy and balanced diet.