(Data) Modeling for your audience

Understanding the reach of your models

Most modeling artefacts (data, process, technology) that are created by humans for a specific purpose, usually outlive this purpose to become more that what their intention was. They will often be red, used or repurposed far beyond the original author's intent or even beyond it's usefulness or applicability just because it is available. For example, a database schema created for one time use by an DBA of a operational system is then used by all kinds of analytics teams to understand the source system, even if the schema does not lend itself well to such purpose. It is often because of the lack of other documentation that people resort to using these models.

But it creates a lot of risks for people who use these models. Reading and understanding a model is critical for using it safely. But it might not be meant for the use case you are looking for. While you can create lively documentation for each model, a better approach is to create a simple labeling scheme so users can easily understand if this model is made to be understood by them. Apart from all kinds of functional classifications of models, a more human centric labelling will go a long way in understanding the risks of using such models.

Counting eyes

My suggestion is a simple classification scheme to help us understand the usefulness of model artefacts. I call it the Eye-counting classification. It assumes that if you can count the number of eyes that your artefact should support, you understand the quality and verboseness that your artefact needs to have. This in turn helps you assess the amount of work, how well structured and documented your artefacts needs to be. It also helps with selecting appropriate modeling methodologies and notations for your artefact.

1. For My Eyes Only (FMEO)These artefacts only have immediate use for the creator, and have no long term usage. Ideal for classifying sketches or doodles. Any notation will do here.

2. For Your Eyes Only(FYEO)These artefacts are there to communicate between 2 persons/specific parties, but are not useful outside these parties. These 2 persons can be the same person over time. Without the communication context and understanding the parties, these artefacts are not really useful. A lot of modeling artifacts used for development in small teams fall in this category

3. For Our Eyes Only(FOEO)These artefacts communicate inside a homogenous group, for example an application development group. The artefacts are maintained by and for the group. Outside the group these artefacts are not really useful. It is important to standardize on approaches well know to the target audience like UML or ER modeling. Larger development teams should strive for this level.

4. For Those Eyes Only(FTEO)These artefacts have a specific target group in mind such as all executives (in organization X), all developers or all business analist. The group needs to be well defined but can be quite large. De creators don't need to be part of this group (usually they aren't). Here verbose models like Fact Based Modeling work quite well. This level is required for analysts or requirement engineering.

5. For All Eyes To See(FAETS)These artefacts are fully self describing, carry their own context/metadata and meta-metadata as to be understandable by arbitrary persons. That does not mean that will be easy, just that all knowledge to read the artefact is provided (or referenced) from the artefact itself. This level of quality is mandated for things like laws and regulations, and hence should apply to all model artefacts that represent such laws and regulations. Here only approaches like de #VDA of the dutch tax authority are valid candidates. In de #VDA we can be not only verbalize the assertions, but we assert the assertions themselves (and we can assert those as well), creating a intricate set of customizable interrelated speech acts #data #technology #communication #quality #regulations #modeling #datamodeling #informationtechnology #regulatory #wetsanalyse #FMB #FactBasedModeling #ORM #UML #ER #FCOIM #speechacts #tax #belastingdienst