Designing any system should start with the bird’s-eye view (the what), followed by the tech-agnostic how, and finally the tech-informed how. Skipping to the last step, or doing a sloppy version of the second, is fine for a POC, but has no place in real-world production.
It still baffles me that systems are built this way. It’s like jumping straight into code without knowing what you're coding.
Maybe we should call this whole modeling style vibe modeling?
Levels of data modeling help organize data into conceptual domains or functional areas, which hopefully leads to business domain ownership. As well has understanding the dependency of entities across domains.
I woulke add that a data modeler not only should ask questions - like you quoted Lartry on, but also truly listen to the answers and stories given by the business side of the organization. If they tell you they need "three hospitals" in their data town plan the response of the data modelers should be why? tell me more? - instead of just shuffling it all upon 1 pile (abstract/generalize)
In other words - data modelers should be curious! , not providing answers when in the stage of modeling
Designing any system should start with the bird’s-eye view (the what), followed by the tech-agnostic how, and finally the tech-informed how. Skipping to the last step, or doing a sloppy version of the second, is fine for a POC, but has no place in real-world production.
It still baffles me that systems are built this way. It’s like jumping straight into code without knowing what you're coding.
Maybe we should call this whole modeling style vibe modeling?
(+ 2 small typos submitted via the form.)
Thanks, fixed the typos.
Yeah, sadly vide modeling pre-dates AI by decades.
Levels of data modeling help organize data into conceptual domains or functional areas, which hopefully leads to business domain ownership. As well has understanding the dependency of entities across domains.
I woulke add that a data modeler not only should ask questions - like you quoted Lartry on, but also truly listen to the answers and stories given by the business side of the organization. If they tell you they need "three hospitals" in their data town plan the response of the data modelers should be why? tell me more? - instead of just shuffling it all upon 1 pile (abstract/generalize)
In other words - data modelers should be curious! , not providing answers when in the stage of modeling
definitely
Is the data modeling example/exercise coming as a separate post? For me this post ends with “now let’s look at an example”.
yes, it will hopefully be up tomorrow
Now over the weekend, cuz some outside things came up
No rush! Just wanted to check if the post was accidentally cut off.