A lot has been talked about microservices type arquitecture, and the API led arquitecture introduced by mulesoft. After reading and studying a lot about it, and looking at the challenges we're facing in Domo, I have the feeling we could and should apply the same arquitecture within domo to organize Datasets.
The api led arquitecture for application integrations talks about creating different layers of abstraction between the different services, which allows for a lot more reuse, more flexibility and an agile methodology (pictures attached).
We had a problem last week integrating time and attendance information and sales information and we decided to create a simple diagram of different datasets one feeding the next, creating different abstraction layers, and creating a lot more modularity and reuse. After implementing it, we started thinking that this approach would benefit all of us going forward, and help organize the sheer number of datasets we have which are getting more difficult to manage.
Whats your view on this ?

