Over the years the demands of multi/omnichannel publishing has led to lots of discussion about “future friendly” structured content. There’s a real tension inherent in that goal, however:
It’s incredibly difficult to predict the kinds of structures that will be necessary or useful in future media, channels, and even different treatments in the same web/print forms.
You avoid some pitfalls by building against underlying, inherent meaning and intent in the content and content types. (“Conceptually, how does a ‘tutorial’ differ from an ‘update’”? etc)
But some structural data decisions come from the ugly, in-the-weeds needs of a given medium or distribution channel, or a particular way of organizing and navigating that can change even if the underlying content stays consistent. That kind of modeling work tends to age poorly.
We’ve worked really hard to call out the parts of a content model that we believe are fundamental to meaning, and the ones that are necessary-but-ephemeral. Also, designing points for expansion into content models where we know needs are likely to grow or evolve in the future.
This is one place where component content strategies really thrive; although there are absolutely editorial tradeoffs to be navigated, adding new component types and assuming that some/all user-facing is actually a composed aggregate of smaller pieces? That goes a long way.
At the most basic level, just having a simple embedding system — the ability to place pieces of structured content like a gallery, a map, an “example,” etc. inside the flow of another piece of structured content — can improve the long-term adaptability of a model dramatically.