Just saw this, which tells you how frequently (or not) that I read more as I should...

Good article. Thanks for it.

We've seen alternatives to relational databases in fits and starts, beginning quite some time ago with object then XML then ... databases. We started thinking relational databases would be replaced. And they haven't. We learned there are reasons for data in certain types of databases and structures. AND we experienced a lack of tools for reporting and other tasks when using other than relational databases that required more programming intensive use experiences... which are not well liked. AND... lest we forget... we've migrated to app execution mode only... yes services and microservices and batch jobs, but these operate within an application context in all cases (even a scheduler application).

What if we ran our applications as "worlds" that continued to learn in specific contexts and shared what was learned? There are classes of systems where this makes plenty of sense, and in which graphs will be the better persistence solution... possibly with ability to reason over the graph to enrich queries, to capture relationships that don't so naturally fit into a relationship/role graph. These will require more than reporting and dashboards... they'll require tools that allow us to view and manage meta things, like schemas of graphs, and then query with some query language to see textually and graphically what exists in populated graphs.

I'm not certain I would go as far as saying "win" vis-a-vis graphs, though I will acknowledge their increase in value and prominence.

As we get beyond first order querying of graphs and the things we organize with them, and we get toward proactive interaction with them in new types of applications, we will by necessity develop ways to view what is in a graph without the silliness of programmatic APIs only ...

Again, thanks for the article, and apologies for my very late read and feedback. I'll look forward to your future rants.