Page 1 of 1

Extenal sources of semantic content assets

Posted: Mon Mar 07, 2022 6:49 pm
by FirstLight
If we can graph our content corpus well, we can query the graphs and discover things previously not visible or practical to do. That’s incredibly powerful, but there’s a lot more.

What about using auxiliary sources of intelligence to mine graphs and create new applications. One theory I’ve floated is using one or several auxiliary sources of intelligence with a graph of our content corpuses – sort of like how GPS works by triangulating them. In a sense, that is what ontologies and graphs do natively within their own structure, but can we do more?

I’ve been contemplating other sources of intelligence we might be able to use to extend or query graphs to achieve the desired result. Let’s examine all our potential sources of intelligence:
  • Obviously, we can create an ontology for our desired domains. For example, in my business the two I’m most interested in are 1) a tax compliance ontology and 2) an ontology for DITA.
  • Next, we can use ontology to graph the content corpus using one or both ontologies. At least I think so. Again, I’m no expert here, but that’s how I understand the relationship and use between ontologies and graphs, with an ontology being pure concepts in nature, and a specific data instance being the graph that is based on the generic ontology model. Unless I am mistaken, I kind of liken the two to the difference between a document schema and a document instance.
So great, now we have a graph of our content corpus and can manually query the graph to learn stuff. But can we do more than that? Consider:
  • Some of us have created content models. Often, these models are in Excel format. Can we ingest and use them?
  • Some of us have content micro-journey maps – specific multi-task scenarios spelled out step-by-step scenarios, some of which cross multiple applications to accomplish a specific goal with an application or across multiple applications. If these were encoded in some machine-consumable format (CSV, XML, JSON, whatever), can we use them with our content graph. If so, how would we use them?
  • Some of our CMS/delivery systems provide extended user features, such as allowing users to create custom content collections. How might we use these in addition to other in-bound personalization intelligence?
  • Some of our CMSs collect a boatload of content usage intelligence, yet another cornucopia of intelligence
  • Of course, if one is using a highly structured DOM-oriented content source, there are self-describing relationships in the content structure such as the hierarchy, sequence, and nesting of collections, sub-collections, topics, relationship tables, and links inherent in the content source itself.
Do we graph these sources or use them to query graphs to generate useful artifacts? Maybe generate a bill-of-materials and organization for a CMS to retrieve and deliver content?

If we really want to bend minds, there’s a class of software that tracks user journeys in real-time that can be anonymized or used with consent. How might that intelligence play into the equation when we think about truly dynamic, non-prescriptive content-as-a-service that is proactive and not failure-mode user assistance?

Go head, ideate! What other sources of intelligence might we discover useful, and how might we use them?