Is EAV (Entity Attribute Value) equal to SPO (Subject Predicate Object)?

The Entity-Attribute-Value model is a quite mature knowledge representation structure in the database community. Especially, Kingsley Idehen is utilizing it quite often to propagate a Semantic-Web-independent Linked Data meme. On the other side, there is the Subject-Predicate-Object form (knowledge representation structure) which we know at its best from simple natural language sentences and which is applied in the design of RDF Model and viewed there as a specialisation of Semantic Networks.

However, I'm wondering about whether EAV is really a good "generalisation" of SPO. As far I understand the EAV model, due its extension to EAV/CR (EAV with Classes and Relationships) it is possible to describe complex substructures, hence it is the basis for a hierarchical structure. The SPO model isn't restricted to that issue, it is the basis for a graph structure. So, I wouldn't insist an equivalence relation of EAV and SPO. I might endorse a subset relation where EAV is a subset of SPO.

What do you think about their relation?

Yes, it is re. my world view.

Subject-Predicate-Object and Entity-Attribute-Value records == 3-tuples. In either case, you can use URI Reference values in slots 1&2, while doing so optionally in slot 3.

When the dust settles, many will find a 3-col spreadsheet saved as CSV or TXT == very powerful data annotation tool and workflow combo re. Linked Data curation.

What is superficially different is the following:

Semantic Web community align SPO and RDF, and by effect the specifics of some granularity associated with RDF specification e.g., typed literals handling.

I hope this clears things up!

RDF is useful re. Linked Data, but that usefulness doesn't extend to the front door i.e. introduction to the basic concept of crafting data representation via hypermedia resources where each Data Object has a URI Reference based Name :-)

When people embrace Linked Data (without RDF distraction) they'll eventually come to understand and appreciate the more granular aspects of RDF starting with the handling of typed literals and its applicability to Locale sensitive data integration etc..


I am saying: if you are speaking to an audience of people that already have experience with EAV in some capacity i.e. most of the pre Web DBMS world (developers, integrators, power-users, and even end-users, start the conversation from EAV not SPO.

There are more DBMS users than there are folks that care about RDF, SPO etc.. Thus, when we encounter this pervasive profile, start the narrative from a point that isn't provincial, confusing, or ultimately repellent.

Not complex, once we understand that RDF, SPO etc.. are just part of a technology continuum.

That's the essence of my point. It's about the audience and not operating in mono-tone :-)