Why isn't RDF more popular within the private sector?

Question

I've learned about the idea of storing data in the form of triples not too long ago. This is essentially what RDF allows and since I've started using this to model my data, I've never really looked back to less interoperable data models. I know RDF has a long history of misconceptions, but so does JSON or HTTP or many other standards out there.

I'm wondering why isn't RDF more popular within the startup/industry scene, because it's definitely popular and has shown its power in academia and life sciences (ex: the Linked Data cloud).

Answer 1

While there is clear power with RDF and other formal ontologies, web technologies are showing a tendency towards simplicity -- things that are easy to code, read, manipulate, etc. RDF has none of those qualities. So while a language like Ruby might evolve on its own, it gains more power, popularity and community when a platform that makes development more simple (like Rails) starts to use it.

Answer 2

RDF is unpopular because it is generally misunderstood. This problem arises (primarily) from how RDF has been presented to the market in general. To understand RDF you have to first understand what Data actually is [1], once you cross that hurdle two things [2][3] will become obvious:

1. RDF is extremely useful in regards to all issues relating to Data

2. RDF has been poorly promoted.

Links:

1. http://slidesha.re/1epEyZ1 -- Understanding Data

2. http://bit.ly/1fluti1 -- What is RDF, Really?

Answer 3

I see three drivers for this true observation:

1.) I'm a very observing information technologist I would say and RDF has been on my radar for I would say 6 years. My sentiment towards RDF until about one year was:

• Very popular in academia
• Complex if you want to do advanced stuff
• Clearly the basis for the Web 3.0 but needs time

2.) Considering the the major architectural paradigm in which at least larger commercial organizations lived and still live: SOA, ESBs. They were just happy to have their interlinked silos and not isolated silos. Their biggest challenge was and still is to integrate systems. Many of these systems is COTS and they have no control and nor interest in the exact details of the employed data models, they just want to import and export data from those systems.

3.) Also consider what major vendors in the IT sector brought to the table: I know of RDF support in Oracle and DB2. I even played around with it on DB2 a bit and my opinion about their efforts is rather crushing: Not really usable. I thinkk that they merely added support for some SemWeb stuff as marketing alibi - just to have it and not being reckoned to have overslept a trend.

I think all these 3 lead to the situation you described. I'm 100% sure this will change and soon be visible in Gartner's Hyper Cycle as big, open and linked data trends mandate a interoperable data model/ serialization format.

Answer 4

I just wanted to add one more point to this old thread, although there are very good answers already.

Graph databases can be very appealing for anyone dealing with unstructured data, and there are "enterprise" products with a lot of features that make the job easier, such as Neo4J, etc. These databases offer the flexibility of graph languages that work well within a closed group (e.g. a company and its subsidiaries).

RDF is a specialized graph language built with sharing in mind. This imposes further restrictions and complexity, such as the use of URIs, the need for choosing an appropriate ontology, aligning with community standards... Such added complexity is welcome by cultural heritage and scientific communities whose core mission is to make their data broadly available. For-profit organizations may not at all be interested in sharing data, so RDF is unnecessarily complicated for their use.