Take away's from Jamie Taylor's (of Google) keynote at ISWC2017 today
- what they are doing basically affirms that we've been researching the right things for 2 decades.
- what he mentioned as open questions are mostly things which we've also been researching for 2 decades.
- there still seems to be a reluctance to use the word "ontology", although that's exactly what it is.
- he basically said that the google graph is of importance all over google's services, though he could have been more explicit about this.
- there was some acknowledgement that the semantic web community's research is central to the developments, though he could have been more explicit also about this.
Some of the things he said which we (the Semantic Web community) know for years:
- you need a good schema, it helps a lot. Questions about the nature of things need to come before modeling (i.e., see foundational ontologies and ontology design patterns).
- entity resolution is hard and of central importance.
- ontologies with some (lightweight?) reasoning help to encode general knowledge (thus avoiding to replicate things all over the place, blowing up the graph)