I read with no small amusement that JSON-LD has begun to descend into its final form — JSON — which, as we’re all relatively aware is descending into it’s final form — the string.
Having worked with JSON-LD for nigh on a year and a half, I’m hardly surprised that we’re seeing this development; JSON-LD purported to bring the semantic web vision with all of the user-friendliness of JSON, which is odd because JSON isn’t any more or less user-friendly than any other serialisation (the cruel might say that it’s XML for those who have an unfounded opinion about XML).
It’s especially clear that JSON is easy to work with in a way JSON-LD really isn’t and it’s this the new trend is addressing; to alias the ‘@’-prefixed keywords with their non-prefixed counterparts. This is because ‘@’ makes everything a bit miserable.
To be honest, from an RDF-point of view, JSON-LD couldn’t get any more miserable because it encourages the kind of syntactic approaches that semantic web technology should be helping us move away from. But, even syntactic approaches are miserable with JSON-LD; the parsers are lacking, so we’re using generic JSON-parsers and we’re having to struggle with weird notation. And sure, it’s a URI because it’s an @id, sorry, id…wait…
I feel a bit sorry for the people behind JSON-LD because it was a good vision, it just didn’t pan out when it was made simpler and simpler and finally ended up being, well, the same as JSON, which was inevitable given the audience and usage.