Martin Davis posed an interesting question recently on his blog, “Is JSON the CSV of the 21st Century?” Generally I would have thought of JSON as a more light-weight and easier to use XML replacement, but his post got me thinking about how we often use JSON (or its spatially-aware cousin) in ways that closely mirror how we would historically use CSV, an easy way to pass information in a human (somewhat) and machine readable way.
As Martin says in his post, JSON does go beyond what CSV offers, by adding additional datatypes, but it also allows for much more powerful ways to model data than a flat CSV file. JSON is a great successor to CSV (even if CSV is probably more popular) especially since it can be mapped directly into native arrays, dictionaries or other types in most programming languages (FME can both read and write JSON and GeoJSON as well!).
If Martin’s guess is correct, that in 40 years, we will no longer heavily use JSON (I would have to agree with him) what will we be using in 40 years? Perhaps it will be an extension of what brought us JSON. Initially JSON was used to load data into javascript programs over a network (it stands for Javascript Object Notation), perhaps the future will bring a cross between JSON (text-based) and SQLite (full relational database with basic SQL in a file)?
Whatever the future brings to simple data formats, I think Martin’s question on JSON’s place today can be answered with a yes. If you are working with CSV, why not take a look at what JSON can do to allow more complex data to be stored (especially take a look at GeoJSON which is far more powerful than using CSV with x and y columns!).
Michael Weisman
