Nice article!

“even though GraphQL isn’t actually a graph query language!”

Isn’t the “graph” in GraphQL about the structure of the query types, i.e. starting at a root node and branching out from there? I seem to recall Lee Byron explaining this and that using the graph form for the query types helps in three different areas.

1. being able to quickly search the graph (tree) for the right query resolvers on the server.
2. modeling the data to fit the graph (tree) form of the DOM, i.e. for component based UI ala React.
3. and to follow a graph structure for the data itself as best as possible.

Also, would I be correct in saying that the connection system was created, because normal pagination can’t properly slice up a graph data structure, because such a structure can be multi-dimensional, whereas classical pagination is 2 dimensional?

And one last thing. If edges do have properties, like you mentioned, the date property of a “like” edge, this could determine the resulting order of the edges, correct? How can a change in ordering be made, say from ascending to descending?

Scott