About a year ago; Peter Murray-Rust showed his research group a web interface that allowed you to type SPARQL into a textarea input and have it evaluated. I had a flashback to people being shown the same thing with SQL years ago. So if SPARQL follows the same pattern, the textareas will disappear so the developers take the complexity of the query language and data model away from the users, then the developers will write enormous libraries (c.f. Object Relational Mapping tools) so they don’t have to deal with the query language either.

Ben O’Sheen recently posted on Linking resources [in Fedora] using RDF, and one part particularly jumped out at me: –

The garden variety query is of the following form:

“Give me the nodes that have some property linking it to a particular node” – i.e. return all the objects in a given collection, find me all the objects that are part of this other object, etc.

I think the common-or-garden query is “I’m interested in uri:foo, show me what you’ve got”, which is the same, but doesn’t require you to know the data model before you make the query. Wouldn’t it be cool to have a tech that gave you the “interesting” sub-graph for any uri? Maybe the developer would have to describe “interestingness” in a class based way, or it could be as specific as templates (I suspect Fresnel could be useful here, but I looked twice and still didn’t really didn’t get it). Whatever solution looks like, I doubt that a query language as general and flexible as SPARQL will be the best basis for it, for the reasons Andy Newman gives – what’s needed is a query language where the result is another graph.