Clojurians Log v2
rdf
reading https://patterns.dataincubator.org/book/data-management-patterns.html
the ideas of how to use named graphs seem almost obvious but I'm glad someone wrote about it since I've seen comments of the tenor that named graphs is something one would almost never need on real life RDF
it seems at least someone is using them 🙂
As far as I know almost everyone doing anything substantial with RDF will use named graphs; it’d be hard not to.
> I use these rarely enough that I re-read the “Named Graphs” section of my book’s “Updating Data with SPARQL” chapter as a review before I assembled the steps below.
I guess I'd been "consuming" quite a bit of Bob's content lately hahah
Yeah I should clarify; if you’re publishing, changing or creating RDF; graphs are important. Much less so for querying as the semantics should ideally just be in the data (i.e. the ?s ?p ?o part of the quads).
We certainly use named graphs extensively
You should however be careful with them; named graphs are a bit of an afterthought (introduced in) SPARQL 1.1. It’s a real shame for example that CONSTRUCT statements can’t construct graphs - there’s an issue discussing adding this to SPARQL 1.2 though.
In particular you shouldn’t give graphs any real semantic meaning. Typically you usually only want to use them for data management; or establishing known/trusted subset(s) of information for processing.
I'm trying to create a lil blog generator
I was thinking of giving every "document" on my blog its own named graph
say, the backing data of "/blog/2020-05-01-title" would live on the named graph "/blog/2020-05-01-title/data"
(not sure about the urls yet)
i think that would correspond to this pattern: https://patterns.dataincubator.org/book/graph-per-resource.html
Yes that makes sense and is pretty common; as it enables you to update the resource through a simple drop and replace of the graph
Am I correct in the recollection that data is only expected to be consistent within a given named graph?