Closed csarven closed 5 years ago
@csarven thanks for your feedback!
The main problem with LDP is that it does not have an abstract model. The only RDF-related specification that does is SPARQL, which has an algebra. That is more important than freedom of implementation, in our opinion, and that is why we have developed an LDT algebra similar to SPARQL (yet to be made public).
Another problem is that there is no standard way to capture application-specific LDP behaviour. From our point of view, it is not desirable for all Linked Data applications to use the same processing model. Rather, the processing should depend on an app-specific component (ontology), which should be available for interchange using standard means.
Homo-iconicity enables meta-programming, which was one of our goals.
PUT
is an Update interaction, which means it generates a SPARQL update consisting of 2 parts: DELETE
and INSERT DATA
. The DELETE
part is taken from ldt:update
value and (?this, <http://linkeddatahub.com/people/Berners-Lee>)
variable binding is applied. INSERT DATA
is generated directly from PUT
request body. The combined update request is then executed on the SPARQL service. Does this help?
@csarven @timbl @sandhawke would love your opinion on the LDT spec draft :) https://atomgraph.github.io/Linked-Data-Templates/