From RDF Stream Processing Community Group
Participants
- Jean-Paul Calbimonte (Chair)
- Alejandro Llaves
- Minh
- Javier Fernandez
- Tara Athan
Agenda
Resources
Minutes
- javier: presents an example of json-ld serialization
- tara: this is a small stream example?
- javier: yes
- tara: another timestamped graph that is in between the recorded at and generated at that arrives in?
- javier: nothing prevents this from happening, nothing enforces sequence in Json-LD
- javier showing examples with internal context, shared context, and no shared context
- javier: showing bnode graph examples
- tara: json-ld may be controversial, but multiple serializations are possible.
- tara: suggestion: modify the dates. example with interleaved graphs, repeated timestamp. may serve as a good exercise.
- jp: does it match the abstract model?
- tara: actually the examples represent 4 timestamped graphs
- jp: to say this in the doc, and include examples after definition.
- tara: explicit example of transmission possibilities of Json-LD
- tara: substream vs window
- minh: could remove the finite form the definition.
- tara: That would work. Suggest to make difference between syntactic terms and relationships
- tara: It is important to distinguish terminology for classes versus property. Stream is a class. Substream is a property. Bounded stream would be another class. Window is also property.
- tara: ordering problem. default ordering function?
- minh: make it more generic.
- minh: associate the predicate to the predicate?
- minh. agree on properties that the ordering functions must comply to
- Tara Athan: For greater generality, it is possible to associate a fixed ordering with a predicate.
- Tara Athan: Some general properties that must be satisfied by any order:
- Tara Athan: 1. if one timestamp "precedes" (in the sense of Allen operators) another timestamp, then it must be less than the other timestamp.
- Tara Athan: 2. if a<c and b>d, then timestamp [a,b] < [c,d]
- Tara Athan: Correction: if a<c and b<d, then timestamp [a,b] < [c,d]
- Minh: something like this should violate the condition: [a,b] smaller [c,d], [c,d] smaller [e,f] but then [e,f] smaller [a,b]
- Tara Athan: Certainly the partial should satisfy the usual mathematical definition of partial order, including transitivity.
- Tara Athan: partial should => partial order should