USE CASE
...
MILESTONES:
- (M1) Development/tweaking of the demo workflows and execution in (a) Kepler and (b) Taverna. -- done but they only work in mockup mode (maybe Lei/Daniel/Saumen to do sth about that)
- (D1) Executable demo-workflows in (a) Kepler (2x3) and (b) Taverna (3)
- TO-DO: LINK THEM from the "Deliverables & Products" pages!
- (M2) Export of provenance information of (M1) from Kepler, Taverna
- (D2) Trace files (in XML? RDF? LOD?) from (a) Kepler and (b) Taverna runs.
- (M3) Shared Conceptual Model designed.
- (D3) Shared Provenance Store (SPS): Conceptual Schema
- (M4) Provenance Store (SPS) implemented and accessible (could reuse an existing e.g. OPM or Linked Data Store)
- (D4) Provenance Repository: Relational (or other implementation) Schema
- (M5) commit-trace-to-PS implemented for (a) Kepler and (b) Taverna.
- (D5) Mapping description from (a) Kepler and (b) Taverna MoPs to the PR schema.
- (M6) Execute simple provenance queries against the PS; visualize results.
- (D6) Description/explanation of the demo-provenance queries (e.g. in QLP? or SPARQL?)
DELIVERABLES:
- SPS architecture & implementation schema
- SPS implemented
- Schema Mappings (via Datalog)
- stitching via OPM artifacts
- references to data need to be ...
- "owl:same-as"
- Data Identification (common naming scheme)
- OPM "profiles" (or just annotations)
Walkthru / Life-cycle
- 0 (wf design / development)
- 1 bind data and parameters
- 2 run wf
- 3 publish wf results (i.e., output data)
- 4 (publish wf specification) --- may be optional / but may be inferrable
- 5 publish input data and params ---
- 6 publish trace
- 7 query trace
-
- These should include at least
- "find all sources/sinks in the provenance graphs, reachable from given node"
- selected "deep provenance" queries, i.e., spanning multiple runs, multiple users etc.
- (see also queries in Ilkay's IPAW paper; especially new Section 3)
- (D7) Publish paper(s) on findings
SPREADSHEET:
Vocabulary Table
Glossary
Conceptual Model
Storage Schema(s)
Implementation A
Implementation B
eScience Paper ideas:
- include distinction: references &D vs actual data *D
- include in traces a second naming scheme for data identifiers