Time/Place
- Time: 12:00pm Eastern Daylight Time US (UTC-4)
- Place: Google-hangout, https://plus.google.com/hangouts/_/event/c1glu6soq43r1rr6ou17qtobug8
Attendees
- Andrew Woods
- Esme Cowles *
- Chris Beer *
- Daniel Davis *
Declan Fleming- A. Soroka
- Benjamin Armintor
Zhiwu XieNeil Jefferies
Note-taker =
Previous note-taker = *
Agenda
- Thought exercise: "What would be the technical "risks" of releasing 4.0 Production *now*"?
- Or another way, "Where do we want to put next sprint's dev energy"?
Review of previous actions
REST API spec plan: Agreement on approach? Next steps?
- Performance scenarios: Focus on use case numbers from Oxford and Stanford?
- Strawman for Grinder scenarios: Performance Testing Scenarios
- Third highest team priority from survey: "Preservation-worthiness"
Previous Actions
From 2014-08-20
- Neil to define initial set of system CI tests
From 2014-08-27
Ben and Adam: Begin drafting REST API spec and what a TCK should cover. Incremental deliverable for next week: what LDP is, how F4 extends it, and bytestreams
Dan: Review existing performance work and put together some scenarios
Esme: Report on status of filesystem serialization once UCSD beta pilot has progressed - Beta Pilot page has been updated with deliverables
Discussion
Adam begins with API:
Base specification LDP plus base ontology but bytestreams
Possibly add transactions (suitable for web scale, not XA, JCR spec)
Thoughts about container:
Objects plus datastreams (fedora)
External to repo
Esme: other objects in repo
Andrew: question do we want to support
Adam: has different guarantees if outside
Adam: Need to think out ramifications for different kinds of outside items
Dan: Could have some overlap with issues about tiered storage
Adam: Not absolutely required for LDP
Andrew: Member, Reference, Indirect Container (need definitions)
Dan: Considerations of asynchronous responses (overlaps burden for remote operations)
Andrew: Need to define proposal for tiers
Adam: Keep core simplest possible
Andrew: We need to review what we have, LDP, is the interaction model logical, clean, does it represent what we expect out of the repository
Adam: Did we make a good guess for the implementation? Test on real world usage.
Adam, Andrew: Specification with dependencies for transactions, versioning, locking, etc.
Andrew to Adam: Can you flesh out all the tiers?
Adam: Will try for week from today
Esme: Should conversation be on Fedora tech list
All: yes
Performance Scenarios (Dan)
- Fixity checking
- Real-world scenarios
- Cache and no-Cache
- Look at direct REST client for Python
- Setting up environment (Scott Prater may have done some work)
- Many folks could contribute tests
RDF and Sparql Update well underway for next period (Esme)
Adam and Ben to detail proposed partitioning of API
All to review "Preservation Worthiness" description for next week's discussion
- Dan to create matrix of performance test payloads (probably here)