This meeting is a hybrid teleconference and IRC chat. Anyone is welcome to join...here's the info:
- Time: 9am Eastern Daylight Time US (UTC-4)
- Via Skype (Dial from anywhere):
- In Skype, select Add a New Contact
- In the search box, type
freeconferencecallhd.8053991200, hit Enter, and add the contact
- Call the contact
- When prompted, bring up the Skype keypad and enter the following by clicking with your mouse – not using your keyboard:
- Via Phone (US Number):
+1(805)399-1200, Access Code:
- Via Skype (Dial from anywhere):
- Join the #duraspace chat room via Freenode Web IRC (enter a unique nick)
- Or point your IRC client to #duraspace on irc.freenode.net
- Notetaker volunteer
- Action items from last week?
- Review recently submitted issues (15 minutes)
- More 4.0 provocation from Chris: What if Fedora objects were named graphs with an optional set of immutable datastreams?
- 3.6 status
- Chris, Ed, Eddie, Adam, Aaron, Dan, Ben, Jonathan
See raw notes on IRC.
Chris: Fedora Object model has not been discussed much as part of 4.0 planning.
Chris: Set aside versioning and other warts in the current design.
Chris: Has come up especially early with Topaz (Ambra)
Chris: Consider about modeling as open Rdf (named graph) instead of a current form.
Chris: Graph eventually related eventuall to immutable bytestreams
Adam: Had a similar discussion with Steve
Adam: Runs into object model.
Adam: Has constructed an OWL ontology.
Chris: Yes there are many facets, Started with storage.
Chris: Very open data model but started with lowest level package.
Adam: RDF pointing to package.
<ajs6f> RDF taking on the functionality of package
Dan: We should understand that we have a balancing act to do to make useful piece of infrastructure
Dan: At Fedora's core it is a bridge between fine-grained and course-grained information systems - but itself is based on the lowest end include course-grained immutable bytestreams.
Chris: Much more open ended model - possibly in OWL but graph oriented regardless.
Adam: Earlier - Facing object model
Adam: Dissemination provides a facade (always dynamic) where you can get the graph or parts.
Ben: Do services go away.
Adam: Not necessarily but need to be rethought.
<barmintor> Well, it depends on what folks we're talking about
Chris: Not a quad store, no big quad store in the sky.
Adam: yes, even linked data.
Adam: Likely internal graph for durability
Adam: Web lacks this notion.
<ajs6f> Web lacks that value, even.
<barmintor> I have a devil's advocate comment re: IDs, and a suggestion about process
Chris: There is a PREMIS owl model.
Dan: Tension between flexibility, openness and interoperability
Eddie: Why have poople not used Fedora maybe we have overconcentrated on openness.
Chris: What has Fedora not been taken up or dropped. What features: Scaling.
Eddie: Get the "whys'" out in the open.
Eddie: We have talked about objects as named graphs.
Adam: We want to do richer Rdf and need to use blank nodes.
<barmintor> When do blue-sky conversations on the committer call become publicly available design propositions that the community can respond to?
<ajs6f> And how?
<barmintor> I thnk this is a necessary step in the resolution of some sticky issues
Dan: What happen when intellectual object are overlayed on Fedora objects.
Eddie: We need to find out what are the real issues.
Eddie: What is Fedora's special sauce?
Chris: Some of the questions about Quad and named graph is intended to serve other uses.
Dan: Reacting to Bens statement: How do we redesign and not just have words in the air.
Ben: How do we engage constituents. Start with forming the use case and publish it. Be willing to reconsider if feedback indicates a new direction.
<barmintor> I DEMAND IT!
Adam: We have and need experimentation.
Ben: Checksum bug.
Ben: Worked out policy problem, tests - does not work with FESL.
Ben: Trying to Springify FESL. So non-web layer policies work.
Chris: Significant change, should be considered, notes that enforcement need not be a both layers.
Chris: Currently a lot of copy and paste code.
Chris: Should the policy enforcement be appropriate to the layer.
Adam: Should not be in the web layer anyway, should be lower.
Ben: If loaded in a Web application is is loading into the Web app context.
<barmintor> Thanks all!