Attendees
- A. Soroka
- Unknown User (escowles@ucsd.edu)
- Eric James
- Andrew Woods
- Benjamin Armintor
- Stefano Cossu
- Peter Gorman
- Nigel Banks
- Nick Ruest
- Mike Stroming
- David Wilcox
- Ed Fugikawa
- Michael Durbin
- Martin Dow
General
Time: 4 PM Eastern
Indicates who took minutes -
- ReadyTalk call-in details:
- U.S.A/Canada toll free: 866-740-1260, participant code: 2257295
- International toll free: http://www.readytalk.com/intl
- Use the above link and input 2257295 and the country you are calling from to get your country's toll-free dial-in number
- Once on the call, enter participant code 2257295
...
- Description of context and problem space
- Discussion of use cases
- inherited from Fedora 3
- new for Fedora 4
- Discussion of proposal
- Does it meet the requirements?
- Is there an impact on the use of CND?
- ...
- Next steps?
Minutes
Context setting
Fedora has history of content modeling (CM).
Currently leveraging inherited type system offered by ModeShape.
- It has some limitations in the F4 context
Art Inst. of Chicago then brought specific needs to the table - Artic began pushing on Compact Node Definition (CND) capabilities and short-comings
This is has raised the question of whether CND meets the needs - What are the gaps?
- What can we do to fill the gaps
Use cases
What use cases have not yet been recorded?
Q: Is F4 CM at the level of F3 ECM
a: In many ways F4 is more expressive
- Note, much of the underlying JCR type system is not being exposed
Q: What is most important to you about CM?
- Eric: Proposal, represent CM in RDF, not via CND
- Ben: Concern that proposal may be too broad in reach
- Ideally, the additional functionality be optional
- Current CND is more than adequate
- Stefano: CND covers 90% of needs
- Drawn to starting point of restrictive types, and opening with mixins
- Is it possible to layer a fedora abstraction over CND?
- Priority is placed on restriction, current F4 is permissive
- Peter: Invested in mixin content models
- Less concerned about restriction
- Interest in being able to tag objects with content models or properties
- Nigel: F3 use of CM is nominal
- Drawn to the idea of new characteristics based on mixins
- Notionally interested in validation of object structure
- Nick: same as Nigel
- Mike Stroming: needs internal CM discussions
- Martin: Interested in repository user roles
- Interested in runtime flexibility, and optional validation
- W3C languages offer powerful expression
- Ed: Resonates with Nigel/Nick/Ben/Martin
- Mike Durbin: Likes ability to enforce relationship rules
- Validation can happen in the repo or in the upper-level application
Summary: People are interested in an easy-to-use, nominal type system.
Restriction and validation is less of a priority to most.
Nigel would like to ensure that if a content model includes a specific datastream, that that datastream exists with the correct mimetype.
Stefano: suggestion to associate access policies to CMs
Martin: suggestion for process moving forward: timeboxing requirement gathering, roles, proposal, etc
CM in F4 meets most of the community need
There is another conversation to be had regarding extension points for validation
Actions