Navigation:
Documentation
Archive



Page Tree:

Child pages
  • 2011-1-12 Conference Call---Deep Dive Content Interoperability

This wiki space contains archival documentation of Project Bamboo, April 2008 - March 2013.

Skip to end of metadata
Go to start of metadata

Notes:  January 12, 2011.   4:00 pm Central.  "Deep Dive Call Regarding Content Management and Interoperability"

Conference Call Participants:
Steve Masover-UC Berkeley
Jonathan Smith-NU
Xin Xiang-NU
Bill Parod-NU
Karen Miller-NU
Martin Mueller-NU
Bob Taylor-NU
Bruce Barton-UW Madison
Jim Muehlenberg-UW Madison
Scott Prater-UW Madison
Jon Dunne-IU
Travis Brown-Maryland
Doug Moncur-ANU
Charles Blair-Chicago

A. Content Format and Protocol

We discussed the difference between content formats and protocols that we will consume from content sources and repositories, and the format we will use to provide the content to workspace applications.

We recommend CMIS (Content Management Interoperability Services) for supplying content to workspace applications. 

CMIS is
- Available on the target systems for Workspace development.
- A very general container format that can be used for any type of content that we would work with.
- Web-service based so it can easily be consumed by a large variety of systems using a range of technologies.
- Undergoing very wide adoption among enterprise content management vendors.
- Does not provide guidance on vocabularies for property names and values, data formats, or format translation.
- Protocol binding for AtomPub and SOAP.
- Relatively language agnostic.

We discussed the role of ATOM standards for interoperability work. ATOM is commonly used for content syndication and is part of the definition of CMIS RESTful access. (CMIS provides SOAP-based standards as well.)

We do want to support access by existing client software where possible, but are not yet clear how far the CMIS services go in supporting existing ATOM clients.

B. Access Models

Jonathan Smith reviewed the two models for managing content transfer and translation from source repositories that NU had used in the recently completed JCR-Connect project. Where possible (Fedora, using web services) we do the translation on the fly and provide immediate access via the JCR model and Apache Jackrabbit web services. 

In the case of a search engine architecture (California Digital Libraries) Northwestern JCR-Connect developres used a local JCR repository as a cache for search results. Northwestern predicts that a similar arrangement using CMIS for Bamboo interoperability services.

Jonathan briefly mentioned the use of Domain Specific Languages (pattern matching, etc, see http://en.wikipedia.org/wiki/Domain-specific_language) to simplify the translation from source repository formats to the node-property model used by both JCR and CMIS.  Smith's  hope is that such techniques can allow us to re-use as much code as we can across connector implementations.

C. Content Indexing and Identifiers

We briefly discussed the issue of content indexing and identifiers.

NU pointed out that (especially when dealing with text, but also images and time based media) it will be common for client software to need to query a segment or portion of a content document.

D. Implementation Technology and Deployment Issues

The participants  discussed development and hosting of interoperability services. The general sentiment of participants on the call was that, where possible, these interoperability services will be developed in Java as OSGi bundles and they will be hosted on the Bamboo infrastructure servers.

The services would access source content repositories remotely and make the content available to workspace systems and clients via CMIS web services.

The comment was made that under some circumstances it may make sense to locate some conversion and content processing closer to the source repositories. This may be  important when  there is complex or heavyweight processing that must be done before the content can be sent on to content consumers.

E. Next Steps for the Content Interoperability Focus Area of CI

As next steps we will need to:

- Identify target repositories and content sources to be used in the initial round of development.  
   (Note:   three collections have now been identified for early work by the Corpora Space team.   Bob Taylor 1/18/2011)  

- Continue to engage with the object design process underway in the Workspaces groups.

- Finalize our recommendations for CMIS developments at the Berkeley F2F meetings. (At the Wednesday morning, January 26, 9:00 am pre-meeting?)   Share recommendations with the larger group of participants at the F2F meeting.

- Identify major milestones/priorities for development work in this focus area for the February 2011-September 2011 period. (Accomplish before or during Berkeley F2F meetings.)

- Identify work assignments among NU, UC-Berkeley and WI-Madison teams in this focus area for the February 2011-September 2011 period. Other work partners are welcome!  (Accomplsih during Berkeley F2F meetings or immediately thereafter.)

- Get to development work!

  • No labels