Using DSpace as a LOR

Challenges to Implementing DSpace

as a LOR

Scott Lesliefor COPPUL Distance Education Group

May 30, 2006

Outline

What a ‘classic’ LOR is ‘supposed’ to do

DSpace-specific challenges as a LOR

General challenges and pitfalls for LORs

In general, LORs appear to

Store actual resources, not just metadata

Be part of a CMS/LMS-centric paradigm

Deal with (hopefully) volatile content

Contain diverse stuff (as no one really knows what a Learning Object is anyways.)

Edutools Comparative Framework

Discovery Tools Aggregation Tools Community &

Evaluation Meta-tagging Content Management Digital Rights

Management & Fulfillment

Presentation and Consortia Issues

Integration and Interoperability

Technical Considerations

cf. http://www.edutools.info/lor/

Evaluative Framework: Details

Discovery Tools– Searching– Browsing– Syndication & Notification

Aggregation Tools– Personal Collections– Content Aggregator and Packaging Tool

Community & Evaluation– Evaluation System– Context Usage Illustrators– Wish Lists

Meta-tagging– Metadata Markup Tool– Schema Support– Indexing Workflow Support– Import and Export Tools– Unique Identifier Support

Content Management– Authoring and Publishing Workflow Support– Version Control & Archiving Functions– Authoring tools

Digital Rights Management & Fulfillment– Digital Rights Management– Payment and Fulfillment

Evaluative Framework: Details(2)

Presentation and Consortia Issues– Customized Look and Feel– Internationalization (I18n)– Multiple Collections– Media Transformation and Display

Integration and Interoperability– Federation and Harvesting– Course Management Integration– API and Web Service support

Technical Considerations– Authentication– Authorization & Personalization– Usage reporting – Operating System– Application/Database Server Requirements– Scalability– Support

Evaluative Framework: Details (3)

Challenges for DSpace as a LOR

‘Perceived’ Shortcomings of DSpace

Limited metadata schema support Lack of support for federated searching Lack of useful browseable subject classification No IMS Content Package support Limited Multi-page Website support Multi-contributor workflows Versioning and content management issues Community evaluation, community building DSpace as a service Others?

Metadata Schema Support

Issue: DSpace currently supports Dublin Core with limited qualifications, whilst most other LORs have used IEEE LOM

Possible Mitigations: Implement LOM in DSpace

– http://cwspace.mit.edu/wiki/MetadataIssueThree Crosswalks

– http://wiki.dspace.org/CrosswalkPlugins – http://www.imsproject.org/metadata/mdbestv1p1.html#Dubl

in2 Disregard, as most of LOM not used anyways

– http://www.ariadne.ac.uk/issue41/godby/

Federated Searching

Issue: Dpsace supports OAI-PMH, a harvesting model, but not a ‘federated search’ mechanism either between DSpace instances or with other repository software

Possible Mitigations: Implement SQI or one of the other federated search interfaces

– http://sourceforge.net/mailarchive/message.php?msg_id=11812130

Use SRU/SRW as mechanism to federate search– http://www.loc.gov/standards/sru/short-topics.html#fed

erated

OKI OSIDs– http://wiki.dspace.org/SakaiIntegration

Disregard, harvestability is enough/better anyways

Browseable Subject Classification

Issue: Currently DSpace only supports browsing by Title, Author and Date, and the only limited way to structure by subject is through ‘Collections’

Possible Mitigations: use ‘keywords’ field for either constrained or

unconstrained (folksonomic) approach; build interface to ‘browse’ by keywords

disregard problem in favour of ‘search’ as main method to retrieve items

IMS Content Packages

Issue: DSpace currently treats IMS Content Packages as binary ‘blobs.’ They need to be previewable on the server, and any existing metadata consumable

Possible Mitigations: work being done in CWSpace to handle Content

Packages, or promise of system extensibility in v 2.0– see http://wiki.dspace.org/PackagerPlugins – http://wiki.dspace.org/XmlNamespaces_2fCwspaceImscp?

action=highlight&value=IMS

Multi-Page Websites

Issue: Many ‘learning objects’ are actually multi-file websites; these need to be EASILY uploaded ‘en masse’ AND EASILY assessable by potential users

Possible Mitigations: Upload each file of the site individually using

multiple upload controls (ICKK!) ??

Multi-contributor workflows

Issue: Not every object has just one contributor, and not every submitter has all the required info nor will for all time

Possible Mitigations: - Disregard, hold to single submitter-to-single-

resource model and use moderation as mechanism to add more metadata

- ??

Versioning and content management

Issue: DSpace has clunky support for multiple versions of content

Possible Mitigations: currently proposed DSpace solution

(manually link multiple records together, use relation field)– http://wiki.dspace.org/VersioningSupport?

action=highlight&value=version ??

Community Feedback

Issue: DSpace has limited support for the community to express the judgments of value of any piece of content

Possible Mitigations: Add these on through third party tools or

possibly extensions to DSpace

DSpace as a service

Issue: LOR content needs to be consumed in CMS, accessed by authoring tools

Possible Mitigations: “Lightweight Network Interface (for CWSpace)” /

WebDav stuff– http://wiki.dspace.org/LightweightNetworkInterfac

e?action=highlight&value=WebDAV

OKI OSIDs / Sakai Integration– http://wiki.dspace.org/SakaiIntegration

Summary - Specific Challenges

If you deploy DSpace “out of the box” as a LOR, then you may find users have a difficult time assessing

the usefulness of certain types of content you may have some interoperability issues with

other LORs you may have some challenges offering access to

the repository contents as a service to specific LMS or authoring tools

the system may prove cumbersome to working with volatile content or typical learning content authoring workflows

Some Important Projects for Dspace as a LOR

DSpace 2.0 Proposal– http://www.dspace.org/conference/presentations/

architecture.ppt

CWSpace (Using DSpace for OpenCourseWare)– http://cwspace.mit.edu/

d+ (Distributed e-Learning Resources Discovery)– http://www.jisc.ac.uk/index.cfm?name=dplus

Using DSpace as a LOR mailing list– http://mailman.mit.edu/mailman/listinfo/dspace-lor

– (seems to have disappeared, contact Alan Wolf, University of Wisconsin; Louse Ratliff, UCLA for more info)

Some existing LORs using DSpace

http://lor.ccconline.cccs.edu/dspace/index.jsp

https://www.dlearn.arizona.edu/index.jsp – http://cwspace.mit.edu/wiki/LorDlearnUnivArizona

https://mspace.lib.umanitoba.ca/handle/1993/122

General Issues for LOR Implementers

The high cost of metadata / Self-service questionable

Intellectual Property/DRM challenges

LOs and increasingly questioned LCMS/LMS paradigm

LO/LORs ignoring innovations from social software

‘Reusability’ Paradox of LOs themselves

If LORs are the answer, what was the problem again?

At the end of the day…What were the original goals behind Learning Object Repositories /

Learning Objects?

– Share learning resources and have them reused to

– increase quality of learning experience

– decrease cost of producing learning content

– Help make resources findable and identifiable that were formerly not being found or well identified

– Make clear what people can do with the resources and give them access to resources they are allowed to use

Are there LORs which are achieving this? Which ones? What’s missing from the ones which aren’t doing this? Time? Training? A different paradigm?