AA4: Wiki-based collaborative work practices

Foster Emerging Wiki-based Collaborative Work Practices in Research Teams.

Executive Summary

This report presents the findings of research undertaken as a part of the DART (Dataset Acquisition, Accessibility and Annotation e-Research Technologies) project, specifically for Work Package AA4, foster emerging wiki-based collaborative work practices in research teams.  The milestones for the work package were:

• Working software with rights management flowing from chosen directory services;
• Wiki content and metadata used to generate semantic web maps of content; and
• Series of discipline specific templates and wiki structures.

These milestones were achieved by utilising an appropriate open source set of tools that would enable researchers to easily access and annotate data within a wiki-like environment. The tool that was chosen as a framework wasPlone.  Plone is a Content Management System based onZope.  Due to Plone’s use of python as a language it is very flexible and easily extendible. 

An agile software development model was followed in the implementation of this work package.  As the work package team developed the skills in Plone and the requirements of the DART project and other potential users became evident, several portals were established using Plone to test the effectiveness of the tools.  Despite the challenges presented during the work package, this work package is has produced a number of operational Plone portals and these are located at: http://plone.jcu.edu.au

Several additional milestones were achieved as part of this work package including:

• Integration with Storage Research Broker (SRB);

As datasets used by researchers can be quite large, a document storage system called SRB was integrated with Plone to provide the capability of handling large datasets.

• Improved Security;

In order to encourage collaboration between members of different institutions it was necessary to implement inter-site authentication.

• Workflow;

Development of workflow within Plone to manage documents.

• Task lists;

Creation of To Do type lists for managing activities.

• WebBlogs.

Ability of researchers to create Blogs of their activity.

This work package has demonstrated that through agile project management of the project at JCU, the work package was able to deliver several significant outcomes in addition to the milestones in the work package.

Project Outcomes

The previous section detailed the process that was undertaken to determine the appropriate tools for collaborative work practices.  This section details the components that were added and integrated with the Plone platform.  It must be noted that throughout the DART project, team members used the Plone service to store information.   The primary milestones were achieved as follows:

R ights Management Flowing from Chosen Directory Service

Rights management has two major requirements:

• To ensure that the correct member has been identified - authentication, and
• That the correctly identified member has the required privileges to access the resources – authorisation.

The rights management issues were address in the context of the Plone implementation, as it was chosen as the product that provided the ‘best fit’ of requirements to resolve the milestones of the AA4 work package.

Initially the TeamSpace and MXMWorkgroups products were identified as possible solutions.  Both allowed site members to manage groups much like content. However the requirement of both products weren't compatible with LDAP and added more complication to the authentication system migration which occurred between Plone 2.1.3 and 2.5.

Authentication Improvement

Plone already had some LDAP support through the LDAPUserFolder plugin but there were some key issues:

• All LDAP groups are printed when mapping LDAP groups to Plone. This can take a very long time and makes choosing the correct group difficult.
• Users cannot be individually mapped to Plone only groups can.
• LDAP users don't appear in Plone when searches are performed.

The LDAP plugin is essential to ensure security as it allowed the users within James Cook University, the host institution to access Plone.  Other users must be created in the LDAP directory or added to Plone own authentication mechanism.  

There were changes made to the Plone source, LDAPUserFolder, and until the changes made to the LDAP plugin are accepted back into the main repository we will maintain patch sets for application to the latest versions of LDAPUserFolder and components.   The changes are especially relevant in assisting in deployment of Plone at institutions which have large LDAP servers and improves integration between LDAP and Plone.

Authorisation

Plone natively allows the owner of content to set permissions to the content at a user and group level, as long as the group is defined in LDAP.  Content that is created can have a predefined set of permissions and generally these permissions are set to owner access only.    When the authentication processes become more complex and include users and groups outside Plone’s LDAP directory then authorisation becomes more complex.  However, authorisation depends on policy and the users defining the security of the content are responsible for granting access to groups and users outside the standard LDAP directory. 

Wiki Content and Metadata used to Generate Semantic Web Maps of Content

Whilst the concept of generating semantic web maps appears to be meritorious, it is complex and doesn’t necessarily reflect the current capability of this type of technology or the complexity required to create the metadata to generate the maps.  Two plugins were researched to identify if a solution was possible.

PloneOntology

PloneOntology replaces keywords with an ontology based solution.  However, the team could not get version 0.9b3 to work and, unsurprisingly, it seemed overly complicated for new users.

PortalTaxonomy

PortalTaxonomy provided a tree like implementation of keywords. In reality it just provided a better keyword selection widget and again only server to unnecessarily increase the complexity of the site.  Ideas from the ontology would be used to allow clustering of search result around similar keywords.

These solutions created work for the users and added complexity to the site without providing significant benefits.  It was decided to utilise a search tool.

Plone Live Search

Plone Live search can be provided with a list of key terms that can be added by the users and will also allow searches within documents.  It may not provide the semantic capabilities of a true semantic search tool but it satisfies 80% of the requirements of the milestone without significant overheads.  Plone Live Search was implemented as the search tool.

Series of Discipline Specific Templates and Wiki Structures

Custom templates were explored, but the overhead of upgrading new products would grow quite substantially with each new template. The overhead grows because of API changes this is particularly relevant since Plone is currently in the process of changing from Zope2 to Zope3.

 A better method would be to generate content types with GenericPloneContent product using the Work Package DA3 metadata schema repository, and to use the much simpler PloneTemplates product. This would push some of the template creation to the users and the rest would generated and easier to update.

An important aspect of achieving this milestone was to remove the barriers to collaboration that was caused by the difficulty in sharing documents with co-workers. To resolve this issue, more advanced work flows were introduced. Work flows within Plone control the default permissions on content and can trigger events. These work flows contained extra states which allowed other researchers on the site to view selected documents in a read-only fashion, as well as provide comments and other input.