Issuu on Google+

Advisory Report on Taxonomy and Metadata for Managing Learning Content A Report to the British Council

By Straits Knowledge

Patrick Lambe, Straits Knowledge; Maish Nichani, PebbleRoad v.1.1

Prepared by: Straits Knowledge, 19A Keong Saik Road, Singapore 089126 Tel/Fax +65 62210383 plambe@straitsknowledge.com

Business Reg no: 52959850K


Taxonomy and Metadata Advisory Report

Contents

1. PROJECT BACKGROUND

3

2. TAXONOMY REVIEW

4

2.1 2.2 2.3 2.4 2.5 2.6

The general taxonomy approach is sound Analysis of British Council use cases What will the facets be used for? Taxonomy usability and maintenance issues Detailed facet analysis and recommendations Summary of facet analysis and recommendations

3. METADATA RECOMMENDATIONS 3.1 3.2 3.3 3.4 3.5

18

The general metadata approach is sound Analysis of British Council use cases What will the metadata be used for? Metadata usability and maintenance issues Metadata recommendations

18 18 19 20 20

4. SYSTEM REVIEW 4.1 4.2 4.3 4.4

27

Analysis of British Council learning content systems The general content management approach is sound System recommendations Player concept

5. STRATEGY AND GOVERNANCE 5.1 5.2 5.3 5.4

27 29 30 35

37

The importance of having strategy and governance Roles and responsibilities Processes Systems

Straits Knowledge January 2010

4 4 5 5 6 15

37 38 38 39

www.straitsknowledge.com

2


Taxonomy and Metadata Advisory Report

1. PROJECT BACKGROUND The British Council develops and manages a range of elearning content objects on different platforms for different audiences and purposes. The volume of content is now at a stage where a taxonomy and metadata solution is required to help British Council staff manage this content more effectively and consistently, and to be able to pull content together quickly to meet internal needs or the requests of partner/client organizations. A faceted taxonomy has been drafted from existing vocabularies in use within the British Council network. The purpose of this report is to assess and give advice on the draft taxonomy and on the implementation of this taxonomy, on the metadata approach that is required, and on the implementation of the taxonomy and metadata within the British Council’s IT environment. These are necessarily high-level recommendations, based on a partial view of the British Council’s environment, and they are intended to orientate the British Council team rather than provide a detailed implementation roadmap. The report is in the following sections: • • • •

Taxonomy review and recommendations Metadata recommendations Systems review and recommendations Content management strategy and governance recommendations

Straits Knowledge January 2010

www.straitsknowledge.com

3


Taxonomy and Metadata Advisory Report

2. TAXONOMY REVIEW 2.1 The general taxonomy approach is sound The British Council has chosen to adopt a faceted approach. This is current best practice in taxonomy development, for any situation where potentially large collections of content need to be filtered and recombined for different purposes and different audiences. The facets work with a metadata framework to enable this filtering to take place. The review of the draft taxonomy in this section covers the twelve facets developed by Mark Hamilton from a variety of sources within the British Council, supplemented by references where appropriate to the professional TEFL literature, and in consultation with colleagues. Each facet represents a way of looking at the content that might serve a particular search or filtering need. Taxonomies are most effective when they have warrant – which means they are derived by reference to the existing vocabularies and categories in use. In practice this usually means a combination of: • content analysis (how is this or similar content already organized?) • usage analysis (how do “typical” users approach the content and how do they frame their enquiries?) • and for technical vocabularies, by reference to existing authorities (eg textbooks, coursebooks, existing internal or external standards). Generally, the vocabularies in the draft taxonomy seem to have respectable warrant in terms of representing likely content and user needs, with some reference to existing authorities (although in some facets – eg Teaching – this could be enhanced).

2.2 Analysis of British Council use cases In addition to analyzing the way that the various content sites are organized and categorized using topics for users, we wanted to understand the way that British Council staff approach the content for specific selection and filtering needs. As part of our analysis (and also to get a better understanding of likely metadata requirements beyond the taxonomy categories) we asked the British Council team (Mark Hamilton, Caroline Meek, Peter Hamilton) to take us through five sample use cases, where the target learning content for this project was being approached to serve specific business needs. The use cases were: (1) Identifying and selecting game-like content for young learners from LearnEnglishKids, organizing it according to themes, and Straits Knowledge January 2010

www.straitsknowledge.com

4


Taxonomy and Metadata Advisory Report

(2)

(3)

(4)

(5)

packaging it for easy use by the British Council’s media partners, to encourage them to use and/or link to British Council resources. Identifying and selecting engaging, adult learning content (podcasts) from LearnEnglishProfessionals, organizing it according to themes, and packaging it for easy use by the Speak Good English Movement as a part of their annual campaign. Identifying and selecting a variety of game-like content from LearnEnglishCentral, and creating an Excel list of links to relevant materials, named and categorized to make it easy for Singapore Press Holdings media partners to select and publish the resources on their sites. Identifying theme-related content (eg Christmas/Winter) to support the Speak Good English Movement from LearnEnglishKids and LearnEnglishCentral. Identifying online content (from within and beyond the British Council’s content collections) systematically mapped against the Singapore Ministry of Education’s English Language syllabus 2010. This syllabus was skills oriented and was a major input to the formation of the “Skills” facet in the draft taxonomy.

General taxonomy-related conclusions from the use case analysis were: •

The faceted approach is essential to be able to serve different search requirements quickly (eg selection by theme/topic, material type, language skill etc). Material needs to be selected and managed across content collections, and so a single consistent taxonomy needs to apply across all content collections (this poses some interesting challenges for “topics” where we would expect some variation in language between eg topics for young learners and topics for professionals). The most commonly leveraged facets across the use cases were Language Level, Topic, Skill and Audience.

2.3 What will the facets be used for? The facets will need to support: • Supporting browse and navigation on public and internal sites • Filtering content by attributes for packaging purposes, or as the results of a search, or for publishing to specific sites • Supporting keyword search (by matching search terms to the taxonomy terms, and supporting with a thesaurus to match common alternate terms used by searchers to taxonomy terms)

2.4 Taxonomy usability and maintenance issues There are some drawbacks to a faceted approach, particularly one using up to twelve facets. “Normal” searchers or browsers for content (whether

Straits Knowledge January 2010

www.straitsknowledge.com

5


Taxonomy and Metadata Advisory Report internal staff or users of public sites) do not consciously “decompose” their enquiries by facets, and being faced with twelve facets to choose from can be confusing and off putting. Good practice is to expose only the 2-3 facets that best represent the most common frames of enquiry among users, and then use the supplementary facets for successive filtering of the enquiry or to perform other content management operations. For example, the British Council’s content on public sites is largely organized by audience type. If content is held on a central repository, an Audiences facet could be deployed to automatically publish relevant content to relevant public sites. This has the advantage of minimizing the manual maintenance of different content sites. In this case, the Audiences facet may not be exposed to the general public for general navigation purposes (although it would be exposed to British Council staff on an internal content site for selection purposes). In addition, the ongoing management of multiple facets can be problematic if the scope and purpose of the facet is not documented and well understood; if the scope or content of the facet is subject to frequent change; if there is ambiguity between facets requiring more manual interventions to “tidy” confusing classifications of content. With this in mind, it is important that the facets: (a) accommodate as far as possible the full range of likely uses of the learning content (even if not currently visible); (b) are as simple and as stable as possible to minimize future maintenance effort; (c) are deployed consistently across all related content collections; and (d) minimize all possible ambiguities and redundancies between facets, to support rapid, accurate and consistent classification by British Council staff. These principles have informed the detailed analysis of the facets given below and resulting recommendations.

2.5 Detailed facet analysis and recommendations This section examines the facets individually. The summary of the analysis, summary recommendations, and overview of how the facets can be used can be found in the next section after this. We have distinguished between: • Primary Facet – represents a common way of starting an enquiry for content • Secondary Facet – represents an infrequent way of starting an enquiry, but may be useful for filtering content for specific needs

Straits Knowledge January 2010

www.straitsknowledge.com

6


Taxonomy and Metadata Advisory Report •

Selection Facet – represents an important but infrequent, specialized query on the content, not likely to be used for secondary filtering (eg type of interactive)

Facet Name Purpose and Potential Use

Likely Audiences Source(s) of Vocabulary Warrant

Possible Issues

Recommendations

Management

Facet Name Purpose and Potential Use Likely Audiences Source(s) of Vocabulary Warrant Possible Issues

Recommendations Management

CEFR Level To enable people to filter content by language ability level – mainly to select content for presentation to other audiences, or for students to select material suited to their levels. Appeared in 3 out of 5 use cases. PRIMARY FACET FOR ALL AUDIENCES. BC staff; teachers; students; educational administrators CEFR for English language learning CEFR is a widely accepted standard in Europe, and can be mapped against other language level description standards. Terminology not transparent, requires knowledge of CEFR; “CEFR” itself will not be transparent to nonTEFL professionals; if other level descriptors are adopted, there will be competing facets (1) Descriptors of the levels should be available in a scope note to each term; (2) consider naming the facet “Language Level”; (3) standardize to CEFR and if other level schemas are used by the BC or key audiences, use a thesaurus to link alternate terms to appropriate CEFR terms. • Who will manage the terminology and the thesaurus of alternate terms? • Intensity of management: light (external standard, not expected to change frequently). Age To enable people to filter content by age range Content managers; teachers; educational administrators n/a n/a Designed to capture schooling levels but otherwise artificial; the same function can be served by the “Audiences” facet; two similar facets can lead to confusion (which ones to tag?) or the need for double tagging. (4) Use the Audiences facet to express age range • None if dropped

Straits Knowledge January 2010

www.straitsknowledge.com

7


Taxonomy and Metadata Advisory Report Facet Name Purpose and Potential Use

Likely Audiences Source(s) of Vocabulary

Warrant Possible Issues

Recommendations

Management

Audiences To enable people and administrators to filter content by demographics (eg to pipe content to different presentation sites). In the use cases, this was the most intensively needed facet, but is primarily directed at internal rather than client use. Appeared in 5 out of 5 use cases. PRIMARY FACET FOR ALL AUDIENCES. Content managers; BC staff; teachers; parents; educational administrators; media partners Academic: from internal British Council sources Professional: drawn from BC World Intranet for references to common customer groups, and BC listserves where BC staff were requesting materials for particular customer groups (eg military, airlines) Social: came from internal discussions in the Singapore office Not needed, if facet is very simple and the language completely transparent Confuses audiences-as-learners with audiences that will deploy the content for learners, resulting in the repetition of education levels for students, teachers and parents; it’s not clear why the professions and content for adults need to be broken down in such detail, unless you have a lot of content covering professions; focus area of content could be handled through a topics facet eg your “Work and Business” category (5) Adopt a simpler facet covering Learners broken down by level (including adult learners), Teachers, Parents; the expectation will be that tagging will always be done for learners, but that tagging for parents and educators will only be done for content specifically designed for those audiences, plus the learner tagging where appropriate • •

Facet Name Purpose and Potential Use

Likely Audiences

Who will manage the terminology and the thesaurus of alternate terms? Intensity of management: light – small facet, new audiences not likely to emerge frequently.

Topics To enable people to filter content by subject matter; provides a controlled vocabulary of topics to which other related terms can be linked through a thesaurus; will support search and browsing by theme. Appeared in 3 out of 5 use cases. PRIMARY FACET FOR ALL AUDIENCES. BC staff; teachers; students; parents; educational

Straits Knowledge January 2010

www.straitsknowledge.com

8


Taxonomy and Metadata Advisory Report Source(s) of Vocabulary

Warrant

Possible Issues

Recommendations

administrators Categorisation developed for a self access centre database in Hong Kong and subsequently adapted and used in Portugal, has been reused around the BC network [suggest starting with the existing topic keywords for content on all BC sites with instructional content, self access topic lists where available; supplementing where necessary from relevant external sites] Terms in use within the British Council network [suggest cross checking for completeness with an external educational content site such as BBC schools] This is one of the potentially more complex facets to create and maintain since “topics” are completely unbounded. It needs to adequately represent/accommodate the inherited topic keywords from existing content collections; could require intensive management because of frequent emergence of new topics, and could become very unwieldy to manage unless a simple way of handling new vocabulary is adopted; will be difficult to maintain topics for different-aged audiences in a single controlled vocabulary. (6) Consider a hybrid semi-controlled system with a controlled, relatively high level and comprehensive set of topic categories (no more than 15) but allow “free” tagging by content creators and content users within those categories – ie when people contribute a tag, they contribute INTO a controlled category (use card sorts with existing topic keywords on all your existing content to derive your master high level list of categories – no “OTHER” category!); (7) Perform regular housekeeping on tags contributed to weed out redundancies, spelling errors etc; (8) Use a thesaurus to link the tags to the controlled categories so that keyword searches can provide related resources (9) Present the tags as sub topics of the master categories in a browse format, but relying on the audience facet to filter and present the tags attached only to that audience’s content (in a public browse format, the browse page should already be tuned/filtered to a particular audience; in an internal browse format, the audience should normally be selected before topic) (10) Have a “suggest” as you type function so

Straits Knowledge January 2010

www.straitsknowledge.com

9


Taxonomy and Metadata Advisory Report

Management

Facet Name Purpose and Potential Use Likely Audiences Source(s) of Vocabulary

Warrant Possible Issues

Recommendations

Management

Facet Name Purpose and Potential Use

Likely Audiences

that taggers can either select existing tags for their content or type their own (11) Use frequently used search terms to create new tags where necessary (12) Use tag clouds as an alternate form of browse navigation for eg most popular tags, most popular search terms linked to content • Who will manage the terminology and the thesaurus of alternate terms? • Intensity of management: moderate to high, mostly housekeeping on tag contributions. Skills To enable people to filter content by language skill being exercised or presented. Appeared in 3 out of 5 use cases. PRIMARY FACET FOR ALL AUDIENCES. BC staff; teachers; students (for self-study); educational administrators; instructional designers Language skills elements drawn from MOE Singapore curriculum, compared with syllabuses and curricula from other countries, cross-checked with English coursebooks; the Singapore one was the most comprehensive, and tested well against example content. Well researched, terms in current professional practice, and tested against content. This is currently a three level hierarchy, which will make it more difficult to browse; third level terms are complex which will make keyword matching in searches more difficult; “workplace skills” potentially very large unless bounded as “workplace communication”, may overlap with Topics facet; “study skills” looks incomplete. (13) Collapse levels 1 and 2 eg “Listening: Recognition”; (14) Simplify the terms at the third level and reduce repetition of terms (eg “techniques”) (15) Use standard textbooks or curricula to check completeness and for warrant. • Who will manage the terminology and the thesaurus of alternate terms? • Intensity of management: light (stable and comprehensive term set). Functions To enable people to rapidly identify content by language function being practiced. Appeared in only 1 use case as a potential secondary facet. SECONDARY FACET. BC staff; teachers; educational administrators; instructional designers.

Straits Knowledge January 2010

www.straitsknowledge.com

10


Taxonomy and Metadata Advisory Report Source(s) of Vocabulary Warrant Possible Issues

Recommendations

Management

Facet Name Purpose and Potential Use

Likely Audiences Source(s) of Vocabulary Warrant Possible Issues

Recommendations

Management

Sourced from Hong Kong self access centre database, enhanced with teacher consultations. Terms are in use in the British Council network. This is a list of 44 items which would make it difficult for rapid navigation and selection. Related functions are not grouped together. (16) Consider clustering into 3-5 groups and creating a two-level facet – use card sorting with teachers and check with students for validation to create them; check with standard textbooks or curricula for warrant. • Who will manage the terminology and the thesaurus of alternate terms? • Intensity of management: light (stable and comprehensive term set). Vocabulary To enable people to rapidly identify vocabulary related content by type or focus of activity. Appeared in 0 out of 5 use cases. SECONDARY FACET. BC staff; Teachers; educational administrators; instructional designers. Came from British Council Hong Kong, and checked against Cambridge vocabulary book. Terms in current professional practice. This is a list of 31 items which is a little long for rapid navigation and selection; seems to combine what is being taught (eg antonyms) with how it is taught (eg word games – is there a clash with Interactivity Facet?) and then some skills work; some apparent overlap with Grammar Facet (17) Should ideally be broken down into two levels, with top categories that make sense to likely users – use card sorts and reference standard textbooks/ curricula for warrant (18) If it makes sense to users (teachers) consider merging into a combined Grammar and Vocabulary Facet • •

Facet Name Purpose and Potential Use Likely Audiences

Who will manage the terminology and the thesaurus of alternate terms? Intensity of management: light (once organized for usability should be fairly stable).

Pronunciation To enable people to rapidly identify pronunciationrelated content and filter by type. Appeared in 0 out of 5 use cases. SECONDARY FACET. BC staff; teachers; students; educational administrators; instructional designers

Straits Knowledge January 2010

www.straitsknowledge.com

11


Taxonomy and Metadata Advisory Report Source(s) of Vocabulary Warrant Possible Issues

Recommendations

Management

Facet Name Purpose and Potential Use

Likely Audiences Source(s) of Vocabulary Warrant Possible Issues

Recommendations

Derived from British Council Hong Kong. Terms in use in the British Council network. Gives a list of 32 items which will not be easy to rapidly navigate and select; at a level of detail that will likely be used only by teachers and instructional designers; could technically be considered a part of the “skills” facet. (19) Consider breaking down into 3-6 related groups of pronunciation focus, using standard textbook/ curricula for warrant; (20) Consider incorporating this top level into the skills (speaking) facet IF it reflects teacher/student expectations and navigation choices (test for usability); (21) Consider not deploying the more detailed layer unless and until the volume of content warrants it. • Who will manage the terminology and the thesaurus of alternate terms? • Intensity of management: light (stable and comprehensive term set). Grammar To enable people to rapidly identify grammarrelated content and filter by grammar area. Appeared as a potential secondary facet in 1 out of 5 use cases. SECONDARY FACET. BC staff; teachers; educational administrators; instructional designers Derived from Hong Kong Terms in use in the British Council network Uneven depth with tenses having three levels and others at one level; this will create navigation complexity; not always clear why some items are in Grammar Facet and not Vocabulary Facet; the “Discourse” category could have some overlap with skills and functions. (22) Seek to balance out the depth of this structure by collapsing the top two layers of the Tenses (eg using compound terms “Tenses: Present) and sorting single layer terms either into clusters or rationalizing with the Vocabulary Facet; use card sorting exercise to check for ambiguity in the “discourse” category vis a vis the skills and functions facets; check for warrant for the terms and structure with standard textbook and/or curriculum (23) If it makes sense to users (teachers) consider merging into a combined Grammar

Straits Knowledge January 2010

www.straitsknowledge.com

12


Taxonomy and Metadata Advisory Report Management

• •

Facet Name Purpose and Potential Use

Likely Audiences Source(s) of Vocabulary Warrant Possible Issues

Recommendations

Management

Facet Name Purpose and Potential Use

Likely Audiences Source(s) of Vocabulary

and Vocabulary Facet Who will manage the terminology and the thesaurus of alternate terms? Intensity of management: light (stable and comprehensive term set).

Text Types To enable people to rapidly identify content by genre of text types used, probably in combination with Reading/Writing skills work or Vocabulary work. Most likely as a subsequent filter applied to results on a skills facet search. Appeared in 0 out of 5 use cases. SECONDARY FACET. BC staff; teachers; students; educational administrators Document from Martin Peacock – Hong Kong? Also supplemented with text types from the Singapore Ministry of Education and web research. Terms in use in Singapore MOE. This is a manageable list from the navigation and selection point of view, but the terms do not unambiguously communicate “text type” – some could also suggest functions, skills, vocab or even topic. (24) Consider the likely usage of this facet, and whether you will have a need to filter on this basis because of the volume of content as well as anticipated need (check with sample users). If the answer is no, then this facet could be dropped until you need it. If the answer is yes, then the terms used will have to be disambiguated to communicate genre quite unambiguously otherwise you will lose precision on navigation and search. (25) Consider whether “Genre” would work better than “Text Type” as a label for this facet. • Who will manage the terminology and the thesaurus of alternate terms? • Intensity of management: could be problematic for ambiguity reasons unless thoroughly tested. Interactives To enable people to rapidly identify content by the form of interaction in the content. Appeared in 2 out of 5 use cases, but it was not clear that the level of granularity proposed here was really required. SELECTION FACET FOR LIMITED AUDIENCES. BC staff; teachers; instructional designers. List of interactive types used in the BC websites.

Straits Knowledge January 2010

www.straitsknowledge.com

13


Taxonomy and Metadata Advisory Report Warrant Possible Issues

Recommendations

Management

Facet Name Purpose and Potential Use Likely Audiences Source(s) of Vocabulary Warrant Possible Issues

Recommendations

Current terminology on the BC sites. This looks like it would function best as a list of templates for instructional designers, rather than as part of a retrieval or browse taxonomy. “Games” is a term that could actually encompass many of the interactivity forms in the rest of the list. From a general user’s perspective, it may be that you will need a “teaching resources” category within the Teaching Facet, which could include eg games, lesson material, lesson plan, reading material, controlled practice exercise; we know from the use cases is that it should be possible to identify “games” quickly and easily. (26) Consider separating interactivity type (for internal use by instructional designers) from a new “learning content type” facet which could be used as a secondary facet for teachers and BC staff; not clear if this needs to be exposed on public sites. • Who will manage the terminology and the thesaurus of alternate terms? • Intensity of management: for interactivity type, simple – pending validation with instructional designers; for instructional content type, moderate (needs to be researched, developed and tested). Teaching Support To enable teachers to identify teacher-related content. Appeared in 0 out of 5 use cases. SELECTION FACET FOR LIMITED AUDIENCES. BC staff; teachers; teacher trainers; educational administrators Teaching English BBC British Council website. Not clear. This is the least well developed of all the facets; combines material related to teaching competencies with teaching resources. It is not clear from the review so far that there is a large amount of content for teacher competencies, meaning there will be limited scope for testing potential usage and usability of a facet among target audiences. (27) Consider using the suggested “instructional content type” facet for some of the topics here. (28) Consider using the BC’s teaching competency framework and standard teacher training curricula as warrant for drafting a more fully developed “teaching competencies” facet; test this and validate it, as developmental/training content for teachers becomes available online

Straits Knowledge January 2010

www.straitsknowledge.com

14


Taxonomy and Metadata Advisory Report

Management

(29) Recommend changing the name to “Teaching Support” to distinguish it from the “teachers/educators” category in the Audiences Facet • Who will manage the terminology and the thesaurus of alternate terms? • Intensity of management: for instructional content type, moderate (needs to be researched, developed and tested); heavy for teaching competencies (more extensive vocabulary, small volume of content to test with target audiences).

2.6 Summary of facet analysis and recommendations Facet

Summary of Recommendations

CEFR Level

Primary navigation facet (public)

Age

Drop, use Audiences Facet instead

Audiences

Primary content distribution facet to different sites – needs radical simplification

Topics

Primary navigation facet (public) – use hybrid categories plus tags

Skills

Primary navigation facet (public) – collapse to two levels, incorporate Pronunciation Facet

Functions

Secondary filtering facet + Advanced Navigation – group into categories

Vocabulary

Secondary filtering facet + Advanced Navigation – group into categories OR incorporate into a combined Grammar and Vocabulary Facet

Pronunciation

Merge with Skills Facet – create higher level list of categories

Straits Knowledge January 2010

www.straitsknowledge.com

15


Taxonomy and Metadata Advisory Report Grammar

Secondary filtering facet + Advanced Navigation – balance out uneven depth of hierarchy and rationalise the overlap with Vocabulary facet OR incorporate into a combined Grammar and Vocabulary Facet

Text Types

Secondary filtering facet + Advanced Navigation – disambiguate terms used to clearly communicate text genre

Interactive Types

Selection facet – is it necessary/useful to expose this on public sites?

Teaching Support

Selection facet – needs more development; can feed content into “Teachers” sections” on public sites

The diagram below shows a visualization of how the facets could be used successively to support guided navigation and filtering on a public site, based on the recommendations made above. It assumes that the Audiences facet is used primarily to channel content to relevant sites, and is not visible to the public. For general navigation purposes it only exposes the Topics, Skills and Level facets to the general user for their first selection, and then each set of results can be supplemented by Function, Grammar&Vocab, Text Types, if necessary. Users would also have access to an “advanced navigation” panel where they could choose any combination of attributes they wanted.

Straits Knowledge January 2010

www.straitsknowledge.com

16


Taxonomy and Metadata Advisory Report

The interface for an internal British Council view of a site would be very similar, except that it would also need to add the Interactives facet to the “advanced navigation� panel.

Straits Knowledge January 2010

www.straitsknowledge.com

17


Taxonomy and Metadata Advisory Report

3. METADATA RECOMMENDATIONS 3.1 The general metadata approach is sound Metadata refers to structured information about a document or piece of content. For example, author, title, subject keywords, date of publication, place of publication, archiving date, access permissions, are all individual pieces of metadata that might be associated with a document. The British Council has decided to follow the Dublin Core standard for its metadata framework. This is a simple, widely adopted baseline metadata standard which can easily be adapted to local needs, and can be supplemented with more specialized standards (such as SCORM) where necessary. Because it is widely adopted, any future need to integrate with, or migrate to a new standard will likely be supported. Dublin Core can be used or supplemented to serve most of the metadata needs with the exception of metadata to track content usage.

3.2 Analysis of British Council use cases Detailed discussion with the British Council team around the use cases referenced in section 2.2 above, revealed the following issues and possible future needs involving metadata support over and above supporting search and retrieval via the taxonomy components of metadata. The types of metadata indicated under each point will be explained in section 3.3 below. (1)

(2)

(3)

(4)

When content is migrated or websites reorganized there is a need to update media partners who are linking to the content with the new URLs; good practice in content management is that metadata should use a Uniform Resource Identifier (URI) which will give permanent and stable access to a content object while its location may change across platforms and migrations (Identification Metadata). When users or partners have made substantial use of a content collection in the past, they will find it helpful to be able to identify newer or more recently contributed content (Identification Metadata using date of publication). It is envisaged that one potential benefit of site membership in the future might be for users to be able to distinguish “their� content (content they have viewed, bookmarked, shared with others) (Usage Metadata). In some cases, there are pieces of content that fit together into a coherent whole or a sequence such as a graduation in level of difficulty (eg in LearnEnglishProfessional), so while the content pieces might be identified and tagged separately, those relationships should also be preserved (Management Metadata).

Straits Knowledge January 2010

www.straitsknowledge.com

18


Taxonomy and Metadata Advisory Report (5) (6)

(7)

It would be useful to be able to identify high value content – eg by tracking the popularity of content through frequency of use, frequency of sharing with others (Usage Metadata). When the British Council offers links to its content to external partners, the content object needs to appear within a structured British Council context (including links to other related content), and not for example as a decontexualised popup box. At the moment, these are standard web page contexts on the public sites, but there is no reason special templates could not be customized for external partners, drawing on the same units of content from the public collection. This implies a possible need for metadata that tells a content management system for a given piece of content what kind of templates or contexts the content should be linked to (Management Metadata). The discussion around the Ministry of Education mapping project (Use Case 5) raised the possibility that although at present the content objects are largely seen as discrete items, there may be situations where it would be useful to define instructional sequences (similar to point 4 above) (Management Metadata).

3.3 What will the metadata be used for? Metadata is used for a number of purposes, and not just for subject categorisation: •

Identification: It identifies content – eg some descriptive metadata captures author and title fields and enables you to distinguish each piece of content as a unique entity.

Management: It helps systems manage content so that it does not have to be managed manually (large volumes of content become rapidly unmanageable using manual methods), – eg administrative and structural metadata capture things like version numbers, archiving and other significant dates, security and access permissions, relationships with other content objects including sequencing, file types, links between certain file types and the preferred applications to “play” them, workflows such as quality management and approval workflows that can vary according to the content, or instructional design workflows.

Retrieval: It aids retrieval of content – eg other descriptive metadata captures things like taxonomy topics, subject keywords, tags, document descriptions.

Usage: It can be used to connect content to other content – on social software sites, behavioural metadata is captured on the fly about the transactions that users have with each document and this is used to infer relationships or connections about the other things the same people transact with. The best known example of this is on Amazon.com, where if you buy a book, you can see the legend

Straits Knowledge January 2010

www.straitsknowledge.com

19


Taxonomy and Metadata Advisory Report “other people who bought this book also bought….” Unlike other metadata, which is usually assigned when the content is created, this metadata is collected by the systems managing the content as it is being used. Our observations from the use cases were that it would be useful for the British Council to plan for specific types of Identification, Management and Usage metadata in addition to the Retrieval Metadata supplied by the taxonomy.

3.4 Metadata usability and maintenance issues There are two major issues with metadata: (a) As with the taxonomy portion of metadata, it should be well planned upfront and designed for minimal maintenance and change in future. (b) As with the taxonomy portion of metadata, the collection or assignment of metadata should be as simple as possible to increase the likelihood of getting complete, accurate and consistent metadata. In practice, this means having only as much metadata as you need in order to serve your purposes, and reducing as far as possible the cognitive load when assigning metadata, and automating as much as possible. Note that (a) (plan so you don’t have to expand or change radically in future) and (b) (deploy the minimum metadata necessary) are in tension with each other and will have to be balanced.

3.5 Metadata recommendations This section presents an illustrative draft of the main elements of a metadata framework and approach for British Council learning content based on Dublin Core, and drawing on the observations from this study. It is not intended to be comprehensive or complete, nor does it claim to be the only way to achieve the same objectives. Identification Metadata

Field Name Identifier

Title

Element Name

Definition

Data Type or Source

Comment

Dublin core:identifier

An unambiguous reference to the resource within a given context.

Text string

Compulsory: System assigned URI reference string

Dublin core:title

A name given to the resource.

Text string

Compulsory: Pick up from content system, should be in instructional design template

Straits Knowledge January 2010

www.straitsknowledge.com

20


Taxonomy and Metadata Advisory Report for creating the content

Publisher

Dublin core:publisher

Name of the person/ entity primarily responsible for making the resource available.

Text string

Compulsory: Pick up from system, should be available from login data

Date Published

Dublin core:date

Date the resource was 'published' to the system.

Date, use format yyyymm-dd

Dublin core:description

An account of the resource.

text string

Compulsory: Pick up from system by default, (today's date) Optional: Can be picked up from instructional design templates; can be useful for aiding search engines which cannot access textual content inside the content (eg audio, video, Flash modules).

Description

Management Metadata – will be heavily influenced by the metadata requirements of the content management system

Field Name Audience

File Type

Element Name

Definition

Data Type or Source

Comment

Dublin core:audience

Audiences for whom the resource is intended or will be useful.

Values come from taxonomy facet: Audience

Compulsory: Select from drop-down list (multiple values possible) or Pick up from content system, should be in instructional design template for creating the content

Dublin core:format

The file format of the resource

Values from list of Internet Media Types [MIME]

Compulsory: Should be picked up automatically from the system; can be used to link to preferred players

Straits Knowledge January 2010

www.straitsknowledge.com

21


Taxonomy and Metadata Advisory Report Child Resource

Dublin core:hasPart

Identifies a related resource that makes up a part of this content resource being described

Value is the URI of the component part

Parent Resource

Dublin core:isPartOf

Identifies a related resource that this resource being described belongs to

Value is the URI of the parent resource

Player

Dublin core:isRequire dBy

Identifies resources that depend on this resource for their function, delivery or coherence

Values are the URIs of the dependent resources

Dependency

Dublin core:requires

Identifies a resource that is needed by this resource for its delivery, function or coherence

Value is the URI of the required resource

Optional: Use to identify resources that cannot play without the identified resource or application

Related Resources

Dublin core:relation

A related resource

Values are the URIs of the related resources

Optional: Can be used to create sequences or suggestions for other content resources.

Optional: Use where a resource can be disaggregated and used as a collective or as separate pieces Optional: Use where a resource can be disaggregated and used as a collective or as separate pieces Optional: Can be used for players or for different presentation templates used by the same resources

Retrieval Metadata

Field Name Topic

Element Name Dublin core:subject

Straits Knowledge January 2010

Definition The subject matter of the resource

Data Type or Source Values come from taxonomy facet: Topics

Comment Compulsory: Select from drop-down list (multiple values possible) – this list will need to draw from the full vocabulary of top level topics plus user and staff tags contributed within those topics

www.straitsknowledge.com

22


Taxonomy and Metadata Advisory Report Language Level

Dublin core:education Level

Levels of progression through an educational or training context, for which the described resource is intended.

Values from taxonomy facet: CEFR Levels

Compulsory: Pick up from content system, should be in instructional design template for creating the content – multiple values should be possible

Skill

Dublin core:subject

Identifies the skills focus being addressed in the resource

Values from taxonomy facet: Skills

Compulsory: Pick up from content system, should be in instructional design template for creating the content – multiple values should be possible

Function

Dublin core:subject

Identifies the language function being taught in the resource

Values from taxonomy facet: Functions

Optional: Pick up from content system, should be in instructional design template for creating the content – multiple values should be possible

Grammar & Vocabulary

Dublin core:subject

Identifies grammar or vocabulary areas being learnt through this resource

Values from taxonomy facet: Grammar & Vocab

Optional: Pick up from content system, should be in instructional design template for creating the content – multiple values should be possible

Text Type

Dublin core:type

Identifies the genre of the text types being presented or exploited in the resource

Values from taxonomy facet: Text Types

Optional: Pick up from content system, should be in instructional design template for creating the content – multiple values should be possible

Teaching Support

Dublin core:subject

Identifies resources that can be used as teacher

Values from taxonomy facet: Teaching

Optional: Pick up from content system, should

Straits Knowledge January 2010

www.straitsknowledge.com

23


Taxonomy and Metadata Advisory Report

Interactive

Dublin core:interactiv eResource

resources, or that refer to teaching competencies

Support

be in instructional design template for creating the content – multiple values should be possible; if teaching resources are split from teaching competencies, then Dublin core:type would be used for the teaching resources and Dublincore:instr uctionalMethod for the teaching competencies

A related resource

Values selected from taxonomy facet: Interactives

Optional: Use for interactives. Pick up from content system, should be in instructional design template for creating the content – multiple values should be possible to cover composite works

Usage Metadata This is metadata which is collected by the computer system on the usage patterns associated with the content. Dublin Core does not natively accommodate this. A records management metadata standard such as NAARMSCA1 can be adapted to provide some relevant fields. The SUSHI Protocol2 - is a standard that defines an automated request and response model for the harvesting of electronic resource usage data utilizing a Web services framework. It is intended to replace the time-consuming usermediated collection of usage data reports.

1

National Archives of Australia Recordkeeping Metadata Standard for Commonwealth Agencies available at http://www.naa.gov.au/recordkeeping/control/rkms/rkms_pt1_2.pdf 2 Standardized Usage Statistics Harvesting Initiative (ANSI/NISO Z39.93-2007) available http://www.niso.org/kst/reports/standards?step=2&gid=None&project_key=2de0e3e04f3 a7e32d45db8ee87574c3c8206ddcb

Straits Knowledge January 2010

www.straitsknowledge.com

24


Taxonomy and Metadata Advisory Report

Field Name Use History

Element Name

Definition

Data Type or Source

Comment

NAARMSCA: use history

The dates and descriptions of all transactions to access and use of a resource from date of publication

Metadata collected should link URI to identify of user, date of transaction and type of transaction

Optional: content system picks up identity of users, links to the record Identifier, and date/time stamps the transactions being monitored

Popularity

Not known

The intensity or frequency of designated user transactions associated with a resource (eg play, recommend, bookmark)

System generated summary of target user transactions

Optional: System can present a summary by resource of transactions

Rating

Not known

The ratings associated with a resource by users

System generated average of user ratings

Compulsory: Pick up from content system, should be in instructional design template for creating the content – multiple values should be possible

The number of possible metadata elements which will need human judgement will place a heavy burden on the content publishing process if the whole metadata template is imposed all at once. However as with the progressive exposure of taxonomy facets for navigation, the metadata required can be assigned in stages. We recommend that when content is created by an instructional designer, the following metadata is assigned in parallel with the instructional design process, using an instructional design template so that it can function as a design document as well as a metadata harvester for identification and retrieval metadata. This will work best if the instructional design and content creation process itself is automated within the same content environment as the content hosting and publishing system or with a seamless transfer of metadata, to preserve transmission of metadata from design to publication without the need for rekeying: • • • • •

Title (compulsory) Description (optional) Audience (compulsory) Topic (compulsory) Language Level (compulsory)

Straits Knowledge January 2010

www.straitsknowledge.com

25


Taxonomy and Metadata Advisory Report • • • • • •

Skill (compulsory) Function (where relevant) Grammar & Vocabulary (where relevant) Text Type (where relevant) Teaching Support (where relevant) Interactive Type (where relevant)

The following metadata elements should be system assigned or generated: • • • • • • •

Identifier (compulsory) Publisher (compulsory) Date Published (compulsory) File Type (compulsory) Use History (compulsory) Popularity (compulsory) Rating (compulsory)

This leaves the following metadata elements to be assigned by a content manager upon publication: • • • • •

Child Resource (where relevant) Parent Resource (where relevant) Player (where relevant) Dependency (where relevant) Related Resource (where relevant)

This is visualized in the diagram below.

Straits Knowledge January 2010

www.straitsknowledge.com

26


Taxonomy and Metadata Advisory Report

4. System Review This section contains an analysis of the current systems in the British Council network that service learning content in the Global English Products suite. The objective of this review is to analyze and recommend how to leverage the taxonomy and metadata recommendations given in the sections above.

4.1 Analysis of British Council learning content systems Given below is a list of English Global Products resources available on the web: Resource URL/Source Learn English Global http://learnenglishkids.britishco uncil.org/

Remarks

This website is in Drupal, using the native Drupal Taxonomy module

Teaching English http://www.teachingenglish.org .uk/

In Drupal, using native Drupal Taxonomy module

Premier skills http://premierskills.britishcounc il.org/ Learn English Central http://www.britishcouncil.org/le arnenglish.htm

In Drupal, using native Drupal taxonomy module

In Obtree, which has limited metadata management capabilities, to be migrated to Drupal

Moodle courses Learning techs http://courses.britishcouncil.org /pub/learningtechnologies/

Courses are done in Moodle. Flash content is integrated with Moodle courses. Moodle allows sequencing of learning content using SCORM, an elearning standard.

TKTE http://courses.britishcouncil.org /pub/tkt/ emoderators http://courses.britishcouncil.org Straits Knowledge January 2010

www.straitsknowledge.com

27


Taxonomy and Metadata Advisory Report /emoderators/ Primary Essentials http://courses.britishcouncil.org /pub/primaryessentials/ Engaging Teachers http://courses.britishcouncil.org /engagingteachers/ Elementary podcasts New – not yet hosted on the web

In Drupal, using native Drupal taxonomy module

Flash activities New – internally created, currently hosted only on test site

Created by in-house activity builder application. The output is a Flash file.

The current configuration is as follows: •

Content is stored in different systems. Content is currently published in Drupal, Moodle, Obtree and using a custom Activity Builder application. These systems are different in that they do not have a single consistent way to author and publish content. They also don’t have a single consistent way of classifying content.

The systems use different ways of managing and applying taxonomy categories. Drupal classifies content using the Drupal Taxonomy Module. The Flash Activity Builder uses its own custom method. This poses a substantial problem for consistent use of categories to organize and access content.

The ability to search for and aggregate related content is limited to the system that holds the content. In the current configuration, one cannot see related content from different systems, e.g., “Christmas” content from both Drupal and Obtree. This significantly reduced the depth of the offering to users, and the ability to manage the learning content as a resource for the British Council as a whole.

More effort is required to maintain different systems. Different processes and workflows are required to use the different systems. This increases the effort required to have the right people with the right skills to administer and manage these systems.

Straits Knowledge January 2010

www.straitsknowledge.com

28


Taxonomy and Metadata Advisory Report 4.2 The general content management approach is sound The current approach of progressively migrating and consolidating currently scattered content onto a single platform (based on Drupal) is sound, for the following reasons: •

It standardizes the authoring process. Different authors in different locations can use the same tools to author their content. This method also gives a single consistent way to attach metadata to the content. A single standard method is easier to teach and administer.

It standardizes the delivery process. Content can be delivered in a consistent way using Drupal templates. This way, the features and functions of content delivery that Drupal provides (related content, commenting, rating, send-to-friend, etc.) can be standardized across products.

It promotes browse and search. A single platform makes it easy for both for visitors and British Council staff to browse and find content in the different systems on the same platform. However, if content is not physically located in the same repository (eg several repositories standardized to Drupal with a common metadata framework) this calls for a search engine that can search through the different systems and aggregate related content. Kevan Cummins is currently exploring one such search engine called “Solr” which is open source. Endeca is an example of a commercial offering in this space.

It promotes reuse. A single platform following a standard taxonomy can help feed related content. A user viewing a language piece on “Christmas” could get related feeds from different systems, say from both Teaching English and Learn English Global. However, this calls for the different systems to expose an Application Programming Interface or API to help build these feeds.

In terms of the Drupal platform itself, we examined the platform for its capability to deliver the requirements set out by our recommendations above for how the taxonomy and metadata might be deployed. Drupal has a very mature taxonomy management capability, and has a large and active user community. Our assessment is that it is capable of achieving at least 80-90% of the requirements we have set out above, and should be able to meet all of them with some additional work. Dublin Core metadata can be collected and assigned to content within Drupal itself without the need for an additional external tagging application. The highlights of our assessment are given in the table below:

Straits Knowledge January 2010

www.straitsknowledge.com

29


Taxonomy and Metadata Advisory Report Strengths

Potential Issues

Drupal taxonomy management allows the addition of synonyms or alternate terms, enhancing the ability to link common search terms to taxonomy terms.

Vocabularies (facets) are assigned to content types, rather than at the content item level, which may reduce some flexibility in the use of facets at content item level.

Drupal can collect and accommodate all the types of metadata we have proposed, covering identification, management, retrieval and usage.

Different types of metadata seem to be handled differently by different modules in Drupal, not all of them from within the Taxonomy management module, which may complicate the metadata management effort somewhat. The current Drupal analytics module is not very advanced in its ability to handle the desired usage metadata.

Drupal taxonomy management allows the contribution of userassigned tags to content, alongside the use of controlled vocabularies.

In considering our recommendation for a hybrid approach on Topics, with user-assigned tags being added to a controlled set of high level topic categories it is not clear that Drupal can presently combine both controlled and uncontrolled terms in the same vocabulary (facet). This needs to be explored.

When exporting content, Drupal can also export all of the critical metadata that belongs to that content, which will allow other systems to leverage the same taxonomy and metadata.

It is not currently known whether Drupal can supply an Application Programming Interface (API) to allow other applications to access content within Drupal. This needs to be explored.

4.3 System recommendations We strongly recommend using a single platform approach to publish and manage the British Council’s learning content, for the reasons listed above. The Drupal platform can be used for this purpose. However, there are 4 different configurations that can be explored.

Straits Knowledge January 2010

www.straitsknowledge.com

30


Taxonomy and Metadata Advisory Report Option 1: Single Drupal installation powering all sites In this configuration, a single instance of Drupal is used to power all BC English Language products.

Option 2: Different Drupal installations, connected together using an API In this configuration, the different installations of Drupal talk to each other using the common API.

Straits Knowledge January 2010

www.straitsknowledge.com

31


Taxonomy and Metadata Advisory Report Option 3: Different Drupal installations, leveraging a single hosted taxonomy and metadata system In this configuration, metadata and taxonomy across all Drupal installations is centrally managed and published to those installations through a centrally hosted taxonomy and metadata management system such as Synaptica, Wordmap or Schemalogic.

Option 4: Different Drupal installations, with all pages exposing metadata on content pages using Dublin Core and using a search engine like Solr or Endeca to leverage the exposed metadata In this configuration, standardised metadata is exposed on every page of content using Dublin Core standards (ie when content is published to a page, its metadata is published with it, within the <META> tags in the html page header code). See official information at Dublin Core: http://dublincore.org/documents/dcq-html/ . This is based on the assumption that there is one content item per page â&#x20AC;&#x201C; it is standard web CMS practice that even if multiple content items are presented in a page, each individual content item has its own detail page, and this is where its unique metadata will be exposed. A search engine such as Endeca (or Solr) can then be used to crawl, classify and aggregate the content across the different systems. The search engine can in turn expose the search content as Really Simple Syndication or RSS feeds, and in Endecaâ&#x20AC;&#x2122;s case it can support guided navigation spanning content across all the distributed repositories.

Straits Knowledge January 2010

www.straitsknowledge.com

32


Taxonomy and Metadata Advisory Report

Weighing up the Options The pros and cons of these 4 configurations are given below: Configurations

Pros

Cons

Option 1: Single Drupal installation powering all sites

Content classification and cataloging can be centrally managed. Content can be mixed and matched in any design to cater to business needs.

Difficult to implement in practice. A single Drupal instance may not be able to partition and sustain different sites.

Option 2: Different Drupal installations, connected together using an API

Gives the ability to search and find related content from other API-enabled sites.

Requires custom programming and standardization. The collection of usage metadata needs additional work to be able to consolidate it across installations.

Straits Knowledge January 2010

www.straitsknowledge.com

33


Taxonomy and Metadata Advisory Report Option 3: Different Drupal installations, leveraging a single hosted taxonomy and metadata system

Taxonomy and metadata is managed centrally using a taxonomy management system, providing complete standardization of metadata.

Very expensive option. Also, requires extensive custom programming to get the Drupal systems to communicate with the taxonomy management system. The collection of usage metadata needs additional work to be able to consolidate it across installations.

Option 4: Different Drupal installations, with all pages exposing metadata on content pages using Dublin Core and using a search engine like Solr or Endeca to leverage the exposed metadata

Most flexible option. Requires all content pages to expose standard metadata as Dublin Core attributes and values.

Commercial search engines such as Endeca can be expensive so commensurate value of the engine needs to be established. There are open source alternatives. Apacheâ&#x20AC;&#x2122;s Solr open source search engine might be able to achieve similar results and is currently being explored by Kevan Cummins. The open source option would require support and custom programming expertise to integrate it with the Drupal system.

The search engine can then pick up the content and also map it to the taxonomy and metadata categories. The search engine can be used as the central search, navigation, find and content aggregation system.

If speed of implementation, support and long-term commitment is required then the commercial option may be the most suitable one. In this option the collection of usage metadata would be more complex and needs to be planned Straits Knowledge January 2010

www.straitsknowledge.com

34


Taxonomy and Metadata Advisory Report for carefully, since content is being accessed through both Drupal and non-Drupal environments.

4.4 Player concept One of the requirements given by Mark Hamilton was the desire to be able to harvest analytics data (eg usage metadata) of the different learning content items. Apart from the usage metadata we have suggested, analytics data could include number of views, time spent on the item, geographical locations, etc. The analytics requirement can be accommodated by using a “British Council Player”. A “player” is a Flash container that plays learning content. But the player can hold more information not visible to the user such as taxonomy and metadata values. It can also be used to collect analytics data. However the harvesting of meaningful usage metadata on learning content relies on two important additional factors: • notwithstanding the use of multiple Drupal installations, there must be only one single source for any individual content item, for which usage metadata is to be collected and analysed; while the content may be published to different platforms, the content and its metadata must come from a single source, otherwise instances of its use cannot be brought together; • usage metadata to be used for analytics purposes need to be aggregated into a single, central analytics module for processing and reporting. The “player” concept is an extension of the principle in our fourth configuration option above, that content should travel with its metadata, allowing for a pragmatic, loosely coupled environment, where common standards are applied across different deployments, but the burden and expense of centralized management is reduced. There are two ways that metadata can travel with its content (a) by being published with the content to an html container page (b) by being published with the content into a British Council Flash “player”, which allows the content to be played in both Drupal and non-Drupal environments. This has the added advantage of being able to harvest usage metadata which can then be aggregated back to a central analytics module. If we put these together we can sketch the following solution.

Straits Knowledge January 2010

www.straitsknowledge.com

35


Taxonomy and Metadata Advisory Report

The concept of the player has the following advantages: • • • •

can ensure consistent BC branding across all products can harvest custom analytics and feed to a central analytics module can aggregate related content from different Drupal installations using standardized metadata can be embeddable in different contexts (which means that learners can share this content via webpages, blogs and wikis) and can be custom-packaged and given to clients (e.g. a custom package of content for an educational or media partner) can be shared across social sites (Facebook, Twitter, etc.)

The player concept can be used with any of the Drupal configurations we listed above, but it is especially useful in helping to manage consistency across multiple Drupal and non-Drupal installations.

Straits Knowledge January 2010

www.straitsknowledge.com

36


Taxonomy and Metadata Advisory Report

5. Strategy and Governance 5.1 The importance of having strategy and governance If the British Council wants to be able to get maximum value, reusability and ease of recombination out of its learning content to serve its different stakeholders, it needs a consistent and standardized way of managing the way the content is produced, stored and accessed. At present the creation and publishing of British Council learning content is ungoverned, with multiple content collections on different platforms, being managed in different ways. As the volume and diversity of content grows, the more problematic it will be to exploit the value of its content in a cost effective way. There is no provision against the possibility of establishing new content platforms, or standards to guide how such platforms might be implemented. It is not sufficient to establish standards for content management. The British Council also needs a strategy for how to move to its desired state where standards are being applied and creating value, and it needs a governance framework to ensure that content continues to be managed effectively and consistently over time. The British Council needs to develop a content management strategy to: •

• •

establish a common, pragmatic set of principles for how learning content should be managed within the British Council, that meets the needs of both British Council content creators and content exploiters; describe the objectives, features and conditions of the desired content management environment including the standards which will be mandated to achieve consistency in content management; set out the steps to be taken to get to the desired state; and identify the resources required to move to the desired state.

The British Council needs to develop a content management governance framework to: •

• •

establish roles and responsibilities to ensure that the content management principles and objectives set out in the strategy are consistently implemented, or adapted to changing circumstances and needs; identify the core processes that need to be put in place; and set out the operating parameters within which technology systems (for creating, publishing, and accessing content) need to operate.

Straits Knowledge January 2010

www.straitsknowledge.com

37


Taxonomy and Metadata Advisory Report 5.2 Roles and responsibilities A governance framework would normally define who is responsible for what, for both implementation and oversight purposes, and how decisions are taken. It is important that: •

responsibilities are clearly described, agreed, and embedded in a policy document which is communicated – and permanently available for reference - to all staff associated with content management tasks; there is a clear description for how the principles and objectives of the content management strategy were arrived at, who is responsible for serving as a point of reference on those principles and objectives, and how decisions get made about changes to the principles, or on other important content management matters; and critical oversight or administrative roles (eg learning content manager, content management policy manager) might be secondary roles; as far as possible they should be associated with structural roles rather than project roles, to ensure against the inadvertent loss of oversight when a project role is discontinued.

5.3 Processes While metadata standards facilitate consistency in how the content is managed within a system, the British Council will also need to define and document some standard processes to ensure that content ends up in systems with the appropriate metadata, and that this metadata is exploited usefully. Key processes will include: • how a content site (eg Drupal) is proposed, approved and set up; • how instructional design gets done and approved; • how metadata is checked for consistency and quality; • how and when analysis is done on search activity and usage metadata; • how new content management enhancements are introduced; • how technology platforms and tools are assessed and approved; • how decisions are taken on amendments to the taxonomy, or changes/expansions to the metadata framework; • how changes are made to the content management principles and objectives; • how issues with compliance to the metadata standards are dealt with; and • how staff working on content management activities are supported with education, awareness and training.

Straits Knowledge January 2010

www.straitsknowledge.com

38


Taxonomy and Metadata Advisory Report 5.4 Systems The technology landscape in content management is still changing rapidly, creating new opportunities as well as complexities. We believe that a single bet on a single technology – even a strong one like Drupal, needs to be hedged. One of the ways that the bet can be hedged is by ensuring that the content itself travels with its metadata, meaning that it can be migrated to new platforms as well as hosted on diverse platforms. No technology platform should be adopted that does not allow the free export and transfer of content with its metadata. Another way of hedging the bet is to draw up and maintain a set of requirements for a technology platform to meet the needs of the content management strategy, principles and objectives, independently of whatever technology platform is currently adopted. This has three major benefits: • •

it allows the current platform to be periodically assessed for suitability; it can help to identify gaps in the current technology where customization, further development, or integration with other tools would help; it can be used to assess competing technologies for suitability as a possible replacement for the current technology as it ages.

Patrick Lambe and Maish Nichani February 8 2010

Straits Knowledge January 2010

www.straitsknowledge.com

39


Taxonomy and metadata advisory report