DCMI Education Community
- **Introduction** (Jon Mason & Stuart Sutton) 
Education Working Group:
Report of Deliberations
30 April 2000
Jon Mason & Stuart Sutton, Co-chairs© 2000 DCMI. All Rights Reserved.
- General Propositions
- Criteria for Evaluating Element and Qualifier Proposals
- General Classes of Metadata Elements Considered
- Element and Qualifier Proposals
- Proposed Audience Element
- Recommendation #1 (Element & Element Qualifier)
- Recommendation #2 (Element & Value Qualifier(s))
- Justification for Audience Element Proposal
- Proposed Standards Element
- Recommendation #1 (Element & Element Qualifiers)
- Recommendation #2 (Relation Element Qualifier)
- Justification for Standards Element Proposal
- Other Proposals
- Referencing the IMS/IEEE Namespace
- InteractityType & InteractivityLevel
- TypicalLearningTime (Duration)
- Quality of a Resource
- Resource Types
- Open Issues
- Teaching Processes/Characteristics (Pedagogy)
- General Statement
- Research
- Level (Course/Academic/Training/Age Level)
- The Face-to-Face Meeting in Melbourne (Kattemingga)
- Attendees at Kattemingga
- Meeting Agenda
Introduction
In August 1999, the Dublin Core™ Advisory Committee (DCAC) formed the DC-Education Working Group (DC-Ed WG) with the charge to discuss and develop a proposal for the use of Dublin Core™ metadata in the description of educational resources. The scope includes educational resources applicable for many national education communities and cross-sectoral communities (e.g., K-12, further and higher education and lifelong learning). Jon Mason of Education Network Australia (EdNA) and Stuart Sutton of the Gateway to Educational Materials (GEM) projects were named co-chairs of the WG. DC-Ed WG discussions began immediately on the listserv in preparation for the birds-of-a-feather (BOF) meeting in Frankfurt in October 1999.
At the Frankfurt meeting, the DC-Ed WG BOF delineated five information areas of particular interest to existing DC-based educational metadata projects:
- Users:
- Grade, age, academic/vocational/training level
- Administrators of the resource
- Student audience (target group, learning context, beneficiary)
- Duration:
- Focus on use time (as opposed to technical duration)
- Learning Processes/Characteristics:
- Student groupings, teaching methods, mechanisms of assessment, learning prerequisites, interactivity type and level, material type from a didactic viewpoint, type of use in a scholastic milieu, difficulty, semantic density, etc.
- Standards:
- National and/or international curricula
- National and/or international content/process standards
- Domain specific standards and benchmarks (e.g., U.S. Departments of Labor & Defense training benchmarks, etc.)
- Quality:
- Unstructured assessments (e.g., third-party reviews/annotations)
- Structured assessments (assessment based on established evaluative criteria)
These five general categories (some more than others) were the principle drivers of subsequent listserv discussions as well as in the development of strawman proposals going into the WGs face-to-face meeting at Kattemingga Lodge outside Melbourne on February 19-20, 2000. The DC-Ed WG discussion summaries and strawman proposals going into the face-to-face meeting in Melbourne can be found at http://www.ischool.washington.edu/sasutton/dc-ed/disc_summ.html.
The minutes of the Melbourne face-to-face meeting can be found at http://www.ischool.washington.edu/sasutton/DC7_DCEd.html.
These minutes were distributed to the face-to-face meeting participants for review and revision from March 1, 2000 through March 7, 2000. On March 7, 2000, the amended minutes were distributed for comment to the full DCEd WG on the DC-Education listserv. On April 15, 2000, the minutes were distributed to DC-General. With this document, the DC-Education WG completes its preliminary charge by submitting its proposal to the DCAC.In addition to memorializing the DCEd WGs proposal to the DCAC, this document outlines both existing issues regarding which the WG could not bring closure, and areas where it has defined needs for future work.
Jon Mason
Stuart Sutton
30 April 2000
Criteria for Evaluating Element and Qualifier Proposals
In assessing each of the five categories of educational information that were defined in Frankfurt and discussed to varying degrees on the DC-Education listserv, the following questions were used at the face-to-face meeting to guide assessment of each category:
Can it be clearly described? Can the semantics of the proposed element or element qualifier be expressed precisely, unambiguously, and briefly?
Is there a clear requirement for it in support of resource discovery in the education domain? Is there a demonstrated need for the proposed element, element qualifier, or value qualifier?
Does it support interoperability? Does it, to the maximum extent possible, support interoperability.
Is it practical? How difficult would it be for people creating metadata to comprehend the semantics of the proposed element or element qualifier and to apply it reasonably in the description of resources.
Does it refine an existing element? If it is a proposal for a new element, can it rationally be handled as effectively as an element or value qualifier for an existing element?
Are there alternative ways of implementing it? Within the conceptual framework of the Dublin Core™ Element Set (i.e., element/element qualifiers and value/value qualifiers), are there alternative ways to achieve the ends sought?
Are there existing implementations or controlled vocabularies, etc., supporting it? Somewhat akin to number 2 above, are there existing implementations for which this solution (element or element qualifier or value qualifier) is needed in support of resource discovery. In similar fashion, are there existing value qualifiers (i.e., controlled vocabularies, thesauri, etc.) that support it.
General Classes of Metadata Elements Considered
In the work at Kattemingga, it was important for the participants to distinguish among the following classes of metadata
elements (in descending order of interoperability):
- DC-Interoperability Elements/Qualifiers: Entities that are judged to be widely useful
across disciplinesi.e., they are Core entities.
- DC-Domain Elements/Qualifiers: Entities that are judged to be useful within a domain,
but perhaps not across domains.
- Common, But Local Elements/Qualifiers: Entities that are intended to be useful for local
applications or in a constrained federation of applications, but, perhaps, not widely used even in a given domain.
Proposed Audience Element:
It was unanimously agreed among members of the Working Group at the face-to-face meeting that the capacity to designate various aspects of the intended users of an educational resource being described is an important function for networked information discovery and retrieval. Frequently, creators and publishers of resources explicitly state the type (class, category, target) of user for whom the resource is intended (designed, appropriate).
In the original discussions on the DC-Education list, it was generally agreed that there are two classes of audience necessary to the education/training domain. These two classes are related and include, in the first instance, those persons, organizations, and other forms of entities that administor or mediate access to the resource by the second class of audiencethe end users for whose benefit the resource being described was designedi.e., the ultimate beneficiaries.
At the face-to-face meeting, there was considerable discussion both in the breakout group looking at the audience proposal and in the subsequent full group session regarding the need for the qualifiers put forward in the original strawman proposali.e., the strawman intermediary and beneficiary qualifiers. There was some diversity of opinion with how best to proceed with qualifiers. As the table below denotes, it was generally (although not unanimously) agreed that there were instances where refinement of the proposed audience element is necessary. However, the breakout group and the generally agreed upon draft proposal changed the name of the strawman qualifier intermediary to mediator and dropped the proposed beneficiary qualifier in the strawman proposal brought to Kattemingga.
Two different implementations achieving the goal were identified and are set out below. Both implementations share the same justifications.
Note: It was noted by the face-to-face participants that a general (unqualified) audience element might be useful in domains beyond education.
Recommendation #1 for an Audience Element
(Element & Element Qualifier)
ELEMENT
NAMEELEMENT
QUALIFIER
NAMEVALUE QUALIFIER
(CONTROLLED VOCABULARIES)DEFINITION
AUDIENCE
A category of user for whom the resource is intended.
category, class, type, target
intended, designed, appropriateMEDIATOR
An entity that mediates access to the resource.
ELEMENT:
- Name: AUDIENCE
- Definition: A category of user for whom the resource is intended.
ELEMENT QUALIFIER:
- Name: MEDIATOR
- Definition: An entity that mediates access to the resource.
VALUE QUALIFIER(S):
Recommendation #2 for an Audience Element
(Element & Value Qualifier(s))
ELEMENT:
- Name: AUDIENCE
- Definition: A category of user for whom the resource is intended.
ELEMENT QUALIFIER(S):
- This second implementation option by design avoids any element qualifiers.
VALUE QUALIFIER(S):
- Name: None defined at the face-to-face meeting
- Definition: One or more controlled vocabularies to express the members of the two categories of audience identified (e.g., DCEd-Med (for a Mediator vocabulary) and DCEd-Ben for an ultimate beneficiary vocabulary)).
- Comment: While the Working Group currently does not propose specific value qualifiers, it recognizes the need for their immediate development. However, it recognizes that examples of both forms of vocabulary are currently in use in existing systems and standards (e.g., see IEEE1484.12 & GEM, etc.).
Justification for Audience Recommendation
Yes. A category of user for whom the resource is intended.
Note #1: The category concept may be variably expressed as class, type, or target.
Note #2: The intended language may be variably expressed as designed or appropriate. However, both intended and designed imply affirmative acts and judgments by the creator of the resource; while the term appropriate carries no such limiting connotation.
Can the proposed element qualifier (mediator) be clearly described? Yes. An entity that mediates access to the resource.- Is there a clear requirement for it in support of resource discovery in the education domain?
Yes. Web-based searches for educational resources are often framed in terms of the particular people for whom those resources are intended; e.g., research has shown that teachers frequently search for resources by grade level (research performed by the Gateway to Educational Materials project (GEM) based on content analysis of the AskERIC archives and survey research of end-users).
Is there a clear requirement for the element qualifier (mediator)? Yes. The element refinement is required for many resources in order to distinguish between the primary audience (i.e., the ultimate beneficiaryusually a student or trainee) and a secondary audience that mediates access to the resource by the primary audience (i.e., usually a teacher or trainer).
Example Searches:
- A school library media specialist needs a lesson plan to teach information literacy to a 4th grade class that is learning about the Lewis and Clark Expedition in United States History.
- An adult education literacy teacher is looking for resources to teach English to Spanish-speaking immigrants to the United States.
- A 10th grade teacher of gifted students is looking for a resource to challenge one of her students who is interested in robotic engineering.- Does it support interoperability?
Yes. However, in order to promote interoperability widely, it will be useful to identify a common vocabulary of audience descriptors. The basis for such a vocabulary (or vocabularies) may be derived through analysis of: (1) existing vocabularies in current use in education-based applications; (2) existing education-specific audience descriptors in thesauri (e.g., the Thesaurus of ERIC Descriptors); and (3) in the emerging IEEE Learning Object Metadata standard specification. The definition of such a vocabulary (or vocabularies) would be an appropriate task of the DC-Education Working Group.
- Is it practical?
Yes. Current experience by existing projects in capturing audience information demonstrates that it is not difficult. This is particularly the case where the resource being described has been expressly intended or designed for a particular audience.
- Does it refine an existing element?
No. There have been discussions elsewhere (e.g., the Description Working Group) regarding audience as an element qualifier for one of the existing DCES. It appears that a general consensus exists that audience should not qualify an existing element.
- Are there alternative ways of implementing it?
No. Working Group members at the face-to-face meeting could think of no ways of achieving the goal except via a new audience element.- Are there existing implementations or controlled vocabularies, etc., supporting it
Yes. Education Network Australia (EdNA) is in the process of implementing its vocabulary; GEM has an existing vocabulary as well as a repository of resources cataloged using audience characterisics; the IEEE LOM contains elements reaching some aspects of audience (which means that the GESTALT, ARIADNE, and IMS projects all provide for audience information); and, various education-specific thesauri contain audience classification descriptors (e.g., the Thesaurus of ERIC Descriptors and the NICEM Thesaurus, to name but a few).**
| Proposed Standards Element: |
It was unanimously agreed among members of the Working Group at the
face-to-face meeting that the capacity to associate the educational resource
being described with organizational, professional, province/state, national,
and international content and process standards is an important function for
networked information discovery and retrieval.
Two different implementations achieving the goal (to varying degrees) were
identified and are set out below: (1) definition of a new education-specific
Standard element, and (2) use of the DC.Relation element with a new element
qualifier (e.g., ConformsTo (or some other token capturing the notion)).
Both implementations share the same justifications.
Recommendation #1 for
a Standard Element
(Element & Element
Qualifiers)
ELEMENT:
- Name: STANDARD
- Definition: A reference to the education
or training standard with which the resource is associated.
ELEMENT QUALIFIER:
- Name: IDENTIFIER
- Definition: Where available, an identifier
(or notation) that serves to uniquely identify the standard being
associated.
- Comment: With increasing frequency,
content and process standards are uniquely identified by means of
structured notation (e.g., in the manner or Dewey Decimal class numbers)
that can support machine processing.
ELEMENT QUALIFIER:
- Name: VERSION
- Definition: Information identifying the
version of the standard being referenced (e.g., a year of publication, a
version number, etc.).
- Comment: Frequently, unique
identification of a content or process standard is a function of
both an identifier and a version. While an identifier within a
given standard scheme may remain constant across versions of the standard,
its content may vary from version to version (i.e., the identifier is
rendered non-unique); therefore, unique identification under such
circumstances may require both an identifier and a version.
VALUE QUALIFIER(S):
- Standards information would be drawn from publicly recognized and
publicly available schemes promulgated by recognized organizational,
professional, academic, vocational, and governmental
bodies.
Recommendation for Standards
#2
(DC.Relation Element Qualifier)
The second proposal relies on the use an element qualifier for the Dublin
Core RELATION element.
-
Definition: The URI of the standard being
associated at whatever level of granularity is required (i.e., standard, benchmark, etc.).<br> </li> -
Comment: A pointer to the external
location of the standard as maintained by the standards-promugating body. However, the reference must point to the exact level of the standard being referenced and not merely to the standards as a whole.<br><br> - The solution needs a mechanism for the relation element to
unambiguously point to discrete locations deeply embedded in a standards document (a pointer to a textual representation is simple; a pointer to a digitally available instantiation is not so simple). <p></p> </li> - In the long term, unique identifiers will solve the problem of needing
to execute a search directly using a reference to a standard. But that is not possible now. <p></p> </li> - Human users will want to search for resources associated with specific
standards. But in the future, machines will be collecting the information for further synthesis. But this is not close enough to be useful for us now. <p></p> - A math teacher is looking for resources that conform to the National
Council of the Teachers of Math (NCTM) standards for teaching 8<sup>th</sup> grade algebra. </li> - A parent engaged in home teaching is looking for resources that meet
a specific reading comprehension standard in his or her state or province. </li> - An aggregator of educational content is searching for resources
associated with a specific content standard. </li><li>Does it support interoperability? <blockquote> <p>Yes. With registries/repositories for content and process standards, varying degrees of interoperability can be achieved. For example, in the United States, an organization named Achieve (formed through the efforts of U.S. state governors) has cross-mapped content standards for over forty states in order to achieve a level of interoperability among resources associated with those standards.</p><li>Is it practical? <blockquote> <p>Yes. Many educational departments at all levels of government wanting to support standards within their jurisdictions will mandate the association of resources to their standards. Many commercial providers and other content creators wanting to make their resources easy to find it, will associate those resources with one (or more) recognized standards.</p><li>Does it refine an existing element? <blockquote> <p>Possibly. As noted above, one of the possible implementations involves a refinement of the existing RELATION element. However, it is not clear that the use of the RELATION element can achieve the goal at this time.</p><li>Are there alternative ways of implementing it? <blockquote> <p>Possibly. As noted above, one of the possible implementations involves a refinement of the existing RELATION element. However, it is not clear that the use of the RELATION element can achieve the goal at this time.</p><li>Are there existing implementations or controlled vocabularies, etc., supporting it? <blockquote> <p>Yes. The standards to be represented exist as formal schemes developed by recognized promulgating bodies–e.g., the US NCTM (National Council of Teachers of Mathematics); the UKs NVQ (National Vocational Qualification); and the Australian Qualifications Framework. Metadata repositories (e.g., GEM) are using an element to capture standards information. </p> - Identify suitable sources of funding.
- Accurately scoping user requirements with regard to the activities and
methods associated with teaching and learning; and </li> - Obtaining real feedback, based upon empirical data where possible, in
terms of existing usage and user requirements.
<p><strong>ELEMENT:</strong></p>
</ol>
<li>
Name: RELATION (already
exists)
<p><strong>ELEMENT QUALIFIER:</strong></p>
<ul>
<li>
Name: ConformsTo (or some other name that
captures the notion)<br>
</li>
<p><strong>VALUE QUALIFIER:</strong></p>
<ul>
<li>None<br>
#2
While proposal #2 comports well with the One-to-One rule, its elegance
masks possible short-term difficulties.
<p>
</p>
Recommendations
<li>Can it be clearly described
<blockquote>
<p>Yes. A reference to the education or training standard with
which the resource is associated.</p>
<p><u>Note</u>: Homonym ProblemThe term standard has multiple meanings
potentially leading to confusion; however, Working Group discussions led
to no better term to express the concept.</p>
<li>Is there a clear requirement for it in support of resource discovery
in the education domain
<blockquote>
<p>Yes. Increasingly, educational resources are either being
designed to meet the needs of specific content and process standards, or
are being associated with such standards after creation. Being able to
search for resources that meet the goals of specific standards is growing
in importance as more educational processes become standards-based.</p>
<p><u>Example Searches</u>:</p>
<ul>
<li>A 4<sup>th</sup> grade teacher in Texas is looking for resources
that support the Texas Essential Skills for 4<sup>th</sup> grade
science.
</li>
It is generally agreed among members of the Working Group at Frankfurt that the capacity to designate various kinds of activities of students, and various kinds of methods used by teachers/trainers in the teaching/learning process is important for networked information discovery and retrieval. However, as chronicled in the section below on open issues, the Working Group members at the face-to-face meeting were unable to resolve issues regarding this category resulting in the decision to make no general recommentations at this time regarding the creation of either a new education-focused element (and possible element qualifiers for such an element ) or element qualifiers for the existing DCES.
However, in conformance with the DCMI commitment to modular metadata, the Working Group members at the face-to-face meeting unanimously recommend the use of the InteractivityType and the InteractivityLevel data elements (Education Category: 5.1 & 5.3) from the IEEE Learning Object Metadata (LOM) draft standard for P1484.12. These two data elements partially fulfill needs identified under the Teaching Processes/Characteristics category at Frankfurt. Representatives from DCMI and IMS/IEEE (IMS is as an instantiation of the IEEE LOM and was represented at the face-to-face meeting) will identify an appropriate technical means to enable referencing the namespace. It was recognized that mixing of namespaces raises important issues of interoperability among metadata registriesissues that will be explored by DCMI and IMS/IEEE.
The table below represents the IEEE P1484.12 data elements 5.1 & 5.3 as set forth in the recently released (5 February 2000) Draft 4.0.
It is generally agreed among members of the Working Group at the face-to-face meeting that the capacity to designate the use time for the resource being described is an important attribute in networked information discovery and retrieval. This category of information was called duration at the Frankfurt meeting. The concept of duration is distinct from the physical size of the resource or its potential technical run time (where such exists). Values for this category represent the approximate or typical time it takes to work with the resource.
In conformance with the DCMI commitment to modular metadata, the Working Group members at the face-to-face meeting unanimously recommend the use of the TypicalLearningTime data element (Education Category: 5.9) from the IEEE Learning Object Metadata (LOM) draft standard for P1484.12 to represent the DC duration concept. Representatives from DCMI and IMS/IEEE (IMS is as an instantiation of the IEEE LOM and was represented at the face-to-face meeting) will identify an appropriate technical means to enable referencing the namespace. It was recognized that mixing of namespaces raises important issues of interoperability among metadata registriesissues that will be explored by DCMI and IMS/IEEE.
The table below represents the IEEE P1484.125.9 as set forth in the recently released (5 February 2000) Draft 4.0.
Through the work at Frankfurt, it was generally agreed among members of the Working Group that the capacity to designate various aspects of the quality of the educational resource being described is an important function for networked information discovery and retrieval for a number of existing projects. However, it was subsequently agreed at the face-to-face meeting in Australia that no formal recommendation regarding the quality category should be put forward at this time. It was suggested that education-related metadata projects needing to express either a quality assessment within the descriptive metadata for a resource or to establish a pointer to such an assessment external to the descriptive metadata, should do so by means of a local element or a local qualifier of the RELATION elementeither alternative being made available to the Dublin Core™ education community through the emerging registry.
There was general agreement at Kattemingga that the proposed DC.Type vocabulary (DCT1just balloted) is inadequate to meet the need for domain specificity while still maintaining the maximum degree of interoperability. It was proposed that an interested group of participants collate an extensive list of education-focused resource types either currently in use or under development. Subsequent efforts will destill an acceptable minimum set of potential interoperability values from the larger listing. Representatives from a group of projects subsequently collated a listing of education resource types and made the listing available through the DC-Education listserv. Discussions of the collated list are ongoing.
Discussion within both the breakout session and the whole Working Group at
Kattemingga on Learning & Teaching Processes/Characteristics was
lively and began with a classification into two broad classes –
activities and methods. Of these two broad classifiers methods
was clearly identified as the most contentious area for debate. However, the
practical constraints of the agreed evaluation criteria resulted in the
Working Group not pursuing developing a recommendation for a DC-Education
metadata element or interoperability qualifier at this stage. While there
appears to be a strong requirement from users that capturing and accommodating
educational metadata is necessary, the lack of any universally accepted
taxonomies in this area is the major issue. Also very much at issue – and a
topic which attracted a fair amount of debate – is where the boundary between
resources on the one hand, and the processes which might operate
on them on the other, is most meaningfully drawn. Such a question was
considered to be more than a philosophical question where digital objects
purposed for educational activity is concerned.
At the Frankfurt meeting the actual category identified was summarised as
Learning Processes/Characteristics. Since that time, discussion on
the DC-Education listserv and at Kattemingga was extended to be more inclusive of
teaching, with pedagogy receiving most attention. Thus, members of the
Working Group agreed that educational practice is more than the
assembly or packaging of resources. Moreover, any assembly and packaging of
resources if describable is potentially a valuable resource in itself for
discovery which may provide a better understanding of learning and teaching
practice. Related to this issue a further question was raised as to whether
there is a privileging of learning objects in currently implemented schema
whereby notions of education tend to reduce to resource-based
learning.
The bottom line group consensus was that Learning & Teaching
Processes / Characteristics
value statements and many overlapping vocabularies based on ideology or local
frameworks.
The Working Group agreed that in order to make further progress on this
issue, a research proposal be suitably framed.
In Frankfurt, October 1999, the DC-Education Working Group identified
Learning Processes/Characteristics as one of five categories which
could classify the kinds of approaches that various educational metadata
projects already implemented had pursued. The five categories were based upon
two independent analyses and first proposed by Stuart Sutton. The categories
were proposed as a representative and broadly inclusive rough cut
classification for educational metadata.
In Kattemingga, the category was broadened to Learning and Teaching
Processes / Characteristics
largely taken place on the mailbase list where pedagogy had occupied most
attention.
However, based on the evaluation criteria for determining interoperability,
members of the Working Group present at the Kattemingga workshop were unable
to make a recommendation concerning possible new elements and qualifiers which
could be seriously considered by the wider DC-Education community or the
DC-Advisory Committee. Nonetheless, it was unanimous that developing a
well-framed research proposal may be a practical strategy to take the issue(s)
forward.
Thus, the Working Group agreed to:
<li>Define suitable research project(s) to remove impediments to useful
implementations of metadata which does or could capture <em>Learning and
Teaching Processes / Characteristics</em>; and
</li>
In framing the project, the following actions from members of the
DC-Education Working Group were considered to be important:
<li>Identifying and/or recommending appropriate controlled vocabularies;
</li>
Part of the discussions on the DC-Education listserv since the face-to-face meeting in Melbourne have focused on the need to identify the level of the students/trainees for which the resource being described is intended. This notion of level might be expressed in a number of ways depending on the context but appears to include information such as grade level, course level, and perhaps age group. In listserv discussions prior to the Melbourne meeting, this aspect of a resource was part of the proposed Audience element. See, for example, the discussion summary (Category: Audience) leading into face-to-face meeting.
While there appears to be general agreement regarding the need to capture this aspect of a resource for effective networked information discovery and retrieval, questions have arisen regarding whether the information appropriately belongs as part of the proposed Audience element (perhaps as an element qualifier) or as a separate DC-Education element.
Since the time-frame was too short for the DC-Education Working Group to discuss and resolve the issues before this submission to the Advisory Committee for review, the general agreement regarding a need to more precisely identify this aspect of an educational resource demands that discussions be ongoing. The Working Group will return to the Advisory Committee with a proposal at a later date.
Rei Atarashi, Katrina Beard, Alan Bevan, Debbie Burgoyne, Warwick Cathro, Denise Curran, Mike Currie, Jack Gilding, Renato Iannella, Margaret Kiley, Jon Mason, Carol McKenny, Stuart Mitchell, Nancy Morgan, Liddy Nevile, Helen Page, Bruce Rigby, Ian W. Roberts, Karen Rollitt, Brendon Ryniker, Kathrin Strati, Stuart Sutton, Frank Tansey, Teanau Tuiono, Heather Watson, Stu Weibel, Robyn White, Jill Wilson