UW Faculty Accomplishments Project
Recommendations Document
Prepared by
The UW iSchool Faculty Accomplishments Team
Bob Boiko Project Director
Suzanne Soroczak, Technical Lead
Aaron Louie, Information Lead
Table of Contents
1 Introduction to the Faculty Accomplishments Project 1
1.1 Goals 1
1.2 Intended Audiences 1
1.3 Definitions 2
1.3.1 Accomplishment 2
1.3.2 Faculty 2
1.3.3 School 2
1.3.4 Vocabulary Concept 2
2 Project Deliverables 2
2.1 Scope 2
2.2 Staffing 3
3 Methodologies and Key Assumptions 4
3.1 Key Concepts 4
3.1.1 VAM Functionality 5
3.2 Emergence 6
3.3 What is an Accomplishment? 7
3.4 Managed Vocabulary 7
4 User Analysis Results 10
4.1 Faculty behavior: Logging Accomplishments 10
4.2 Summary of Results 10
4.2.1 Sort order 10
4.2.2 Categories 10
4.2.3 Labels 11
4.2.4 Other findings 11
4.3 Design Recommendations 11
4.3.1 Faculty 12
4.3.2 Administration 13
4.3.3 Assistants 14
4.3.4 System Analysis 15
4.4 4.4 Scenarios of Use 19
4.4.1 Faculty member 19
4.4.2 Executive 20
4.5 User Interface Recommendations 21
5 Vocabulary Management Results 22
5.1 Vocabulary Discovery Toolkit 22
5.1.1 Document Cracker 22
5.1.2 Part of Speech Tagger 22
5.1.3 Phrase Extractor 23
5.2 When is the Managed Vocabulary Successful? 23
5.3 Key Questions 23
5.4 Vocabulary Management Experiment 24
5.4.1 User Analysis 24
6 Data Model Recommendations 26
6.1 User Data Model 26
6.1.1 unique faculty id 26
6.1.2 faculty name 26
6.1.3 title 26
6.1.4 school/department 26
6.1.5 email address 26
6.1.6 current list of accomplishments 27
6.2 Accomplishment Data Model 27
6.3 Accomplishment Metadata Elements 27
6.3.1 Open Elements 27
6.3.2 Controlled Elements 27
6.3.3 Open Lists 27
6.3.4 Managed Terms 28
7 Required Functionality 29
7.1 CV/Resume import 29
7.2 Web-based Input 30
7.2.1 Cut & Paste 30
7.2.2 Spell checker 30
7.2.3 String normalization 30
7.2.4 Multi-Item add 30
7.3 Suggest A Topic 30
7.3.1 Text tokenizer 31
7.3.2 Part-of-Speech (POS) tagger 31
7.3.3 Noun Phrase Extractor 31
7.4 Reporting Data 31
7.4.1 Aggregate Reports 31
7.4.2 Ego-centric Reports 32
7.5 Search 32
8 Recommended Functionality 32
8.1 CatGrid 32
9 References 33
ii
1 Introduction to the Faculty Accomplishments Project
The faculty accomplishments project is sponsored by the UW Provost’s Office, the School of Nursing, and the Information School to make recommendations to implement a university-wide faculty accomplishments system (FAS) for tracking the achievements of the UW faculty. The system would be used by the faculty at the UW to record their accomplishments over the course of their employment with the university. Deans and university administration would make use of this information to prepare reports on departmental and university activities and advertise the accomplishments of the faculty. Other uses for a system of this type would be to reduce accomplishment data redundancy, facilitate finding expertise on campus, and supporting the faculty tenure process.
1.1 Goals
Defining and describing what an accomplishment is. Our system must somehow instantiate a concept of accomplishment that is wide enough to encompass the whole faculty, yet confined enough to be comparable across disciplines. The system must show how such accomplishments can be created and later accessed.
Acknowledging, rewarding, and promoting accomplishments. The system we propose must define and instantiate these terms and propose an interface for accomplishing them.
Categorizing accomplishments in ways that are meaningful to the users of the system. The proposed system must show how the users of the system can find the types of information they most want from the accomplishments that the system contains. For example, users might want to see accomplishments in the light of the expertise or interests they imply.
Enabling maximal high quality submissions from faculty. The system must include feasible and reasonable means for encouraging faculty to log their accomplishments in a high-quality way and categorize them to the extent necessary for the later use of these accomplishments.
1.2 Intended Audiences
Deans and department chairs. To begin, we will assume the main need of this audience is to track, understand, and reward the accomplishments of their faculty. To succeed, the system must be useful and usable by this audience.
Support staff. These people include the administrative staff of academic units and the university as a whole. We assume that their primary goal will be quickly finding faculty accomplishments that merit attention and easily creating the press releases or other outputs that describe these accomplishments. We will decide the exact list of positions and outputs that we will recognize and serve in this version of the system when the project begins.
Faculty. In addition to their key position as the providers of input, faculty members are an important audience for the output of the system. To be successful, the system must allow faculty to manage their own accomplishments and to find the accomplishments of others.
1.3 Definitions
1.3.1 Accomplishment
A completed or ongoing activity engaged in by a faculty member which has a desired impact in the world, whether inside or outside of the university context. In this document the terms accomplishment and achievement will be used interchangeably.
1.3.2 Faculty
A faculty member will be defined as university employees with the following titles: dean, professor, associate professor, assistant professor, adjunct professor, lecturer, etc…
1.3.3 School
A self-governing administrative unit on campus.
1.3.4 Vocabulary Concept
A free-text representation of an idea often best captured in a noun-phrase format. May be referred to as a vocabulary node.
2 Project Deliverables
Below are the major tangible outputs that we intend to produce in this project. These outputs will be created to encompass the information that we gather from both the School of Nursing and the iSchool and will include, as much as is feasible, our recommendations for creating a system that encompasses the entire university.
§ A full prototype of the system as designed. The prototype will give the impression of moving fully through the system from the perspective of the primary audiences. It will not perform the tasks that it illustrates.
§ Working models of key concepts. We will create working code for a number of the parts of the project that cannot be tested in a non working form. For example, we may need to create working data entry screens to test the ides of emergence and active review. For these parts, we will deliver the code and user interface as separate applications or embed them in the final prototype. The models can be modified and used in the later implementation of the system.
§ Implementation considerations document. We will create a document that can contribute directly to the specification of the system to be implemented. In this document, we will collect the relevant research we have done, include the specifications we have created for the system, and provide a stock of advice and implementation directives for those who will implement the system.
2.1 Scope
The domain of information for this system is faculty accomplishments. We intend the definition of accomplishment to emerge from the kinds of accomplishments that faculty log. Thus, we must maintain a somewhat open definition of the information that the system will handle. Still, some constraints are evident at the start of the project:
§ All the content the system contains will be digital.
§ The system will not duplicate information held in an accessible form elsewhere in the university.
§ The system will contain information that describes an accomplishment (name, type, date, and so on) and will contain information that categorizes accomplishments (keywords, department of origin, and so on).
§ It is possible for the definition of an accomplishment to emerge organically from the ways that faculty themselves describe their activities.
§ There will still be a need for someone to actively review, refine, and synthesize the raw discriminations made by faculty in order for the definition of an accomplishment to become sharp and span disciplines.
§ The system will be "the one place" for faculty to manage their accomplishments. As such, it will have an established and supported place in the weekly or quarterly workflow of faculty.
§ The system will have to balance between the faculty's rights to direct how their information is distributed and the needs of the other audiences to know and use information about faculty accomplishments. In any case, the policy and rules governing use of the information must be clearly stated.
§ To be useful, the system must "look right" from the perspective of each discipline that uses it.
§ We are aiming for maximum automation to deal with the accomplishments that have already been logged somewhere (in resumes and so on) and a minimum of data entry on the part of faculty to enter new accomplishments.
§ The project will be run under the auspices of the newly formed Content Management Evaluation Lab which will begin operations in the Information School in the Fall Quarter of 2002.
§ The project will address two academic units: the School of Nursing, and the iSchool but will be designed for scalability to a larger base of units and eventually to the entire university.
§
2.2 Staffing
We will work through a complete task and staff structure for the project as we begin the planning phase. Initially, we assume that to accomplish this project we use the following staff:
§ Bob Boiko, iSchool Lecturer, will serve as the project director. He will assemble the team and oversee its activities.
§ Suzanne Soroczak, a funded half-time Information School PhD-level graduate assistant, will serve as project manager, tracking the scope, schedule, and budget of the project. She will also serve as technical lead, performing or overseeing all of the software development involved in the project.
§ Aaron Louie, a funded half-time masters-level graduate assistant, will serve as information lead. He will perform or oversee the knowledge organization and negotiation involved in the project.
§ Other students will participate in the project on a volunteer or hourly basis as their skills and the needs of the project dictate.
3 Methodologies and Key Assumptions
3.1 Key Concepts
In addition to the concepts of LUGs, managed vocabulary, and a VM defined earlier, here are some other key concepts of the managed vocabulary system.
Concept / UseVocabulary / Our notion of a vocabulary is a set of phrases that taken together completely describe the subject matter (in this case accomplishments) from the perspective of our various LUGs
We will work toward a single vocabulary.
Phrase / One or more words that together have some specific meaning. A phrase ought to be understandable to one or more LUGs without further explanation.
Suggestions / We intend that any phrase, definition, or relationship be modifiable by users as they create and search for accomplishments.
We call this class of proposed changes suggestions. They suggest changes to the vocabulary that the VM can review and act upon. We will instrument the suggestion system such that suggesters can be more or less directly enabled to have their suggestions enacted.
Managements / The system will have a set of functions to solicit and receive suggestions, and evolve the vocabulary based on them. We will call the set of these features managements.
The system must allow the right set of vocabulary managements to best facilitate the movement toward a complete and coherent vocabulary.
Surveys / We will have a set of solicitation tools that allow the VM to easily canvass suggestions from the appropriate individuals in the appropriate LUGs.
Centrality index / For each phrase in use by each LUG, we will calculate a number that best approximates our understanding of how important that phrase is to that group. This number will be use to create displays of phrases where the most important are the most prominent.
Acknowledging that no calculation can fully capture importance, we will allow the VM to manually offset the calculated value as needed.
Phrase Subscriber / If a user has ever made a suggestion on a phrase, she is a subscriber. As the phrase is worked upon we will include these subscribers in the workflow.
Phrase User / If a user has ever chosen a phrase, she is a user. As the phrase is worked upon we will include these subscribers in the workflow.
3.1.1 VAM Functionality
3.1.1.1 Add Node
Allows for the addition of unique node ids. Normalized input concept for sentence case and punctuation. Tests spelling with a spell checker. Possibly uses stemming or other Natural Language Processing utility to test for duplicate entries in the system.
3.1.1.2 Add Facets
Each concept/node will contain a list of general facets including, but not limited to:
· title,
· description,
· parent or hierarchy information,
· language usage group,
· and other relations.
3.1.1.3 Delete Node
Allow for the deletion of the unique node id and associated facets from the system. Also deletes references to the deleted node.
3.1.1.4 Find Node
A full-text search for any node or facet. Includes a search for unique node ids.
3.1.1.5 Tree View
The ability to view the vocabulary in a collapsible, hierarchical tree view.
3.1.1.6 Administration
The administration function should tie into the Vocabulary Discovery Tool which is used to managed the growth of a vocabulary through an interactive question-and-answer process. The administration functionality should allow part a of the vocabulary tree to be submitted to the question-and-answer process for vocabulary refinement.
3.1.1.7 Statistics/Utility
3.1.1.7.1 Node Count
Count the number of nodes in the vocabulary.
3.1.1.7.2 Check for circular references.
Check the parent-child relationships for circular references in the vocabulary. A circular reference is one that refers to itself.
3.1.1.7.3 Text Reporting
The ability to print our a text list of the current vocabulary.
3.1.1.7.4 Export
The ability to export the vocabulary in another format, XML for example.
3.1.1.7.5 Import
The ability to import a text list of hierarchical vocabulary terms.
3.2 Emergence
According to the Institute for the Study of Coherence and Emergence (http://emergence.org/):