A Taxonomic Approach to theOrganization of Penn State Web

SpaceMichael Pelikan, Jin Ma, Margaret Smith, and James Leous

The Pennsylvania State University

A Taxonomic Approach to the Organization of Penn State Web Space – p. 1

Forming the Posse

� Formed by Dr. Russell Vaught, AssociateVice Provost for Information Technology as aresult of his Web report to the Provost

� His assignment:

� make finding specific pages easier

� simplify Web content management tasks

� come up with something durable

� Vaught held a series of one–on–onemeetings in 2003, pulled together smallgroup in 2004

A Taxonomic Approach to the Organization of Penn State Web Space – p. 2

The Team

� James Leous, Manager, Research Programmer,Information Technology Services

� Jin Ma, Metadata Librarian

� Sam Haldeman, Program Coordinator, InformationTechnology Services

� Richard Pearce, Director, Business and Finance,Auxiliary and Business Services

� Margaret Smith, Manager of Publications, InformationTechnology Services

� Michael Pelikan, Technology Initiatives Librarian

� Russell Vaught, Associate Vice Provost forInformation Technology A Taxonomic Approach to the Organization of Penn State Web Space – p. 3

The Penn State WebExtent of Penn State Web

� 24 campus locations outside of UniversityPark

� 12 colleges

� Graduate School

� Medical School

� Law School

� more than 1,000,000 public, non–personalweb pages

� “first stabs” at Content Management poppingup like mushrooms!

A Taxonomic Approach to the Organization of Penn State Web Space – p. 4

Initial FindingsGroup’s initial findings

� near–total lack of control on:

� page titles

� link labels

� standardized or authoritative form forunits, departments, etc.

A Taxonomic Approach to the Organization of Penn State Web Space – p. 5

Initial Findings, cont’dGroup’s initial findings (cont’d)

� increasing adoption of CMS systems acrossPenn State

� dozens? hundreds?

� ever–evolving search engines with differingtreatment of tags

� Need to be CMS and search engineagnostic!

A Taxonomic Approach to the Organization of Penn State Web Space – p. 6

Initial TargetsThe Group’s initial targets (2004):

� Rules to govern authorized form for name forany entity in the University

A Taxonomic Approach to the Organization of Penn State Web Space – p. 7

Initial TargetsThe Group’s initial targets (2004):

� Rules to govern authorized form for name forany entity in the University

� a taxonomic rendering of those terms, whereuseful

A Taxonomic Approach to the Organization of Penn State Web Space – p. 7

Initial TargetsThe Group’s initial targets (2004):

� Rules to govern authorized form for name forany entity in the University

� a taxonomic rendering of those terms, whereuseful

� a plan to derive benefit from such terms,even in the present environment

A Taxonomic Approach to the Organization of Penn State Web Space – p. 7

Initial TargetsThe Group’s initial targets (2004):

� Rules to govern authorized form for name forany entity in the University

� a taxonomic rendering of those terms, whereuseful

� a plan to derive benefit from such terms,even in the present environment

� a plan anticipating their broader applicationin foreseeable Web management systems

A Taxonomic Approach to the Organization of Penn State Web Space – p. 7

Initial TargetsThe Group’s initial targets (2004):

� Rules to govern authorized form for name forany entity in the University

� a taxonomic rendering of those terms, whereuseful

� a plan to derive benefit from such terms,even in the present environment

� a plan anticipating their broader applicationin foreseeable Web management systems

� a set of practices to ease their maintenanceover time

A Taxonomic Approach to the Organization of Penn State Web Space – p. 7

Tools c. 2004Potential implements to wield, at hand in 2004:

� Google Appliance

� LDAP

� Zope–based CMS effort

� concept of CMS as both a means and anopportunity to implement

A Taxonomic Approach to the Organization of Penn State Web Space – p. 8

Google Fly in the OintmentIn the first year of the “Tags” committee work,Penn State purchased a Google Appliance. Thismeant:

� The PageRank

� �

algorithm determinesquality of search

� Keywords/META tags were much lessimportant

Overall, our constituents were much “happier”with the quality of the searches, but we had aharder time with “fixing” pages.We suggested a “Sponsored Links” on theGoogle Appliance.

A Taxonomic Approach to the Organization of Penn State Web Space – p. 9

The Fix is OnWithout a “Sponsored Links” option the processof making the search find the “proper” academicdepartment is very manual. Could the UniversityDirectory (LDAP) help us?LDAP is populated by our CACTUS Accountsdatabase. The LDAP table comes from severalauthorities:

� Student Data from Enrollment Management

� Employee Data from Office of Human Resources

� Medical School/Center information from MSHMC

� Departmental Information from Call Center

A Taxonomic Approach to the Organization of Penn State Web Space – p. 10

Combined SearchCan we create a combined search page whichincorporates Google result set with a return fromLDAP?

� Such a page would satisfy our need forgetting the “right” answer if LDAPdepartmental entries were properlypopulated.

� The key is to have the proper synonyms forsearches

� Can EduOrg LDAP schema help us?

A Taxonomic Approach to the Organization of Penn State Web Space – p. 11

Controlling TerminologyPredictable content retrieval

� based upon data in elements or fields

� common or mappable elements necessarybut not sufficient by themselves!

Need for consistently applied terms

� controlled vocabulary

� authority control, if feasible

A Taxonomic Approach to the Organization of Penn State Web Space – p. 12

Formal Names, Legal Names

� Departments, Colleges, Locations, Units, etc.

� The So and So Institute of the Donor–NameCollege of Subject Area of the PennsylvaniaState University

� Non–intuitive unit names:

� General Stores (office supplies)

� Fleet Services (rental cars)

A Taxonomic Approach to the Organization of Penn State Web Space – p. 13

OwnershipWho “owns” the authorized terms?

� As of now, list is maintained by the Office ofPublications

� The Call Center provides a means to updatesome information (Consulting and SupportServices, a unit of Information TechnologyServices)

A Taxonomic Approach to the Organization of Penn State Web Space – p. 14

Formal TermsDistillation of complicated, formal terms

� must be done by hand

� must use consistent practices

� must result in unique terms

� or must represent context for non–uniqueterms

� meaning thesaural or taxonomicrepresentation is indispensable

� need for mechanism by which commonnames or “nicknames” resolve to authorizedterm

A Taxonomic Approach to the Organization of Penn State Web Space – p. 15

Tools c. 2005Implements to wield, at hand in 2005:

� The “ear” and interest of Google (??)

� Day Communiqué, licensed by Penn StateLibraries, but for University–wide use

� Library local LDAP implementation

� Engagement of Old Main as aUniversity–level strategic project

A Taxonomic Approach to the Organization of Penn State Web Space – p. 16

GoogleThe “ear” of Google

� Sponsored Links?

� Database connectors (including LDAP?)?

� Will they let us at the controls??

A Taxonomic Approach to the Organization of Penn State Web Space – p. 17

GoogleThe “ear” of Google

� Sponsored Links?

� Database connectors (including LDAP?)?

� Will they let us at the controls??

� “Naaaaaaah!”

A Taxonomic Approach to the Organization of Penn State Web Space – p. 17

Day Communiqué

� “Content Bus”

� LDAP connector

� Workflow modules

� Workflow elements are Content

� “Glossary”

� pre–populated check boxes

� capture locations, Organizational Units,etc.

� representation of hierarchical relation

� <dcterms:isPartOf>ISPARTOF</dcterms:isPartOf>

� <dcterms:hasPart>HASPART</dcterms:hasPart>

� metadata is ContentA Taxonomic Approach to the Organization of Penn State Web Space – p. 18

Old SchoolGood old fashioned Library work

� round up, wrangle and corral Unit,Department, College, Location names, etc.

� populate LDAP OUs with standardized,authorized form (OUs as DNs == main entry)

� populate LDAP “nickname” elements withaccess points resolving to authorized form

� similar authorized labels needed for page“function” (resource, transaction, etc.)

A Taxonomic Approach to the Organization of Penn State Web Space – p. 19

ProgressThe Taxonomic Tags Group is making substantialprogress!

� How can this be?

� small, interdisciplinary group

� a blank sheet of paper

� fortunate convergence of circumstancesand technology

� sheer, innocent audacity

A Taxonomic Approach to the Organization of Penn State Web Space – p. 20

Questions and Discussion

Questions and Discussion

A Taxonomic Approach to the Organization of Penn State Web Space – p. 21