XML Metadata Virtually Making Sense of Our Records of Reality Owen Ambur, Co-Chair XML Community of Practice (xmlCoP) NCC-AIIM Educational Seminar Rosslyn, Virginia November 4, 2004 Data v. Reality Dreaming v. Acting "An information system (e.g., database) is a model of a small, finite subset of the real world... We expect certain correspondences between constructs in the information system and the real world." William Kent Data and Reality (p.1) Records v. "Nonrecords" Databases v. E-records Authenticity Reliability Integrity Usability Databases Are Not Good Record-Keeping Systems Tuples Lack Integrity Every DB Has Administrator Insiders Have Motives Privacy & Security Not Primary Manipulation of Data Is Question: If the IT system on which you are spending the taxpayers' money does not create "records," which of those four attributes do you intend to ignore? Data v. Metadata Context is Key Metadata = Data About Data Data Without Context Is Meaningless, Or Worse Records Require Context e.g, "Four Corners" of the Document Separation of Presentation from Content = Unreliable Records Lacking Integrity = Invitation to Waste/Fraud/Abuse Metadata Provides Context for E-records Presentation Is Critical ... for Important Records Internal v. "External" Metadata In Other Contexts, Metadata = Data Parsed (copied) into Databases Indices, Queries, Analyses, Manipulation Having Our Cake & Eating It Too Almost Heaven ... Extensible Markup Language! Open Standard Bridges Gap Between Documents (Records) & Data E-records Data/Metadata Readily Shared/Reused in Other Contexts Automated Auditing of Databases Myriad Generalized and Specialized Indices Easy Linkage/Extension of Internal/External Metadata Integrity of Original E-Record Remains Inviolate Metadata = Management Information Technology & Electronic Records In the Context of IT Systems and E-records, the Terms "Metatada" and "Management" are Virtually Synonymous With the Proper Metadata, Nearly Anything that May Need to be Done with or to Any E-record Can Be Automated Little or No Metadata = Little or No Management The Question Is: How Much "Management" (i.e., Metadata) Any Particular Record or Class (Series) of Records May Warrant before Reaching the Point of Diminishing Returns Costs/Benefits May Accrue to Different Individuals/Organizations Metadata & Human Behavior Wishes & Dreams v. Reality Commonly Held & Oft-Proven Belief People Won't Provide Metadata If Takes More than a Few Moments Unsuprising Since May Be No Immediate Benefit to Them Solution Measure Performance Against Metadata Standard Make Results Salient Leave It to Stakeholders Deal with Results E-Records Metadata ISO 15489 Attributes of a Record Integrity Reliability Usability Authenticity Attributes = Metadata Specify E-Records Metadata Standard Default as Much as Possible from System Parameters Autoclassify E-records & Autogenerate Metatada Measure & Report Human Performance Relevant Efforts E-Records Metadata eGov eRecords Initiative http://www.whitehouse.gov/omb/egov/pres_init/internal/records.htm NARA's Electronic Records Archives (ERA) Project http://www.archives.gov/electronic_records_archives/whats_new.html Metadata Encoding & Transmission Standard (METS) http://www.loc.gov/standards/mets/ Interagency Committee on Government Information (ICGI) http://www.cio.gov/documents/ICGI.html Categorization of Information Work Group Web Content Management Work Group Web Content Management Standards Group Integrated Enterprise Content Management (IECM) Standard Mike Connor, Adobe & Paul Fontaine, FAA So What Else Is New? What the World Still Needs Questions, Comments, Suggestions? Contact Information Owen Ambur, Co-Chair XML Community of Practice (xmlCoP) http://xml.gov/ Owen_Ambur@ios.doi.gov