Tagging: It’s the Interface Stupid!
Who I am • Over 25 years in the business of organized information • Founder & Principal, Taxonomy Strategies • Director, Solutions Architecture, Interwoven • VP, Infoware, Metacode Technologies • Program Manager, Getty Foundation • Manager, Pricewaterhouse • Assistant Director for Technical Services, Hampshire College • Chief, Technical Services, Paul Weiss Rifkind Wharton & Garrison • Metadata & taxonomies community leadership. • President, American Society for Information Science & Technology • Trustee, Dublin Core Metadata Initiative • Co-Founder, Networked Knowledge Organization Systems/Services • Adviser, National Research Council Computer Science and Telecommunications Board • Reviewer, National Science Foundation Division of Information and Intelligent Systems
Government Chelan County Public Utilities District Commodity Futures Trading Commission Federal Aviation Administration Federal Reserve Bank of Atlanta Head Start Infocomm Development Authority of Singapore (http://mysearch.internet.gov.sg/) NASA (nasataxonomy.jpl.nasa.gov) U.S. Defense Intelligence Agency U.S.D.A. Economic Research Service U.S.D.A. e-Government Program (www.usda.gov) U.S. Dept of Education ERIC U.S. D.H.S. Citizenship and Immigration Services U.S. Environmental Protection Agency U.S. Forest Service U.S. GSA Office of Citizen Services (www.usa.gov) U.S. Small Business Administration U.S. Social Security Administration Commercial Agency.com Amway Albertsons Allstate Insurance Baker Hughes BHP Billiton Blue Shield of California Campbell Soup Company Capital One Debevoise & Plimpton Dell Halliburton Hewlett Packard Microsoft Motorola Oracle PeopleSoft Pricewaterhouse Coopers Siderean Software Sprint Time Inc. NGO’s Dewy Decimal Classification European Committee for Standardization IDEAlliance International Monetary Fund National Association of Realtors OCLC Recent & current projects
What I do Organize Stuff
For us, taxonomy work includes: • Metadata Scheme. Data fields for describing content so that it can be found and used. • Vocabularies. Collections of terms that are used to specify some of the metadata properties. • Relationships between content, fields or terms (hierarchical, equivalence, & associative) • Some vocabularies are big & hierarchical, some are small and flat. • Application Profile. Formal representation of metadata & vocabularies.
Agenda • Content Tagging • Tagging Interface
Tagging Overview • Tagging is better than the words that happen to occur in a piece of content. • All tagging is useful • End user tagging • Tagging by librarians • Automated tagging by OS and algorithms • Content should be tagged throughout its lifecycle, each time the content is handled and used so that it accrues value or its significance is diminished.
MS Office: File Properties How many people fill this in?
Organize How many people click on this?
What is social tagging? • End user tagging • Easy, intuitive tagging interfaces • Almost instantaneous feedback • Enables people to tag & re-tag content • … in response to seeing their tags in context with other tags. • Emergent categories • Resembles open card sort process in which patterns emerge • … rather than validating categories using closed card sorts.
Social tagging innovators • flickr founders • Caterina Fake • Stewart Butterfield • del.icio.us founder • Joshua Schachter • del.icio.us & flickr are now both part of Yahoo! • As of April 2006 flickr had 130 million photos posted by 3 million registered users.
Agenda • Content Tagging • Tagging Interface
Requirements for a tagging interface • Automated form fill-in (automatically fills in known data) • Tagging precedents (see tags already assigned by others) • Controlled vocabularies, e.g., with pull-down list • Multi-valued tags • Geo-tagging • Group tagging • Clean-up tag tools, e.g., alpha list • Batch editing • Share/Don’t share (Public/Private) • Identified owner (who can be emailed) • Almost immediate feedback, e.g., tag cloud
Manual form fill-in w/ check boxes, pull-down lists, etc. Auto keyword & summarization Form fill-in: Automatically filled-in known data
Auto-categorization Rules & pattern matching Parse & lookup (recognize names) Form fill-in: Automatically filled-in known data
Bulk tagging • ID collection of related content items by pattern or context • Then, apply same attributes to all content items
Tag a folder • Drag & drop content items into folder • Then, content items inherit properties of folder
Create Content Add Metadata Publish Review & Improve Review & Improve Workflow • Approve & improve mindset
Interactive rewards • Almost instantaneous exposure of tags in simple user interfaces on the web provides positive reinforcement for user tagging that simply did not exist before. • For example, • Most popular • Tag clouds • Alerts
Most popular • Another example is most emailed from, e.g., the NY Times.
Alerts • New (content selected by date) • Subscriptions (content selected by tags) • Interest (content selected by other people) • Individual (content selected for you by other people)
Summary • There are lessons to be learned from web tagging about how to get good metadata in document and content management applications. • Document and content management system tagging must be simple, and it must be almost instantaneously easier to find relevant work products.
Questions? Joseph A. Busch415-377-7912 email@example.com
Tagging Overview • Tagging, any kind of tagging is better than the words that happen to occur in a piece of content. End user tagging is useful, so is tagging by librarians, as are tags automatically assigned by operating systems and language processing algorithms. Content should be tagged throughout its lifecycle, each time the content is handled and used so that it accrues value or its significance is diminished. • Almost instantaneous exposure of tags in simple user interfaces on the web provides positive reinforcement for user tagging that simply did not exist before. It should not be surprising that a good user interface improves usability. • As content users flock to websites that help to organize the content on the web, advertisements and value added content services follow. The bottleneck in the semantic web has been not enough tagged content. The end user tagging revolution may begin to address this shortcoming. • There are lessons to be learned from web tagging about how to get good metadata in document and content management applications. Document and content management system tagging must be simple, and it must be almost instantaneously easier to find relevant work products.