Notes from 45th International STC Conference
Anaheim, California, May 17-20, 1998

Building the Treasure House:
Creating Knowledge Bases for the World Wide Web

Jack Massa
Guidance Communications, Inc., Tucker, GA

Abstract: This workshop explores the components of a Web knowledge base and the steps needed to plan and build one. Includes interactive discussions and a writing exercise.

• Note: 16-page handout available... you can download it by contacting Jack Massa via the above Web site or give me a call (or e-mail note) and I'll send it to you.
• Definition of knowledge base: "A database that contains information about human experience in a particular application and data resulting from solution of problems previously solved." -- IBM Dictionary of Computing, 1987
• Benefits of a knowledge base
   
• Central repository of product info (both internal and external)
• Quick reference
• Easy update
• Accessibility (24 x 7 availability of info)
• Searchability
• Cheapest form of widespread distribution
• Reusability of information
• Reduced tech support costs... improves the bottom line!
• Technical communicators tend to focus on helping new users, rather than on providing easy-access help for advanced users who only need help when things go wrong.
• Key elements in developing a knowledge base
   
• Buy-in from management and from customer support
• The database itself (content)
• Organization/structure/outline
• Tie-in to tech support ($)
• Maintenance plan
• Test plan (including usability testing)
• Work flow process
• Technical validation of information
• Tools
   
• HTML code
• Web design tools
• Link management... it's important to check links regularly and keep them current
• Server
• Search engine
• Search engine is a critical component
   
• Greatly increases usability of information
• Tools are available, depending on the server
• Free search engines are available on the Internet and can be downloaded
• Abstract returned by Search can be controlled by META tags. META tags go in the header section of the HTML code; they are read by the browser (and thus are "hits" for search engines) but are not displayed.
• Printability is an important user consideration. This is a strong argument for some scrolling (i.e., multiple-screen segments of data) rather than "tunneling in" down a vertical hierarchy of single screens. If the user is likely to have to print out the material to resolve a technical problem, he/she isn't going to want to separately call up several screens and have to hit Print a half dozen times. A continuous page design, including a reasonable amount of scrolling, prevents this. On the other hand, excessive scrolling is still verboten; the user will rapidly tire of that.
• Some HTML tips:
   
• Use narrow columns for readability. These can be achieved using TABLE tags.
• Make liberal use of subheadings.
• Use graphics wherever applicable.
• Place critical information near the top. People searching for solutions are generally impatient.
• Typical content of knowledge bases
   
• Product overviews
• Existing documentation
• Frequently asked questions (FAQs)
• Trouble reports/fixes/workarounds
• Tips
• Troubleshooting solutions
• Error messages
• Article content
   
• Article ID
• Date of last update
• "Pertains to" info
• Legal disclaimers
• User feedback tool... "Did this help you?"
• Workshop exercise: Given the following case record from the technical support call logs of Acme Robot Company, write a knowledge base article that would help the customer to solve the problem without calling technical support. Use the template on the next page as an aid.
  
  Note: Exercise is available with the handout.