WebWord.com > Interviews > Understanding the VA HyperFAQ (29-May-2001)

 
If you want to know when new interviews go online,
subscribe to the WebWord.com Usability Newsletter!

Understanding the VA HyperFAQ

An interview with Basil White, the driving force behind the VA HyperFAQ

Conducted via email by John S. Rhodes (29-May-2001)


An Interesting FAQ Indeed

What is the VA HyperFAQ? Why is it interesting? How is it different than other FAQs (Frequently Asked Questions)?

The VA HyperFAQ is an alternative navigation system for www.va.gov.  It indexes the top 50% of all hits to va.gov by asking simple yes-no questions of users until the appropriate content is concise enough to list on a single page of links. 


In other places, you have said that this FAQ is an instance of a binary tree? What does that mean? Why is that good? Are there limitations? 

It's a binary tree because all of the navigation to the final index pages uses yes-no buttons in response to simple questions. 


How did you build the VA HyperFAQ? What tools did you use? Specifically, what technology makes the tool work? 

I used EditPad, a shareware ASCII editor available at jgsoft.com.  There's no database or scripting.  The whole site is 440K, and indexes 2.5 million hits per month worth of hits.  That's the power of binary navigation systems: every additional question doubles the amount of web pages you can index.     


How did you map out the various questions and answers for the VA HyperFAQ? That is, how did you do the information architecture since the FAQ includes over 200 pages? 

I asked our web operations staff to schedule an automatic report of the top 200 web pages on www.va.gov for every month.  Every time I update the HyperFAQ, I build a list of every page that made any of the top 200 reports for the last six months.  

I print each page to make sure that the pages are live and that I'm using the latest locations of redirected pages.  Then I read all of the content to try to form a simple yes-no question that will allow me to divide the stack of web pages as close to two even stacks as possible.  For www.va.gov, that question is "I'm looking for information about a benefit."  

Then I read each of those two stacks to try to form yes-no questions that divide those stacks in half.  I continue forming questions until each stack is small enough to list as a brief page of links.  A few pages belong on both sides of a yes-no question.  For those pages, I just photocopy the page and put it on both stacks.  These pages are rare, and duplicating them doesn't seem to have a significant effect on the density or complexity of the site. 


Web Applications, Desktop Applications

The VA HyperFAQ looks and feels like a desktop application. Users click on buttons and they walk down a certain logical path. Was that intentional or is it a side effect of the design or information architecture? 

The intention of the design was to minimize the cognitive burden on the user.  The HyperFAQ, like other FAQs on the web, are frequently used by people for whom the primary navigation system has failed. For those people, you want to minimize the cognitive burden.  Structurally, there's no simpler cognitive burden than a yes-no question.  There are poorly-formed, confusing yes-no questions, but those are the fault of syntax, not structure. 


On the heels of the previous question, do you think web applications should operate like desktop applications? Why? Why not? 

Personally, I don't care if web applications function like desktop applications, online help files, or the mechanical tic-tac-toe chicken at Coney Island. What's important to me is that the application is easy to use and is customer-focused.  In the case of the VA HyperFAQ, that means indexing content for the veteran as effectively as possible. 


Lessons Learned

What recommendations would you make for other people trying to emulate the VA HyperFAQ? Lessons learned? Tips and tricks? 

I've learned that building interfaces based on a customer goal hierarchy not only improves usability, but it actually eliminates rework, because all of the non-interface changes to the product occur in the lower levels of the design.  Screen and command changes end up affecting Section 3.2.2.4 or some other bottom-level section of the design. 

I build products in the context of goals.  Goal 1, 2, 3 becomes Module 1, 2, 3, becomes Chapters 1, 2, 3 of the manual.  I've found no better way to reassure customers that their goals are Number One. 


How can people learn more about binary trees? How can people learn more about building effective FAQs?

I teach a workshop to Federal IT designers at Customer-Focused IT Workshop (.doc link) which mentions how to develop IT products and their accompanying rhetoric in the context of user goals. I also recommend the free and low-cost resources of the Usability SIG of the Society for Technical Communication.


What next?

  • Feel free to create a link to this interview.

 

If you want to know when new interviews go online,
subscribe to the WebWord.com Usability Newsletter!

 

Home | Moving WebWord | Cool Books | Hot Web Sites
Reports
Newsletter Archive | Interviews | About WebWord.com


Contact John S. Rhodes, WebWord.com Editor and Webmaster

URL: http://webword.com/interviews/white.html

© 2001 by WebWord.com. All rights reserved.
Do not reproduce or redistribute any material from this document,
in whole or in part, without explicit written permission from WebWord.com.