Creating user-friendly electronic information systems

The future of any automated information systems, let them be World Wide Web servers or online public access catalogs (OPACs), will have to include "interactive assistance" features. This article reviews the definition of information systems, describes the concept of interactive assistance, describes how it relates to information systems, and points out a few prototypical examples.

For the purposes of this article, any information system is a collection of organized value-added data. In our world, information systems abound. They include things such as books, billboards, libraries, the dash board of your car, or a World Wide Web server.

The effectiveness of information systems is directly related to its size and how well it satisfies its users' information needs. This effectiveness can be characterized by the system's readability, browsability, searchability, and finally, interactive assistance. Readability connotes well implemented graphic design and visual appeal or practicality. The dash board of your car had better be very easy to read. A magazine article had better be inviting or it is less likely to read. Traditional card catalog cards rated very poorly when it came to readability.

As the size of an information system grows so does the need for browsability. In other words, as the number of items of the information system increases, so does the need to logically organize and classify the items in the collection. Tables of contents in books serve this purpose as well as the top-most subject classifications of services like Yahoo or the ten primary subjects of the Dewey Decimal Classification System.

Searchability overcomes some of the problems of purely classified systems by providing direct access to particular items as well as creating dynamic collections of otherwise dissimilar items. The function is similar to the indexes of books. The downside of searchability is it requires users to articulate their information need and formulate it in terms of the systems' query language.

No matter how well a large information system exemplifies readability, browsability, and searchability, there are always to going to be people who cannot locate the information they seek even though it exists within the information system. The purpose of interactive assistance is to reduce this possibility. The concept is nothing new. After all, it is one of the primary purposes of librarianship. It is usually called "reference," and has been implemented through face-to-face interviews and telephone communication.

While the largest and most familiar information systems like Yahoo, our OPACs, C|Net, or A2Z exemplify varying degrees of readability, browsability, and searchability, they could all use a healthy dose of interactive assistance. Interactive assistance is not just online help or instructions on "how to use this book." Interactive assistance provides customized help for particular users in particular situations. Put another way, interactive assistance adds the next level of intelligence to information systems. Interactive assistance does not necessarily mean artificial intelligence or expert systems, but that would be nice. Interactive assistance is simply a way of specialized help for specialized situations.

Interactive assistance can be proactive or reactive. Proactive interactive assistance queries users for the information needs. It analyses the answers to the queries and either formulates possible solutions to the information need or continues the query process. A reactive assistance model would only provide possible solutions after being asked questions by the users. The difference between these two models is similar to the difference between browsability and searchability. Both browsability and proactive assistance layout ready-made solutions or information paths. Searchability and reactive assistance require the users to articulate their needs and translate them into language of the system. Like browsability and searchability, ideally, elements of both proactive and reactive assistance are desirable in the implementation of any large, automated information system.

With the exception of Microsoft's "Bob" and Apple Computer's Apple Guide, the concept has barely made its way into the world of computers and the Internet. Since the dawn of computers people have tried to create artificial intelligence agents for the purposes of solving problems, such as locating information from an information system. There are a number of texts describing these sorts of systems for libraries. Knowledge-based Systems For General Reference Work by Richardson is the most scholarly to date. Expert Systems For Reference And Information Retrieval by Alberico and Micco offers pointers to a multitude of examples.

With the advent of CDROM-based bibliographic indexes "expert reference services" were being provided for these media as well. The work of Myers from the University of Houston is an excellent example. The service, called Reference Expert, queried users and suggested bibliographic indexes possibly satisfying the information need. Based on some of this research and other projects, it has been determined a library's bibliographic descriptions of collections could benefit from enhancements if qualitative judgments supplemented catalog entries. Such judgments could then be factored into an automated decision-making process.

With the advent of the Internet, providing interactive assistance manifested itself through the use of e-mail. This has almost become a ubiquitous means for providing reference. Another, more innovative solution, was the creation of MOOs and MUDs. These text-based, virtual realities were populated by people with common needs and desires. MOOs and MUDs always seemed popular with the gaming crowd, but because of its text-based orientation it never seemed to cause very much of a stir.

Similar work has been done with video-conferencing. One of the more successful projects has used CU-SeeMe to supplement reference services at the Shapiro Undergraduate science library at the University of Michigan. Another such experiment was the See You See A Librarian project.

The big thing these days is the World Wide Web. Everybody's individual WWW server represents an information system and the larger they get the more they necessitate interactive assistance to make them more useful. The first hints of interactive assistance efforts include relevance ranking and meta-search engines. While these applications hint at interactive assistance, they have a way to go before they provide anything remotely like "proactive" assistance.

Ask Alcuin, proto-typed by myself in 1995, is a Perl script whose purpose is to put into practice the concept of interactive assistance. Using a question-and-answer process similar to a reference interview, Ask Alcuin tries to identify a user's information need and translate that need into the query language of remote and local databases. The program is functional in that it executes correctly, but the method it uses convert information needs into queries needs to be updated and refined. Enhancements could include the log of user needs so repeat users do not have to re-build a profile with the application. Furthermore, the program needs to be updated in terms of the search engines it queries and how it queries them.

Meta-search engines are those Internet applications allowing you to input queries into a field, select various databases, and submit your query. The main strength of these applications is their ability to translate your query into something many databases understand. Examples include SavvySearch, All4one, Cyber411, Inference Find!, and MetaFind. These applications all work with varying degrees of success. The better ones collate the results, remove the duplicates, and list the results is some sort of order. None of these engines help you refine your queries.

WebArcher, a program developed by ClearWay Technologies, is billed as

the first 'Desktop Internet Search Expert' - an unobtrusive desktop program that has built-in expertise about dozens of Internet databases, search engines, and catalogs. It helps you search the way a librarian would: you tell Web Archer what you're looking for, and it uses its knowledge of the Internet to help you find your target quickly and easily.

Using WebArcher is an easy 3-step process. You begin my inserting your query into an empty field. Any text will do, but it does accept specialized search syntaxes. Next, you select the sort of search you want to do: general, news/stocks, telephone numbers, software, etc. Finally, you select "Go!" WebArcher then sends your query off to seemingly random sites across the Internet in search of answers. Subsequent searches for the same query are sent to alternative sites.

Obviously, WebArcher knows how to formulate your simple search queries into the syntax of the remote engines. This is an excellent idea similar to the meta-search engines described previously. Saying it "helps you search the way a librarian would" may be stretching things a bit. And again,the whole process could be improved with a bit of a question/answer process.

In summary, if librarians want to participate in providing information services electronically, then librarians must create useful electronic information systems. These systems, depending on their size, ought to exemplify varying degrees of readability, browsability, searchability, and interactive assistance. Interactive assistance, as specialized help for specialized situations, can take many forms: telephone conversations, video conferencing, or expert systems. The implementation of these electronic services will be a challenge for the profession, especially considering the current "do more with less" mind set, but with perseverance and determination we can make it happen.

Further reading

  1. John V. Richardson, Knowledge-based systems for general reference work: applications, problems, and progress (San Diego: Academic Press, 1995)
  2. Ralph Alberico and Mary Micco, Expert systems for reference and information retrieval (Westport, CT: Meckler, 1990)
  3. Video conferencing and reference service - http://sunsite.berkeley.edu/~emorgan/see-a-librarian
  4. Ask Alcuin - http://www.lib.ncsu.edu/staff/morgan/alcuin/ask-alcuin.cgi
  5. WebArcher - http://www.clearway.com/WebArcher/

Creator: Eric Lease Morgan <eric_morgan@infomotions.com>
Source: This is a pre-edited edited copy for Eric Lease Morgan, "Creating User-Friendly Electronic Information Systems" Computers In Libraries. 17(8):31-33, September 1997.
Date created: 1997-07-10
Date updated: 2004-11-14
Subject(s): interactive assistance; information systems design; expert systems;
URL: http://infomotions.com/musings/information-systems/