UNDER CONSTRUCTION

[In a working MSOLAR system, this would be a page of explanation for interested users (bypassed by those wanting simply to use the system).  To jump to what the users' version would look like, click here.  This section above the line is meant to promote discussion of the whole idea behind MSOLAR.

A system like this takes into account two current trends:

• the growing presence of worthwhile content on the WWW; and
• the growing interest in "virtual" libraries, courses and universities.

By extending reference access in time and space, such a system expands library service by:

• Reaching some users who never opt to consult a living reference librarian.
• Reducing the workload on reference staff, siphoning off certain routine questions (in this regard, MSOLAR is a potential tool for IRC, as much or more than SSHR).
• Serving traditional extension and outreach patrons through a new medium.
• Playing a role in the "virtual university" by offering a "virtual reference" desk.
• Providing ready reference tools for referral to "chat" users.

What is the point of MSOLAR?

There are already numerous "virtual reference desks" on the Web: some examples include the Michigan Electronic Library, or UPenn's Reference Shelf page.  MSOLAR differs in one major way: those sites start with lists of tools and force the user to select among those resources, while MSOLAR starts with the user's need and guides the user to a relevant resource.

If you look at the system and try it out, keep a few questions in mind:

• Who should form the target audience for MSOLAR?  Students after hours, users in campus offices, extension participants, visitors from the Internet?
• Should the list of suggested sources consist only of Web-based tools?  Such a list of resources has the advantage of fully serving remote users, but several disadvantages: it fails to inform users on campus about good print tools, and leaves gaps where only print tools are available.
• Does this kind of access point depress use of traditional desks, or reach a different clientele?
• What are the tradeoffs in using natural language vs. controlled vocabulary for instructions to users?
• How taxing is the work of maintaining such a site, and how can that be delegated? What happens if reference units are divided about its utility?

I expect you will think of more. I'm less interested right now in a critique of specific pages, or their layout, or the URLs selected, than I am in the overall concept of a tool that takes advantage of HTML to lead users to resources, especially Web resources.

This project is based on a similar project called SCHOLAR ("Swarthmore College Hypertext On-Line Assistant for Reference") that was written during 1995-96 for a liberal arts college library. The content for MSOLAR has been edited to reflect the MSU collections. The original SCHOLAR site may still be available: if you want to see it, try clicking here.

The user page would begin here:]

What is the MSOLAR system?

MSOLAR, an experimental set of Web  pages, is an effort to provide a minimal level of ready reference assistance when reference librarians (or even the library itself) are not accessible: between semesters, at three in the morning, during blizzards, from remote locations, and so on. Thanks to HTML and the Internet, some of the advice library users receive from live reference librarians is available here.

How does MSOLAR work?

Like a reference librarian, MSOLAR will ask questions, to help you define and narrow your question. Then a simple question-and-answer tree structure will lead you to suggested reference sources. When possible, Web-accessible sources will be included.  Print tools available in the M.S.U.  Libraries will be listed as well (with their call numbers). Locations are indicated by abbreviated names, such as "IRC" for the Information Referral Center on the main level of the Main Library, "SSHR" for the Social Sciences and Humanities Reference collection (basement level of the west wing of the Main Library), or "Business" for the Business Library.  

MSOLAR asks: "What would you like to find?"

A good reference librarian uses a "reference interview" to help you define exactly what you need, then connects that need to available tools in the library. Open-ended questions--ones that CANNOT be answered "yes" or "no"--are the best questions for this task, but open-ended questions are hard to handle in a tree structure like the one used here. Instead, MSOLAR relies on simple questions with multiple-choice answers.

How extensive is MSOLAR?

You will get better results from MSOLAR if you understand its limits. The selection of Web-accessible tools reflects the realities of copyright and the information market-place.  Some of the best free materials come from the U.S. government, and therefore emphasize politics, the law and the U.S. economy. MSOLAR also limits itself to a selection of "best" sources, only a few hundred titles.  By contrast, the library reference collections contain thousands of  titles. MSOLAR can be thought of as relying on the so-called "80/20" rule, which argues that eighty percent of  reference questions can be answered using twenty percent of the available reference tools.

What if my question is too complex for MSOLAR?

MSOLAR is best suited for "ready reference" questions that can be answered by simple facts, figures and names. Many kinds of questions are too sophisticated for this system. If you are defining a topic for a substantial research project, MSOLAR won't be much help.

When will reference computers replace human librarians?

Not soon. A good reference librarian goes through a graduate school library program to learn what kind of information is found in libraries, how it is organized, and how library users can articulate what they need. A really good reference librarian often has years of experience, too. If reference searches were truly this easy, librarians would be out of work. If anything, the explosion of information available in digital form and on the Internet makes the advice of librarians more, not less, important.

Is anyone looking into better computer-based reference systems?

In the mid-1980s, there was a lot of interest in library use of "expert systems" and "artificial intelligence." After a while, it became clear that computer libraries like the ones on "Star Trek" are a long way off. If you want to read a bit about ongoing library discussions, take a look at:

• Karen F. Smith, "Robot at the Reference Desk?" in College and Research Libraries, vol. 47, no. 5, pp. 486-90 (September 1986)
• Kenneth Quinn, "Expert System Shells: What to Look For" in Reference Services Review, vol. 18, no.1, pp. 83-86 (1990)
• Charles W. Bailey, Jr., "The Intelligent Reference Information System Project. A Merger of CD-ROM LAN and Expert System Technologies" in Information Technology and Libraries, vol. 11, no. 3, pp. 237-44 (September 1992)
• James Gunn, "'Dreams Written Out': Libraries in Science Fiction" in Wilson Library Bulletin, vol. 69, no. 6, pp. 26-29 (February 1995)

Return to MSOLAR's first page.

Click here to try MSOLAR in its test version.