Computerized information is now arguably as ubiquitous as printed 
information. Increasingly, a large proportion of the electronic 
information that faculty, researchers, students and librarians use is 
housed, not on personal or even local institutional computers, but on 
computers that are miles away and oceans apart.
     Gopher is a navigational tool developed by the University of 
Minnesota for exploring the Internet. Users can access local 
information on their institutional gopher and connect to other gophers, 
networks and resources throughout the world. Locally-built 
institutional gophers are becoming increasingly important for helping 
patrons navigate the Internet and connect to information in a variety 
of subject areas. 
     Today, dozens of institutions have already created gopher or Web 
servers that contain local information and pointers to related Internet 
information. Many more institutions are planning to do so. Our 
experiences in creating and maintaining a biomedical gopher at the 
Yale-New Haven Medical Center have taught us lessons worth sharing 
with those who are planning or maintaining gopher or Web servers. 
We've also reached some conclusions about future directions for the 
organization and compilation of Internet information. Our experiences 
were in the biomedical area, but the process is the same for building 
and maintaining any gopher.

HOW TO DEVELOP A GOPHER
     Consider these four major areas in developing a gopher--
organization, promotion, training and maintenance.

ORGANIZATION
     Our guidelines for organizing information grew out of 
brainstorming sessions among computing and library professionals to 
plan the university gopher [1] and meetings of a committee of 
computing professionals, librarians and faculty devoted to creating 
and maintaining a local gopher [2]. Several guidelines emerged from 
these groups.

* Organize the gopher by the discipline or subject area. Organizing by 
discipline allows users to connect by a familiar field of study, 
encourages general use and facilitates use by faculty and researchers.

* General information is then organized under several useful headings, 
such as events and calendars, grants information, university or 
institutional information and other Internet resources.

* Redundancy is desirable because it helps users find what they are 
looking for. Each information source need not be restricted to a single 
location in the gopher, since there is no overhead in having the same 
information source appear in several gopher submenus. This provides 
multiple pathways to information sources and permits general 
information sources to 
be listed under several specific submenus. 

* It is easier for people to find information using longer menus with 
fewer levels than short menus with many levels to get to an 
information source [3]. At the same time, it is best to create menus 
that can be viewed on a single screen [4]. Lengthy menus, however, are 
appropriate for alphabetized lists such as biomedical disciplines, 
diseases or items sorted by date, such as journal issues. 

* Standardize menu features in fixed positions because consistency is 
important. For example, information about the gopher could be in the 
first position, searching in the second, etc. The most important and 
frequently used items should be higher up in the menu. Include a top-
level menu of shortcuts to the most popular services, such as 
telephone directories, schedules of events and the library catalog and 
databases, since this facilitates access to them.

* Specify the originating source of the information to help manage 
user expectations. For example, if a user knows the information is 
coming from a gopher in Australia that is not working, local support 
staff will not receive complaints they cannot address.

PROMOTION
     As with any major new service, it is important to promote a gopher 
using a variety of approaches.

* Write articles about the gopher for organizational newsletters. A 
regular column that contains information about and additions to the 
gopher could be added to relevant newsletters. The recent hype about 
the "Information Highway" and the many articles appearing in popular 
magazines about the Internet have piqued people's curiosity. Include 
references to these articles in publicity.

* Hold workshops and seminars to promote the gopher. Have an 
information fair and spotlight the gopher. 

* Create a user guide, which not only helps with training but 
legitimizes information.

* The gopher could be a menu item on public and personal workstations. 
At Yale, we have added our gopher as a selection on NetMenu [8], the 
navigation tool to the campus network that runs on both public and 
individuals' computers.

* Word of mouth is also an important way to promote the gopher. 
Discuss the gopher with faculty and research staff and as part of 
library and information systems seminars and orientations. Suggest 
that librarians recommend the gopher as an information resource at 
the reference or information desk. If they used the gopher to answer a 
reference question, they can explain to the patron how the information 
was found.

* Encourage people to contribute information to the gopher. 
Contributors tend to become strong promoters. This also helps develop 
a critical mass of local information and spurs others to contribute 
their own information.

* Announce the new gopher on the Internet using forums such as Usenet 
news groups and LISTSERVs. And, of course, register the gopher with 
the University of Minnesota. (Send electronic mail to 
gopher@boombox.micro.umn.edu. Include the server name, host name, 
port number and administrative contact.)

TRAINING
     To help users become familiar with how to find information on the 
gopher:

* Create seminars and organize them by subject. In our case, we 
offered a gopher seminar for clinicians emphasizing clinical 
applications, another one for basic scientists emphasizing science 
resources, one for public health students about finding international 
health resources, and so forth.

* Offer point-of-use instruction by helping users use the gopher on 
public access terminals and by giving demonstrations in offices and 
laboratories.

* Create an outline of the information in the gopher. Have a printed 
copy for new gopher users, as well as an electronic version on the 
gopher.

* Create online help files, since most computer users prefer help when 
and where they need it (onscreen). Some users prefer to refer to a user 
guide, however.

* Tell users how to submit information during seminars and training 
sessions. 

MAINTENANCE
     An important step in maintaining quality control for a gopher is to 
monitor all non-local sources to be sure they still work and to add new 
sources as they become available.

* Connections should be checked weekly. It is faster to check 
connections at times of day when there is relatively low traffic on the 
Internet, such as morning in the U.S. Eastern Time zone. The weekly 
checks can be done by a clerical person, with problems referred to the 
gopher administrator for resolution.

* Search the Internet monthly for new information sources. A number 
of gophers have created "subject trees" that list gophers by subject. 
This is one useful place to look for new information resources. Another 
good way to find new sources is to perform searches on relevant 
subject terms using search tools such as veronica [5], archie [6] and 
jughead [7].

* Scan literature on the Internet and review library, Internet and 
computer magazines and Internet directories. 

* Subscribe to subject-oriented LISTSERVs to learn about new 
information sources.

* Collect suggestions from users and library and information systems 
staff. 

* Form a committee of content experts to review and recommend 
sources and suggest changes in the gopher's organization. 

* Monitor the status of alternative platforms such as the World-Wide 
Web (WWW). Right now, gopher is the best tool for the majority of 
users, but the searching and display advantages of Mosaic, other net 
browsers and WWW make them highly desirable platforms for patrons 
with high-speed network connections.

ECONOMICS OF RUNNING A GOPHER SERVER
     Economics will probably play a role in determining whether to 
create a gopher server. Table 1 displays our estimates of the resources 
needed to develop and maintain a gopher. 

DEVELOPMENT COSTS
     We found development costs to be relatively straightforward and 
low, given the vast amounts of information the gopher will make 
accessible. The gopher server software is free to nonprofit 
organizations and can be downloaded from the University of 
Minnesota's gopher. Once downloaded, the server software can usually 
be installed on any existing machine running a wide variety of 
operating systems, including UNIX, VMS, Macintosh, DOS, OS/2, MVS and 
VM/CMS. Hard disk requirements depend on the amount of local 
information that will reside on the gopher server. Gopher data will 
ultimately require roughly double the hard disk space of the data 
itself, due to space needed for an index. The actual gopher server 
software uses very little hard disk space; the UNIX version, for 
example, is less than 5MB. We estimate that it will take most 
programmers approximately one to two weeks to set up a new gopher 
server. 
     It is useful to form a committee of computing and library staff, 
faculty and students to help define the gopher's intended scope and 
recommend information sources. In addition, a content expert with the 
ability to comb the Internet for relevant information sources may 
spend as much as a week compiling a list of sources to include 
initially. 
     Finally, a librarian or computing support staff member can be 
trained by a senior programmer to construct the gopher menus, mount 
local information and add links to existing Internet information 
sources. This task can easily take the equivalent of two weeks of full-
time work.

MAINTENANCE COSTS
     Maintaining a gopher is straightforward but time-consuming, and 
accounts for the abundance of obsolete gopher servers on the Internet. 
A programmer is needed on an ongoing, infrequent basis for such tasks 
as maintaining and upgrading the server software, building WAIS 
indices that underlie gopher searches, and automating the addition of 
new local information sources.
     Links to Internet information sources tend to change or disappear 
without notice. We have assigned an administrative assistant the 
weekly task of monitoring our links to Internet information sources. 
This takes two to three hours, and a computing staff member then 
must spend an additional hour or so to update or remove any reported 
broken links. 
     Information is being placed on gophers at a rapid and increasing 
rate. We have a content expert devote one or two days each month to 
seeking new information sources on the Internet, adding them to our 
gopher, documenting the changes in an online gopher file, and updating 
online and paper outlines of the information in our gopher. New 
information sources that have been suggested by other librarians, 
computing staff, and patrons are also added to the gopher at this time.
     The committee we formed to determine our gopher's initial content 
and structure also laid out guidelines for adding Internet information 
to the gopher. Because the School of Medicine already has Yale's 
NetMenu [8] as a front-end menu to major information services 
(MEDLINE, the NIH Guide, online catalogs, NEXIS, etc.), the content 
committee agreed to place very few restrictions on the information 
covered in our gopher. We include any databases or information sources 
that are of potential interest, but not likely to be harmful to members 
of our community [2].

OUTCOMES AND FUTURE DIRECTIONS
     We found there are several positive outcomes from developing a 
gopher, some of which accrue from gaining access to external 
information resources, and some of which are the result of mounting 
local information electronically.

GAINING ACCESS TO EXTERNAL INFORMATION
     A gopher server advances our missions of research and education by 
giving patrons easy access to a variety of information resources. 
These information resources include library catalogs, subject gophers, 
clinical guidelines and organizational gophers such as the World Health 
Organization, U.S. government agencies and the vast array of college 
and university gophers around the world. In addition, a gopher 
encourages cooperation among institutions worldwide, because local 
users gain access to people and information at other institutions and 
collaborate to mount additional information. 
     A local gopher improves existing access to Internet resources in 
several ways. It simplifies the process of connecting, allows sites to 
tailor organization of materials to patrons' needs and offers tools such 
as veronica, archie and jughead to facilitate searching for Internet 
information.
     Finally, a local gopher helps patrons understand the Internet by 
permitting users to focus on information rather than deal with 
intimidating connection and retrieval tools (such as FTP and telnet). 

MOUNTING LOCAL INFORMATION
     A local gopher also permits an institution to offer its users 
widespread access to local information resources and databases. Our 
gopher contains information previously available only in paper form, 
such as the Medical Center's Weekly Schedule of Events, the Faculty 
Research Interests Database and information from the Office of 
Research Affairs. Mounting local information is the most compelling 
reason to develop a local gopher server. The information directly 
benefits local users and cannot be found on external gopher servers. 
Distinguishing between the goal of mounting local information and that 
of providing access to information on the Internet can help an 
institution determine a focus for a gopher. It may choose to develop its 
own organizational scheme for Internet information or concentrate on 
mounting local data and simply point to existing gophers to access 
information elsewhere on the Internet. 
     Development costs for a gopher server are moderate. Maintenance, 
however, can be taxing. Maintaining local data is often easier than 
maintaining links to other gopher servers because it is easy to 
allocate maintenance to the information providers. Providers are 
naturally motivated to update their own information and tend to focus 
more on the benefits of mounting information (increasing access to 
their information, reducing or eliminating the costs of paper 
distribution) and less on added tasks such as maintenance. 

LINKING TO EXTERNAL INFORMATION
     Maintaining links to external gopher servers is a task that tends to 
be delegated to a few already busy individuals. This helps explain why 
many gopher servers become neglected after an initial period of 
careful maintenance. Most organizations take care to keep their own 
data and links to important external gophers up-to-date, but it is easy 
to let other external links wither. 
     If this is true, then why are so many libraries and schools creating 
gophers that all link to the same Internet information? At least 15 
institutions have mounted the NIH Guide to Grants and Contracts on 
their own gopher servers, despite the fact that they could simply link 
to the Guide on the NIH's gopher server. When the National Library of 
Medicine reorganized its gopher last year, how many hours did gopher 
administrators around the world spend fixing their links to 
information on the NLM gopher? And how many gopher servers' links to 
the NLM gopher remain broken? Why are we all reinventing the wheel?
     One explanation might be a need not just for access to information, 
but for access to organized and well-maintained information. We have 
used the gopher's ability to collect usage statistics to track local and 
Internet use of the information in our gopher. This helps us learn which 
local and external information is most valuable to our users and how 
much specific local information is being used by outside users. 
     We discovered a dramatic increase in Internet usage in October 
1993, when we formally announced our gopher over the Internet. Since 
then, external Internet users have accounted for between about one-
quarter and one-half of overall usage of our gopher, but the greatest 
draw for external users seems to be not local information but our 
_organization and maintenance of links to external information_. 
External use of the discipline and disease section of our gopher has 
fluctuated between 53 percent and 76 percent of accesses, as 
compared to a range of seven percent to 38 percent for the Weekly 
Schedule of Events. 
     Several other explanations for the abundance of gopher servers are 
possible. The desire to pioneer or at least not be left behind by a new 
technological development, pressure from users and the legacy of 
printed media all play a part in the proliferation of gopher servers. 
Gopher and the World-Wide Web are buzzwords these days. They are 
tools that create easy access to a wealth of important information, 
and sometimes organizations think they are remiss if they do not make 
these tools and the resulting information available to their users. 
Furthermore, organizations often have information they believe should 
be made available electronically via gopher, offering a compelling 
argument for creating one's own gopher server. Once local information 
is mounted, it is only natural to link to external information.
     If the Internet community has not begun to question the 
inefficiency of having many gophers maintain many of the same links 
to Internet information, it is at least in part because it has always 
been necessary for each organization to maintain its own collection of 
printed materials. For instance, what medical library would consider 
itself complete without its own paper subscription to the _Journal of 
the American Medical Association_? Indeed it is prudent, not wasteful, 
to maintain multiple copies of the more popular books and journals in 
the same physical space. Therefore, we tend not to question the 
redundancy of having every gopher maintain its own links to 
information. 

STREAMLINING ORGANIZATION AND MAINTENANCE
     Although these factors help explain why so many gopher servers are 
all pointing at the same information on the Internet, are they 
compelling enough to justify the time required to maintain existing 
links and unearth and evaluate new ones? It is time to consider a 
different solution.
     One custom that has evolved among gopher administrators has been 
voluntarily apportioning maintenance of global gopher resources. Most 
gopher sites want access to libraries, telephone directories, subject 
trees and the complete list of all gophers on the Internet. Rather than 
have every gopher site compile and maintain its own lists, some 
gopher administrators have volunteered to develop and maintain 
definitive lists of links. Administrators at Yale University, the 
University of Michigan and the University of Texas at Dallas maintain 
the list of libraries. Notre Dame University maintains the list of 
telephone directories. Rice University and Michigan State University 
maintain a list of subject trees gleaned from searches of discipline-
specific gophers at more than a dozen institutions. And Washington and 
Lee University maintains the complete list of all gophers. 

SUMMARY AND CONCLUSIONS
     To develop a successful gopher, you must train users and 
contributors and pay careful attention to the gopher's organization, 
promotion and maintenance. Start-up costs are relatively low and 
finite. 
     It is difficult to maintain a gopher, however, without acquiring 
additional staff or redeploying existing staff members. Maintaining 
local data is critical and easy to justify. Maintaining links to external 
gopher data can be very time-consuming and is more difficult to 
justify. 
     Currently, much redundant effort is expended worldwide by gopher 
servers maintaining links to the same resources. In the biomedical 
field, two solutions to this problem are creating one definitive 
biomedical gopher server or systematically apportioning responsibility 
for updating links to biomedical information on the Internet among 
several biomedical sites. 
     Either of these solutions would free libraries, universities and 
other biomedical organizations to develop, mount and maintain 
additional biomedical resources. Other disciplines will need to make 
the same decisions as gophers and Web servers proliferate.

ACKNOWLEDGMENTS
     The authors are indebted to Cindy Crooker, Marie-Christine Mahe, 
John Paton and Nancy Roderer for the helpful suggestions they made 
during preparation of this article.

REFERENCES
[1] Long, P.L., personal communication, 1993.

[2] Marone, R.K., S. Grajek, M. Helenius, S. Powsner, M. Shifman and L. 
Stone-Infeld. Information Sources Committee Report, _Yale School of 
Medicine Technical Report_. (September 1993).

[3] Shneiderman, B. _Designing the User Interface: Strategies for 
Effective Human-Computer Interaction_, 2nd ed. Addison-Wesley 
Publishing Company, Inc., 1992.

[4] Geller, M., M. Lippert, M. Manoff, K. Morgan and C. Snowden. "GARP's 
Good Gopher Rules." _MIT Libraries Gopher Implementation Team 
Report_ (April 21, 1994).

[5]  Krol, E. _The Whole Internet User's Guide & Catalog_, 2nd ed. 
Sebastopol, CA: O'Reilly & Associates, Inc., 1994, pp. 247-251.

[6] Ibid., pp. 187-207.

[7] Ibid., p. 251.

[8] Shifman, M.A., J. I. Clyman, J.A. Paton, S.M. Powsner, N.K. Roderer and 
P.L. Miller. "NetMenu: Experience In The Implementation of An 
Institutional Menu of Information Sources." _Proceedings of the 
Seventeenth Annual Symposium on Computer Applications in Medical 
Care_. Washington, DC, November 1993, pp. 554-558.

[9] Kahn, R.E. "The Role of Government in the Evolution of the Internet." 
_Communications of the ACM_ 37, No. 8 (August 1994): pp. 15-19. 

[10] Riley, R.A. and B.L. Shipman. "Building and Maintaining A Library 
Gopher: Traditional Skills Applied to Emerging Resources." Presented 
at the 94th Annual Meeting of the Medical Library Association, San 
Antonio, TX, May 1994.


Communications to the authors should be addressed to Susan Grajek, 
Ph.D., Associate Director, Office of Academic Computing, Yale 
University, or R. Kenny Marone, M.L.S., Associate Director, 
Cushing/Whitney Medical Library, Yale University, School of Medicine, 
333 Cedar Street, P.O. Box 208000, New Haven, CT  06520-8000; 
Internet--susan.grajek@yale.edu or Internet--
marone@biomed.med.yale.edu.

Home Page

Home Page

Issue Contents

To Subscribe

Top

Copyright © 1995, Information Today, Inc. All rights reserved.
Feedback
[This site created for best results under Netscape.]