Campus-Wide Information Systems: User Publication of Documents

Copyright 1993 CAUSE. From _CAUSE/EFFECT_ Volume 16, Number 4, Winter 
1993. Permission to copy or disseminate all or part of this material is 
granted provided that the copies are not made or distributed for 
commercial advantage, the CAUSE copyright and its date appear, and 
notice is given that copying is by permission of CAUSE, the association 
for managing and using information resources in higher education. To 
disseminate otherwise, or to republish, requires written permission. For 
further information, contact Julia Rudy at CAUSE, 4840 Pearl East 
Circle, Suite 302E, Boulder, CO 80301 USA; 303-939-0308; 
e-mail: jrudy@CAUSE.colorado.edu


                     CAMPUS-WIDE INFORMATION SYSTEMS: 
                      USER PUBLICATION OF DOCUMENTS
                     by Beth Ardoin and William Weems

ABSTRACT: Two years ago, the Office of Academic Computing at the 
University of Texas-Houston Health Science Center began to implement a 
campus-wide system for the dissemination of information within the 
University. In the last eighteen months that system has been tied into a 
world-wide system via the Internet. This article explains the growth of 
that system and how the placement of information into the system is 
managed.


During the last six years, the University of Texas-Houston Health 
Science Center has developed an open-systems network, based on the 
Transmission Control Protocol/Internet Protocol (TCP/IP), that 
currently interconnects over 3,300 computers. This network continues 
to expand at a rate of about 200 nodes per month and is part of the 
larger Texas Medical Center Network that includes over 5,300 Internet 
nodes. The goal is to create an integrated information system that 
maximizes the sharing of all computer-based information. This system 
must provide information from local as well as international sources 
directly to individuals as needed, and must evolve as technology and 
standards change. It must also allow information to be easily added to 
the system by its users.

The University's TCP/IP-based network began in response to 
federal initiatives designed to standardize on a single transport 
protocol for research funded by the National Science Foundation and 
the National Institutes of Health. Concurrently, most UNIX 
workstations that were commercially available were being shipped with 
Ethernet ports and TCP/IP software. We began to realize that it might 
be possible to use the basic TCP/IP protocols to provide a set of 
common information services for virtually any type of computer. These 
services included the simple mail transfer protocol (SMTP) for 
electronic mail, "telnet" capability for terminal access to remote 
hosts, the file transfer protocol (FTP) for the exchange of ASCII 
(American Standard Code for Information Interchange) and binary files, 
and the network file system (NFS) for client/server-based file 
exchange. As TCP/IP-based local area networks (LANs) started to appear 
within the Texas Medical Center, the University of Texas Office of 
Telecommunications provided links for these LANs to become a part of 
the Internet. This access allows the same protocols for information 
sharing that had been used locally to be used in accessing the 
Internet.

With several existing LANs and wide-area networks on campus and 
the promise of more, it was obvious that there was a demand by 
students, faculty, and staff for Internet access. A number of 
administrators and users were realizing the feasibility and 
convenience of "posting" documents on the Internet rather than 
dispersing the information using traditional methods. Thus, the idea 
of implementing a campus-wide information system (CWIS) emerged.

Homegrown CWIS implemented

To meet the immediate demand for the distribution of ASCII-based 
documents, the Office of Academic Computing (OAC) developed an in-
house CWIS consisting of script files forming a menu-based system. The 
system was accessible only through computer terminals, either via an 
anonymous telnet session or dial-up phone lines. The menus and 
associated processes were "hard-coded" using UNIX script commands, and 
were not standardized. Protocols such as FTP, REXEC, and NFS were 
driven by script or C programs to automatically obtain information 
from other Internet nodes and incorporate them into ASCII files for 
distribution. However, no attempts were made to integrate our CWIS, as 
a whole, with other campus systems.

The initial CWIS clearly demonstrated to the campus community 
that a distribution system that primarily delivered ASCII text had 
valuable potential. In particular, it was recognized that (1) users 
would always know where to look for information, (2) desired 
information could be accessed from any Internet node, (3) updated 
information would be immediately available, and (4) publication costs 
would be drastically reduced. Although these strong points were 
appreciated, the system definitely had shortcomings. It suffered in 
that (1) it could not be efficiently maintained, (2) it did not adhere 
to a user interface that was standard for similar information systems, 
(3) its existence was not readily visible to the world, (4) it 
provided no standard method to transparently access other functionally 
similar information resources, and (5) it was not a client/server-
based system that coupled the resources of personal computers with 
those of information servers.

WAIS introduced

As these advantages and limitations were being considered, the 
Wide Area Information Servers (WAIS) system was developed by Thinking 
Machines Corporation, Apple Computer, Dow Jones, and KPMG Peat 
Marwick. This client/server-based system permits clients on both 
personal and larger computers to access text, images, voice, or 
formatted documents located on WAIS servers. As a searching tool, WAIS 
makes finding relevant text in a document easy, because the WAIS 
search engine, not the user, does the work. The WAIS indexing process 
categorizes the words and makes all words of the entire text 
accessible for search, with exclusion of articles, conjunctions, etc. 
Combinations of words can even be used. In cases where phrases are 
submitted for search, the WAIS search engine responds with as close a 
match to the entire phrase as it is able to make, then lists all 
references that contain each of the individual words.

While WAIS is an excellent tool for searching if you are looking 
for specific data, the system of WAIS servers also has weaknesses. For 
example, the list of WAIS sources has grown by leaps and bounds. 
Although there is a list of registered sources, no hierarchical menu 
exists that groups the sources by subject. Also, you cannot just 
peruse the text without first trying a word search. If you just want 
to read a small portion of the text, WAIS is not the tool to use.

For the University, WAIS became a dynamic tool. The following are 
a few examples of how the WAIS system has demonstrated its usefulness.

  *  The University's Office of Information Services (which oversees 
administrative systems) works with the Human Resources department to 
pull together information on all employees, including names, 
departments, home and work addresses, phone numbers, and so forth. 
This information, used with the WAIS tool, allows anyone with Internet 
access to search the campus phone book to locate an employee or 
employees by name or other descriptive information.

  *  The Office of Research Services gathers information on all active 
labs, research being done, papers written, animals and chemicals used, 
and so forth, and compiles this information into the Catalog of 
Research Expertise. With the WAIS system, researchers on or off campus 
could find UT-H labs working with any variety of chemicals, 
techniques, or animals.

  *  The Purchasing and Receiving departments have used a similar 
process to help researchers and staff locate hazardous material 
numbers.

As a stand-alone system, however, WAIS does not provide a basic, 
integrated CWIS system. In particular, the protocol does not provide 
for a standard user interface that appropriately organizes documents 
and services so that they can readily be browsed according to subject. 
The protocol also does not specify a standard, intuitive method for 
adding available sources to WAIS clients. To partially resolve these 
problems, text- and X-windows-based WAIS clients were implemented on 
the host supporting our initial CWIS. This approach allowed users to 
be directed by the CWIS menus to appropriate WAIS sources. However, it 
negated the important advantages associated with having clients 
running on personal computers so that information is directly and 
transparently delivered to their desktops.

Gopher protocol adopted

Since the demand for a more functional, expanded CWIS continued 
to grow, we systematically began to investigate the efforts of other 
universities and groups to develop a standard CWIS protocol. As a 
result of this investigation, we concluded that the Gopher protocol 
for document distribution developed by the University of Minnesota 
would be used to implement a new CWIS for the Health Science Center.

This protocol was selected for a number of reasons. For one, 
users see the world-wide networked system of Gopher servers as a 
single hierarchical system of documents, directories, full-text search 
tools, and other services. Second, it is a client/server-based system 
that enables users to easily and transparently access servers anywhere 
on the Internet and functionally transfer data to their personal 
workstations. Third, it incorporates a well-defined set of standards 
for defining data types, establishing connections between servers, 
transferring data, linking menu entries to data files, etc. Fourth, it 
is becoming a de facto standard for colleges, universities, and other 
organizations that utilize the Internet for the rapid exchange of 
digital information. Finally, Gopher is a recommended protocol for 
document distribution within the NSF implementation plan for the 
Interagency Interim NREN.

Our Gopher-based CWIS was activated in March of 1992. Users were 
immediately impressed with the system's ease of use and simple design. 
Gopher requires virtually no training for access or use of the system. 
One of its best selling points is its connectivity. Gopher servers, 
once listed by the Minnesota Gopher, are all capable of being 
interconnected. To move from one Gopher server to another, the user 
merely chooses a different topic. The addressing and other 
connectivity issues are all invisible to the user. It is as though any 
information available via any Gopher server resides on the user's 
personal computer. Files can be mailed, saved to disk, or printed by 
an average user, without training.

Converting our CWIS to Gopher necessitated the creation of an 
editorial staff and new document publishing tools.

Editorial staff function required

Although the OAC had only seven staff members, training and 
experience among this group were diverse. Two had newspaper or 
yearbook experience (layout/design), four were programmers, one was a 
systems expert in UNIX, three had expertise in PCs, one was a Mac 
expert, several had integration experience, and three had teaching 
degrees. With these overlapping skills, we were able to draw on the 
strengths of each to implement the Gopher server.

A team was formed for that purpose, made up of an "editor-in-
chief," who headed the team and oversaw all work, and other staff who 
had responsibility for various pieces of the whole. It was imperative 
that the group work as a team, rather than each holding tightly to his 
or her own piece of the project. In addition to the editor, the team 
included programming and design functions. With such a team, we 
expected the end product would be usable, effective, and aesthetic.

The assistant to the OAC director took on the responsibility of 
editor-in-chief, because of her background in publication, while the 
director assisted in editing and creating the menu structure. We felt 
it was more important for the editor-in-chief to have strong 
communications and information skills than to have programming skills. 
This helped avoid the possibility that the end product would be geared 
only toward computer-proficient users. The CWIS had to be easy to use 
to attract non-traditional computer users.

The editor-in-chief became responsible for gathering information 
for inclusion, assisting information providers in preparing their 
information properly, creating the overall vision of the CWIS, and 
structuring the menu. She also found that creating handouts of 
instructions for information providers and teaching 1-1/2 hour courses 
on how to access CWIS greatly increased use and interest, so these 
were added to her responsibilities. The computer category on the CWIS 
was structured and is still maintained today by OAC's assistant 
director, who heads the programming staff.

Once the Gopher server was up, the editorial staff (consisting 
primarily of the editor-in-chief from the development team and the 
director) assumed several ongoing responsibilities; primary among them 
was maintaining a state-of-the-art electronic publishing environment. 
This requires the staff to evaluate and implement document 
distribution protocols, publishing servers, and reader clients. 
Appropriate software tools to support publishing activities are also 
developed as needed. These tools are requested by the editorial staff 
and developed by the programming staff.

A second responsibility is the layout and production of the CWIS. 
This includes assisting the information providers in defining 
appropriate text for distribution, as well as the menu and submenu 
categories. When campus groups become interested in publishing new 
information on the CWIS, some editing requirements are necessary. 
While we agree to hand-hold the information providers in formatting 
the documents for publication, it is important to note that the 
responsibility for producing ASCII documents lies in their hands, not 
the OAC staff's. As a starting point, a list of standards was created 
(see sidebar for a sample of some of these).

When ASCII documents are appropriately formatted, the editorial 
staff assists in creating or modifying existing menus and assigning 
menu ownership. Owners of information are often provided with 
appropriate tools that enable them to self-publish their material via 
the CWIS. If necessary, the staff will enter the information for the 
owner until tools can be created.

The editorial staff is also responsible for monitoring the overall 
look and feel of the CWIS. These activities include (1) confirming 
that menu pointers to remote information sources remain functional, 
(2) ensuring the general availability and timeliness of documents, (3) 
checking the functionality of WAIS sources, and (4) guaranteeing the 
adherence to certain basic standards. These standards include 
conformance to the formatting guidelines for ASCII text, the inclusion 
of ownership information within documents, conformance to copyright 
law, and the general readability of documents.

Finally, the editorial staff has had to assume certain public 
relations responsibilities. The "reader" base had to be expanded by 
advertising, assisting users, analyzing usage trends, and responding 
to user suggestions. The publishing service also had to be actively 
promoted by contacting possible information providers, demonstrating 
both user and publishing aspects to various individuals, and 
responding to queries about the publishing capabilities of the CWIS.

Publishing tools developed

Publishing tools have been developed in ANSI, C, and Bourne 
scripts to enable users to directly publish information from either 
their desktops or from mainframe hosts. One of our publishing tools is 
called Mailman. Mailman was developed to permit users to publish ASCII 
documents directly, via e-mail. The process begins when a document is 
created for publication. Before publishing, the document must be 
converted to ASCII format. Then it can be mailed to the user "Gopher" 
(or another specified userID) on a computer that is running the Gopher 
server daemon. (The Gopher daemon is a utility program running in 
background on the server, awaiting queries from Gopher clients.) Mail 
addressed to this designated userID that is received by the host is 
sent to the Mailman program.

Mailman checks to determine if the person sending the message 
from that particular host is in a list of approved user/host pairs 
with permission to remotely publish information. If the userID at the 
specified host/domain is included on this list, Mailman checks the 
subject field of the message to determine where in the hierarchical 
Gopher system the information is to be placed. In other words, the 
Mailman program checks to see that the userID, host/domain name, and 
subject heading all match. If all checks out, the ASCII file is 
published. As an added security for quality work, the Mailman program 
then sends e-mail to the editor-in-chief, with information on which 
files have been updated. This allows for checking of quality and 
format. A diagram of this process is shown in Figure 1.
(Figures not available electronically)

A second publishing tool created by the programming staff, 
MakeWAIS, was produced to convert information from mainframe 
computers. When a report is generated, either at set intervals or on 
demand, it is converted to an ASCII text file from the appropriate 
database, located on the mainframe computer. The ASCII file is then 
retrieved by a UNIX-based computer, using the FTP protocol. Once the 
data are on the UNIX server, the MakeWAIS tool indexes the data into a 
full-text WAIS source. As on all WAIS servers, the file is then 
completely searchable.

When MakeWAIS creates a WAIS index, the program produces several 
files. These files need to be moved to proper areas in the Gopher 
server. The MakeWAIS tool moves all of these files. If the information 
becomes obsolete, the CleanWAIS tool retrieves all of the various 
files for that database and removes them from the system. This process 
is illustrated in Figure 2.

These tools, plus the UT-Houston Gopher client (Ugopher), are 
available for download in the anonymous FTP archives on: oac.hsc. 
uth.tmc.edu

As the system develops further, new tools will need to be created 
to help users publish with ease. It is our opinion that any new tool 
requiring an entire pamphlet of instructions should not be given to a 
user, but rather thrown into file 13. The Gopher protocol is an 
elegant model of ease and usability that should be copied.

Moving forward

The University's campus-wide information system changes daily and 
will continue to do so. In the eighteen months we have used these 
tools for our CWIS, we have found that interest has blossomed. Small 
classes of ten to twenty employees are held regularly on campus to 
introduce the system. We feel it is important to show the system to 
all levels of employees in the actual environment where they will use 
it. Also, the small classes allow for interaction and questions, 
spawning new interest. We hope with each class given that one or two 
more people will find the system an aid to their jobs, and that they 
will identify more information to include in the CWIS.

The one-on-one work with information providers is still 
necessary. Although it means that the CWIS staff need to learn several 
systems, in this age of rapid technological change, this is seen as an 
asset. Since more and more of our interaction is done via e-mail, 
these meetings allow the OAC staff and the information providers to 
match faces and voices with e-mail addresses. This, we think, greatly 
enhances the team feeling.

We have also added an additional staff person to upload 
information and assist in the day-to-day office work, to make more 
time available for classes, hand-holding, and interest-building.

The UT-Houston campus-wide information system is still growing 
and developing; we are still adding information into the CWIS via 
Gopher and will do so for some time. At the same time, we have already 
begun building a sister system in World Wide Web ... but that is the 
subject of another article.

******************************************************************

UT-H Health Science Center's Guidelines 
for Preparing Text for the CWIS


Avoid the use of tabs; instead use spaces.

ASCII tabs are not converted to a constant number of spaces by 
all display devices. Thus, the use of ASCII tabs often causes text to 
be displayed in a jumbled fashion.


Delete bullets or other special symbols such as Greek characters 
and/or replace with standard ASCII characters.

Character-based terminals can only display ASCII symbols.


Single-space lines of text, but double-space between paragraphs.

This maximizes the amount of text that can be displayed per 
screen while preserving readability.


Use a non-proportional font, and limit line length to a maximum of 78 
characters per line.

If all characters do not take the same amount of space, it is 
difficult to ascertain whether each line has less than 80 characters. 
Lines that exceed the 80-character limit will either wrap improperly 
or run off the screen.


Place headings in capital letters and text in upper and lower case.

This enhances the identification of the beginnings of thoughts 
and divisions of chapters, sub-chapters, etc.


Place identifiers around headings (dashes or equal signs).

This enhances the readability of the material and prepares the 
text for conversion to a WAIS index, if desired.


Prepare text in a neat, orderly, and easy-to-read format for both 
paper and electronic copies.

For first drafts we suggest submitting both a hard copy and an 
electronic copy, in case there are errors in the formatting.

******************************************************************

Beth Ardoin is the Administrative Assistant for the Office of Academic 
Computing at the University of Texas-Houston Health Science Center. 
Her background is in education and publishing. She is editor of the 
campus-wide information system and directs the distribution of 
information on the American Physiological Society Gopher on the 
Internet. She has a bachelor's degree in education and communications 
from the University of Southwestern Louisiana.

William Weems is Director of the Office of Academic Computing (OAC) at 
the University of Texas-Houston Health Science Center. The OAC 
coordinates the academic computing efforts of the six schools that 
make up the University of Texas-Houston. He has played a major role in 
developing the Texas Medical Center network, which includes forty-one 
separate institutions. Dr. Weems holds a Ph.D. in physiology and 
biophysics from the University of Illinois.

************************************************************************