NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/bit/listserv/cdromlan/1674

Internet Message Format | 1992-11-11 | 32.5 KB

Path: sparky!uunet!stanford.edu!bu.edu!wupost!uwm.edu!zaphod.mps.ohio-state.edu!darwin.sura.net!paladin.american.edu!auvm!STMARYTX.BITNET!ACADMARG
From: [e-mail]
Newsgroups: bit.listserv.cdromlan
Subject: Re: Book chapter on network installation
Message-ID: <[e-mail]>
Date: 12 Nov 92 02:02:00 GMT
Sender: "CDROMLAN@IDBSU - Use of CDROM Products in Lan Environments"
              <[e-mail]>
Lines: 712
Comments: Gated by [e-mail]
Original_To:  JNET%"[e-mail]"
Original_cc:  ACADMARG

I have gotten a lot of requests for copies of the book chapter on
installation of our CD-ROM network and several requests that it be
posted to the list.  It is attached below and it is pretty long so
delete now if you don't really want to read about this.  If you do,
here it is!

Margaret




Talking 'Bout a Revolution:

CD-ROM Networking Inside and Outside the Library

with Novell, the Infoserver 150, and the V-server Gateway


by

Margaret Sylvia

  St. Mary's University Academic Library



Introduction

     Installation of our CD-ROM network at St. Mary's University
Academic Library has created a revolution in the way students and
faculty do research.  There are over 1000 uses per week of the
network during the regular semester despite the fact that we are
a small university with about 4000 students overall, 800 graduate
students, and 650 law students.  This usage is only a count of
logins to a particular database since we cannot count the number
of searches done once a database is in use.  Our print indexes
are going relatively untouched except for subjects which don't
have good coverage on the network and librarian-mediated online
searching has dropped off to a trickle.



Background

     St. Mary's University is the oldest and largest Catholic
University in Texas founded in 1852 by Marianist brothers and
priests.  The student body is culturally and ethnically diverse
and the university's efforts help promote  the graduation of
female and minority students.  One of the driving forces behind
the push for a CD-ROM network was the institution of a doctoral
program in Counseling at the University where previously, all
graduate programs had been at the Master's level.  This was the
first Ph.D. program in the city except for the doctorates in
biological sciences awarded by the University of Texas Health
Science Center at San Antonio.

     The Academic Library began adding individual CD-ROM
workstations several years ago to aid students in their research
and these became very popular almost immediately.  During peak
research times, there would often be several students in line
waiting for a CD-ROM workstation even though the equivalent print
indexes were only a few steps away.  Whenever this was pointed
out to these patrons, most often they indicated that they would
rather wait for the workstation to be free than use the print
index.  It became obvious to us all too quickly that using the
CD-ROM indexes was the preferred method of doing research for
journal articles and that there were too few workstations to
handle the volume of research that the students were undertaking.


Alternatives

     We began by researching the various possibilities of
bringing journal indexing to students over a network.  One option
we considered was loading journal index tapes (such as those
available from Wilson) onto our library computer which handles
the online catalog.  But we found that we did not have the size
or power to do this with our current computer and we were unable
to afford to migrate to a larger one at the time.

     We also considered loading the tapes on the campus
mainframe, a VAX 6000 series, but found that purchasing a search
engine to run them was out of our price range.  In addition, we
were hoping to keep as many of our computer resources inhouse as
possible since we have found that when they leave our direct
control, they are often not treated as first priority when
problems arise.  When we discovered that the other options were
impossible to implement, we began to plan very seriously for
building a CD-ROM network.



Goals

     One of our principal goals was that our students at off-
campus sites should be able to dial in to the network in order to
search the CD-ROMs.  We wanted to offer these students the
convenience of doing some of their research from their homes or
offices.  Our online catalog was already available for dial-in
access.   Many of our users are non-traditional and off-campus
students and we strive to give them the same research advantages
as the full-time on-campus students.

     Secondly, we wanted to keep control of the network within
the library.  Some of the options for networking that are
available make use of the campus mainframe as the network server.
Since other departments have other priorities than the library,
we wanted the network to be a top priority for us and the best
way to do that was to keep the fileserver within the library.

     Of course, we wanted the network to be reliable and
affordable.  It had to be easy for the students to use and offer
the ability to limit the number of simultaneous users for
different databases if necessary.  The vendor had to be
reputable, financially sound, and have a good track record of
installation and support.  Also, different CD-ROM search software
has different requirements regarding the installation of DOS
extensions and we wanted the network we purchased to support all
calls to the CD through DOS extensions, directly to the CD, or
any variation of this.

     Next, we wanted the network to be easy to administer.  Since
we would be unable to hire a fulltime system administrator, one
of the librarians would have to take on this job in addition to
other duties.  Regarding dial-in access, we wanted a network
which would accept any sort of hardware dialing-in and which
would not require any sort of special software to be loaded onto
the remote machine.  Anything from a dumb terminal, a Macintosh,
an IBM XT clone, or a 486, any of which might be found in a
campus office, should be able to dial-in and use the same
procedures on the network.  This would ease the problem of
administration of the network and also eliminate political
problems which might be caused by limiting use of the network to
people with certain types of equipment.

     Another goal which would also ease the administration of the
network was to have dial-in access take place through the campus
fiber optic network which was already in place, rather than
installing new modems in the library for this purpose.  The St.
Mary's campus is connected by a campus-wide fiber-optic network.
This procedure would enable the network to have an extra level of
dial-in security since users on the campus network would need a
password assigned by the campus computing center.  In this way,
the library network administrator need not assign individual
passwords and yet remain able to limit dial-in access to our
students and faculty if necessary.

     Finally, we wanted to be sure that all the CD-ROM software
we were using would run on the network and be accessible through
the gateway.  These included PsycLIT and ERIC from Silverplatter,
Periodical Abstracts and ABI/Inform from University Microfilms
International, Government Documents Monthly Catalog from Marcive,
Books in Print Plus from Bowker, and the National Trade Data Bank
from the Government Printing Office.  Since we were aware that
search software can access the CD in different ways and not all
networks support all kind of search software, we wanted to be
sure that different kinds of software would be supported.  In
this way we could be sure that not only would our current
subscriptions work but also all future items we might want to add
to the network would be usable.



Network Affordability

     The choice of CD's we placed on the network was basically
driven by what we had already chosen for our standalone
workstations.  With the exception of Infotrac's Academic Index,
we retained all our subscriptions and simply upgraded to a
network version.  We considered attempting to consolidate all our
subscriptions with a single vendor in order to reduce the number
of search interfaces users had to deal with, however this was not
possible.  Certain databases, such as Books in Print, were only
available from a single source and no suitable substitutes were
available.

     One major ongoing expense of a network is the subscription
price for the CD-ROMs themselves.  One way that we approached
this problem was to cancel a very expensive CD-ROM index
(Infotrac Academic Index) for which we had two individual
subscriptions.  We redirected that money to finance network
subscriptions to our remaining CDs and to purchase a network
subscription to the less expensive Periodical Abstracts from UMI
as a replacement for the Academic Index.  This strategy enabled
us to stay within our CD-ROM budget and still purchase needed
network subscriptions. Though Infotrac Academic Index was popular
with our students, Periodical Abstracts has served the same
purpose and is easier to use in our networking situation.   Also,
the interface is identical to that of ABI/Inform which we already
owned.  This meant that in discarding Infotrac we had fewer
search interfaces to deal with in teaching the use of the
network.

     In the earliest days of networking CD's, prices and
contracts were often negotiated individually by each customer.
When we went to our vendors, however, the contract and pricing
agreements were already in place and we had only to decide what
we could afford.  Some items, such as the National Trade
Databank, from the U.S. government, have no restrictions on
networking.  Others, such as PsycLIT from Silverplatter, are very
tightly controlled and many options are non-negotiable.

     Another big expense involved in building a network is the
outlay for hardware.  In our case, the library director wrote a
grant for approximately $50,000 which was funded by the George W.
Brackenridge Foundation in San Antonio.  This generous foundation
has worked with us on several grants to improve access to
information in the library and we are very proud of the progress
we have been able to make with their help.  Hardware maintenance
cost is the final factor in this equation.  We requested free
maintenance from our vendors for the remainder of our fiscal year
and for next year the cost will be added to our regular computer
maintenance budget.

     Preliminary planning for the network was done during the
grant writing process, however, the actual work of choosing a
particular vendor and configuration was not accomplished until
after the funding had been received.  Personnel involved in this
task included the director of the library and myself, the future
systems administrator.  Advice was also sought from our small
computer expert and others in the campus computer center, however
they had no vote in the actual choice of vendor or system.   That
decision was made by me with the concurrence of the library
director.



Choosing the Networking Hardware and Software

     Since our first goal was to have dial-in access to the
network, we looked first for hardware and software which would
deliver this service.  We wanted searchers with any kind of
hardware to be able to dial in to the network.   Secondly, we
wanted searchers to be able to use the network in pretty much the
same way an inhouse user would, without having to load the search
software on their machines and without doing any remapping of
their keyboards.

     In looking for the right kind of hardware and software to
meet our needs, I went through a learning process.  I was already
fairly familiar with CD-ROMs and how they worked on an individual
workstation but I had never set up a network before and I was not
familiar with all that was involved.  One of the first things I
did was to join a LISTSERV group on BITNET called CDROM-LAN.
This is a group of system administrators of CD-ROM networks who
share their thoughts and concerns with each other by electronic
mail.  This gave me a feeling for the views of people who already
owned and administrated these networks.  I was also able to use
the group as a sounding board for my questions and decisions on
various options.   Of all the information gathering I did on
networking, the most crucial information came from this group
since they were able to tell me what really worked and what did
not.

     After looking at the possibilities of various kinds of
networking options, including those from Meridian, CBIS, Ebsco,
Lantastic, Optinet, SCSI Express, Logicraft, Multiplatter and
others, we eventually discovered the choices which seemed to meet
all our needs as outlined above.  Virtual Microsystems produces
the V-server gateway which works with the DEC Infoserver, and
Gandalf offers the Starport server running Meridian software.
Either of these options seemed to fit well with our goals and
needs, particularly the goal of making the network available to
off-campus users.

     We requested and received permission to dial into systems
which had been set up by each of the companies in order to test
the systems and their response times.  The DEC Infoserver system
responded significantly faster in this test and, because their
bid was also somewhat lower in total cost, we chose Virtual
Microsystems as our vendor.



Network Configuration

     Originally the DEC Infoserver ran only on Pathworks but now
runs on Novell Netware also.  We chose to run Novell since we
already had another Novell network in the library thus giving us
inhouse expertise to exploit.  Also, it is possible that in the
future we may want to connect the two networks and this would be
simpler if both were running on the same operating system.

     Since Virtual Microsystems installs only the DEC Infoserver
and the V-server gateway, we used a local vendor to set up our
Novell network in the library.  Taking advice from the CDROM-LAN
group, I chose 10baseT unshielded twisted pair cabling and a star
configuration for the network inside the library in order to
reduce both the cabling cost and the difficulty of locating and
repairing network wiring problems.  The 10baseT cable is very
inexpensive so even though a star configuration requires more
wire, we saved money over using ethernet cable in a bus
configuration.  The star configuration makes it easier to
determine which connector is at fault when a workstation fails to
connect to the network.  This configuration also makes it easy to
take a workstation offline and to add other workstations without
interrupting the workflow on the network.

    For the remaining hardware needed, I chose 8-bit DEC DE101-AA
network cards, monochrome monitors, and 286 clones with 20
megabyte hard drives for the workstations.  If there had been
sufficient funds, I would have picked 16-bit cards, color
monitors, and 386's.  I also purchased a 386 with a 105 megabyte
hard drive for the Novell fileserver though I would have
preferred a 486 with a larger hard drive, possibly even dual hard
drives to allow mirroring.  However, price considerations entered
into the picture.  We have since upgraded the 286's to 386SX's in
order to take advantage of the 386's improved memory
capabilities.

    One item which was not purchased with our original network
order but which should be purchased at the earliest opportunity
is an uninterruptible power supply for the fileserver.  The
Novell vendor did not recommend this to us initially though I
wish he had.   We found that it is really a necessity since
Novell crashes in case of even a momentary power failure.  An
uninterruptible power supply allows Novell to shut down normally
in case of a power failure and also helps smooth out power surges
and dips.  An uninterruptible power supply has recently been
added to our system so that we no longer worry about even
momentary power outages.

     For network printing, we reused the original printers which
came from our individual CD-ROM workstations.  Electronic
switchboxes were purchased and three printers are shared by nine
public workstations.  Another item on my wish list is new
printers for the network.  The printers we have now are older and
are beginning to show the strain of constant use.




Placement of the Network in the Library

     Where to place the network workstations in the library was
the object of some discussion among the librarians.   Our
individual CD-ROM workstations had been placed on tables beside
the wall only a few feet from the reference desk alongside the
paper indexes.  This arrangement allowed the reference librarian
on duty to monitor their use and help users with problems in
searching without leaving the reference area.  The setup was a
useful one but unfortunately, there was not enough room in the
area for the extra workstations we wanted to add for the network.


     Discussion centered on the possibility of locating the
network workstations in the Reference area by rearranging the
index tables or using a separate conference room on the main
floor.  The advantages of using the conference room for the
network stations were that it was already wired electrically for
the appropriate number of computers, we could do bibliographic
instruction there using a computer-projection system, and the
door could be closed to keep the main floor quieter since the
printers and bibliographic instruction are somewhat noisy.  A
disadvantage of using the separate room were that the
workstations would be out of sight of the reference desk.  Our
reference librarians found that much of their reference work
involved helping users with search strategy and choosing the
appropriate database for searching.

   The network was finally placed on the main floor within the
reference and index area.  The reference desk and index tables
were rearranged to make room for the new workstations.   The need
to have a reference librarian on hand to oversee operations
outweighed the other advantages of having the network in a
separate room.  In fact, having the network right in the middle
of the main floor adjacent to the index tables, the reference
desk, and the online catalog is really the optimal place for it.
Students are drawn to it as they enter the library and begin
experimenting with it even before they are quite sure what it is.

     The fileserver and CD drives are not located in the public
area with the workstations but are located in the technical
services area of the library about 20 yards away.  There was also
some thought given to placing the fileserver in the public area
but in a locked cabinet.  However the amount of heat generated by
the fileserver, the Infoserver, the V-server gateway, and the CD-
ROM towers would have required lots of ventilation, perhaps even
a small air conditioner in the cabinet.



Network Installation

     Installation of the network was scheduled to take place
during the interterm between the fall and spring semesters to
allow for the least disruption of student and faculty library
research activities.  After deciding on the placement, the type,
and number of terminals and printers we could afford,  we used
our on-campus physical plant electricians to rewire the area and
install enough outlets for all the equipment.   An independent
contractor was hired to pull the 10baseT wiring for the network
itself.  The workstations were set up with their network cards
and DEC Repeaters were used as hubs to bring all the 10baseT
wiring together and into the fileserver.

     The Novell license and the Novell fileserver were purchased
from and installed by a local vendor who has a contract with the
University to supply us with microcomputers.   Finally, Virtual
Microsystems arrived to install the Dec Infoserver with its CD-
ROM towers of drives and the V-Server gateway.

     This seems to imply that all went smoothly in the
installation, when of course it did not.  The original wiring
contractor did not get all the 10baseT wiring installed in time
and never got the ethernet cable pulled to the computer center
which was to serve the gateway.  We did not have the right
network card for the Novell fileserver so we were still on the
phone with this problem while Virtual was trying to install the
Infoserver and CD drives.

     Books in Print Plus was late sending their networking
software so this was not installed by Virtual Microsystems along
with all the other software which went in quite easily.  Of
course, BIP+ turned out to be extremely tricky to install so I
spent many hours on the phone to Bowker and Virtual trying to
work this out.  We eventually had to upgrade all our workstations
to 386SX's in order to meet the memory requirements of BIP+ on
the network.

     The menuing software which Virtual sold and installed for us
worked fine except that they installed it with all users having
the same name and rights (supervisory level).  This worked for
counting usage but did not work when we wanted to limit the
number of simultaneous users of certain CD's.  I had to
reconfigure the menuing software (Direct Access) to give each
workstation a different user name and to limit supervisor rights.
Due to many problems with wiring between the buildings, it took
five months to get our ethernet line run to the computer center
to complete the gateway service to remote users.  Everything else
went in fairly easily and, except for the ethernet cabling and
gateway problems, these were fairly minor things.  My nightmare
was the idea that we had forgotten to purchase some important
item which would cost about $10,000 and without which nothing
else would work.

     After we finally had the ethernet line run to connect our
gateway to the campus network we were able to have the gateway
installed.  This was probably the most harrowing experience of
all.  A technician from Virtual Microsystems made a second trip
to the campus to complete our installation, however, he was not
able to get all our databases to work through the gateway.  At
the end of the day, he promised that the remainder of the
installation could easily be finished by them during a dial-in
session in the next few days.

     In fact, this was a real problem.  The first technician to
dial in not only failed to fix any of the gateway problems but he
crashed the in-house network by changing important software
parameters without telling us.  Next, one of the dial-in ports on
the gateway failed.  Unfortunately, it happened to be the first
in line and this caused all ports to fail.  After much
telephoning and still further dialing-in by Virtual, it was
determined that they needed to finish the installation on-site.

     Another trip by a technician from Virtual Microsystems to
the campus solved most of the remaining problems except for a
few.  We are still working with a problem with the Silverplatter
software in that it writes large temporary files to the virtual
disk as searching progresses and sometimes this causes the user
to get a "diskquota exceeded" message and to be kicked off the
system.  Even giving users a fairly large diskquota does not seem
to solve the problem.  One solution we are considering is to run
the gateway software directly from the Infoserver 150 instead of
from our campus VAX so that the diskquota can be turned off on
the Infoserver.  The other problem with the V-server is that the
most current version of searching software from University
Microfilms International does not work through the gateway and we
are having to run an older version for this reason.  We have been
promised that the next version of UMI software due out in
December 1992 will work with the system.

     A final problem with the V-server gateway was that the
telecommunications software we were using as the campus standard
(Procomm) does not have a true VT terminal emulation so the
functions keys fail to perform correctly on the gateway.  After
much discussion with Virtual Microsystems about this and after I
had been assured many times during the purchase of the gateway
that Procomm would work with it, it was eventually determined
that Procomm would not work properly.  I was rescued by another
gateway administrator from the CDROM-LAN group who sent me a
specially tailored Kermit program.  Since Kermit is in the public
domain we are able to distribute this to our users for gateway
access.



Guides to Use of the Network

     One thing that concerned us about having off-campus users on
the network was that we would have no way of training them to
begin searching.  We decided to make as much of our training
material available online as we could.  We already had short
guides to most of our databases which only needed minor updating
to handle variations such as new printing procedures on the
network.  Other guides needed to be produced, such as one for
Books in Print Plus and Periodical Abstracts, since we had never
had these available to the public before.

     For inhouse use, we placed copies of all guides in clear
plastic pocket pages and made duplicate notebooks of these for
placement next to each workstation inside the library.   In
addition, all guides were made available online on the network
for the use of those patrons who would be dialing in from remote
locations.  There is also an option on the Direct Access network
menuing software for online help for each menu item.  Context
sensitive help screens were devised for each database.

     Finally, we received interactive tutorials from UMI and
Silverplatter which we loaded on the network.  Some of these
training aids and guides have gotten fairly heavy use and others
have not.  In the future we plan on taking a look at some of the
more effective and useful ways of training people to search on
the network.



Remote Use

     Remote use of the network is possible through our campus-
wide network.  Due to the problem with our regular
telecommunications software program Procomm noted above, users
must obtain a customized copy of Kermit from the library in order
to use the network properly on a PC from a remote location.
Users dial into the campus network, login normally, activate the
Kermit program, and proceed to a menu on which one of the options
is the CD-ROM network.   From this point, remote network use is
no different from network use in the library except that searches
which are downloaded are placed in the user's VAX account instead
of on their local hard disk.  Users can log sessions to their
local hard disks with Kermit or print searches if a printer is
attached to their workstation.

     One problem that we did not anticipate is that all students
do not automatically receive an account to use the campus
network.  At this time, professors must request that their
classes receive computer accounts.  We are working to get all
students assigned a computer account at registration.



Usage and Impact

     Usage patterns for the network reflects the usage we had
seen on standalone workstations.  Our general database
(Periodical Abstracts) is by far the most heavily used, PsycLIT
and ABI/Inform are the next most heavily used, followed by the
National Trade Databank, ERIC, Marcive Government Documents
Catalog, and Books in Print Plus.   Mediated online searching has
fallen off dramatically since the installation of the network and
our paper indexes related to these items are much less used now.
Interlibrary loan use had risen somewhat though we suspect this
may be due in part to a large influx of Ph.D. students rather
than simply the installation of the CD-ROM network.  Reference
work is beginning to consist largely of teaching users about the
contents and use of databases on the network.

     The gateway has had a greater impact on the computer center
than we anticipated since formerly fewer students were interested
in having a computer account and in dialing into the campus
network.  Plans are being made in the computer center for
expanding the main campus VAX and for generating student computer
accounts automatically upon registration.



What's Next

    Future automation plans in the library include the
installation of Lexis/Nexis for student and faculty use.
Lexis/Nexis is an online fulltext database containing current
news, business, legal and other information.  This service will
expand our users' access to current information tremendously and
it is available for a relatively small fee in educational
institutions.

     We plan to investigate the purchase of other CD-ROM indexes
and fulltext databases.  Currently, we are using Periodical
Abstracts, ERIC, PsycLIT, ABI/Inform, Books in Print Plus,
Marcive Government Documents Catalog, and the National Trade
Databank.  A religion index, an index to literary criticisms, an
engineering index, and perhaps a fulltext local newspaper or
newspaper index would be welcome additions.

     For bibliographic instruction on the network, we would like
to purchase a computer projection system and a portable PC with a
CD-ROM drive which could be used to instruct remote users at
their sites.   Equipping this PC with a modem would also make it
possible to display the remote use of the system if a phone line
hookup is active at the remote site.



Summary

     Before installing a CD-ROM network, the need for the network
and goals to be accomplished should be established.   The
hardware and software chosen for the network should be able to
meet the established goals and criteria for reliability,
affordability, and ease of use.  Careful consideration must be
given to placement of the network workstations within the library
so that the workstations are accessible to the users and visible
to the reference librarians, while the fileserver and CD drives
must be in a secure location.   Installation should be scheduled
so that the full attention of the systems administrator can be
given to installation and so that disruption of services will be
kept to a minimum.   Teaching aids and bibliographic instruction
must be planned with all users in mind, including those in other
locations if remote use is a possibility.










-----------------------------------------------------------------

Biographical Information

     Margaret Sylvia is an Assistant Professor at St. Mary's
University Academic Library in San Antonio, Texas.  She has been
with the library since 1989 as the Acquisitions & Collection
Development Department Head and since 1991 has also taken on the
responsibilities of Local Area Network Systems Administrator.
Ms. Sylvia has a B.S. in Psychology from Texas A&M University and
an M.L.S. from the University of Texas at Austin.  Her research
interests include library automation, particularly in the area of
CD-ROM technology and also in the area of library acquisitions.

-----------------------------------------------------------------



Margaret Sylvia
Acquisitions & Collection Development Department Head
Local Area Network Systems Administrator

St. Mary's University Academic Library
One Camino Santa Maria
San Antonio, Texas 78228-8608

voice: [phone]
fax:   [phone]

Bitnet: [e-mail]
Internet: [e-mail]








                            Bibliography

Cutright, Patricia & Kenneth M. Girrard. "Remote Access to CD-ROM
     for the Distance Learner."  CD-ROM Professional 4:6 (Nov.
     1991), 80-82.

Desmarais, Norman. CD-ROM Local Area Networks: A User's Guide.
     Westport: Meckler, 1991.

Jaffe, Lee David & Steven G. Watkins.  "CD-ROM Hardware
     Configurations: Selection and Design." CD-ROM Professional
     5:1 (Jan. 1992), 62-68.

Jensen, Mary Brandt.  "CD-ROM Licenses: What's in the Fine or
     Nonexistent Print May Surprise You."  CD-ROM Professional
     4:2 (Mar 1991), 13-16.

LaGuardia, Cheryl, Churck Huber & Carol Doyle.  "CD-ROM
     Networking in ARL Academic Libraries: A Survey."  CD-ROM
     Professional 4:2 (Mar. 1991), 36-39.

McCusker, Tom.  "Now Playing: Compact Disks on LANs." Datamation
     36:20 (Oct. 15, 1990), 55-56.

McQueen, Howard. "Networking CD-ROMs: Implementation
     Considerations." Laserdisk Professional 3:2 (Mar. 1990), 13-
     16.

McQueen, Howard. "Remote Dial-In Patron Access to CD-ROM LANs."
     CD-ROM Professional 3:4 (July 1990), 20-23.

Morgan, James Jay. "Expansion and Testing of a Meridian CD-ROM
     Network."  Public-Access Computer Systems Review 1:1 (1990),
     34-42.

Musgrave, Bill "Network Management: Keeping the Connection."
     Datamation 33:17 (Sept. 1, 1987), 98-107.

Perratore, Ed. "Networking CD-ROMs:  The Power of Shared Access."
     PC Magazine (Dec. 31, 1991), 333-363.

Rosen, Linda. "CD-Networks and CD-ROM:  Distributing Data on
     Disk." Online 14:4 (July 1990), 102-105.

Rutherford, John. "Improving CD-ROM Management through
     Networking." CD-ROM Professional 3:5 (Sept. 1990), 20-27.

Smith, Steve. "A CD-ROM LAN Utilizing the CBIS CD Connection
     System." Public-Access Computer Systems Review 1:2 (1990),
     50-61.

Wilson, Thomas C. "Zen and the Art of CD-ROM Network License
     Negotiation." Public-Access Computer Systems Review 1:2
     (1990), 4-14.