robertj@tekgen.BV.TEK.COM (Robert Jaquiss) (11/29/89)
SPECIAL INTEREST GROUP ON CD-ROM
APPLICATIONS & TECHNOLOGY
October 24, 1989 Meeting
1. Introduction
This month's meeting was held at the Washington Convention
Center in conjunction with the Federal Computer Conference.
Jerry McFaul brought the meeting to order, and declared the theme
of the meeting to be "Exploring CD-ROM Options." It is important
for users to be aware of the broad spectrum of options available
for developing and distributing information on CD-ROM. It is
also important that the vendor community understand the needs and
desires of the users. Those present were invited to sign up as
member of SIGCAT, if they had not already done so. SIGCAT meets
every other month; the next meeting is scheduled for December 7
and will be held at the U.S. Geological Survey, in Reston, VA.
2. A Perspective on CD-ROM
The first guest speaker was Chris Andrews, President of
UniDisc, Incorporated. Mr. Andrews has been in the electronic
information industry for ten years. Based on his own experience,
Mr. Andrews summarized the history of CD-ROM and presented his
forecasts of future trends.
CD-ROM was an outgrowth of the publication of recorded music
on Compact Disc. Library applications were the first to be put
on CD-ROM. Because of this association with bibliographic
indexes and catalogues, CD-ROM became type-cast for static
databases. The essence of CD-ROM is that it is another
information distribution standard, just as the book is. While a
book is designed to be used linearly, from front to back, CD-ROM
can be used like an online retrieval system, where information
can be accessed by content.
The major area of success for CD-ROM has been in cutting the
cost of printing. If a 300 page document, such as a reference
manual, must be distributed to some 200 users, not all of the
users will need all of the document. Some users will only need a
few pages, some will need more. Typically, from 30% to 40% of
the document is excess, which only gets in the user's way. If
the same document is published on CD-ROM, then the user only
needs to print the pages he uses regularly. This is known as on-
demand printing. Other areas of success have been (a)
distributing information quickly, (b) promoting control and
standardization, and (c) selling things.
Mr. Andrews gave the following perspective on the costs of
CD-ROM compared to other distribution media:
Cost/
Medium Mbyte
CD-ROM <$0.01
paper $4.00
fiche $0.10
35mm. slide $1.25
floppy disk $2.00
WORM $0.10
The above comparison is only for the cost of replication, with
the cost of replicating a CD-ROM disc given as $2.00. The other
part of the cost perspective is that the start-up costs for CD-
ROM publishing are high. Pre-mastering software and work-
station will cost about $100,000, and can run much higher. The
keys to cost-effective CD-ROM publishing are: (a) high frequency
of access, (b) high volume of data, and (c) a large number of
users. It is also important to look critically at each step of
the preparation process and to cost it out independently. Base
in-house preparation versus contracting out decisions should on
cost.
Future trends in will see CD-ROM publication as a hybrid of
book and music publication. CD-ROM drives will get less
expensive (around $500) as local area networks compete with
stand-alone drives. Newer technology will allow publishers to
master CD-ROMs in house. Multi-media ROM will become
increasingly important, but will add to the cost of the user
work-station.
Mr. Andrews concluded by describing UniDisc and its current
projects. UniDisc Incorporated is a consulting firm which serves
as an intermediary between the CD-ROM publisher, and the CD-ROM
hardware and services vendors. One of UniDisc's current projects
is the "Guiness Disc of Records." It is a fun, educational tool
containing 300 color slides, 30 minutes of sound, 20 animations,
and 5000 records. Anyone interested in obtaining a copy should
call (800) 326-0092, Extension 42. Another UniDisc product is a
CD-ROM disc sharing service called Alpha ROM. Publishers can
have their files mastered on an Alpha ROM disc at the rate of $90
for 100 pages, or 25 pages for $25. Please, no confidential
data. For further information call:
Mr. Chris Andrews, President
UniDisc Incorporated
(415) 969-0606
or (415) 962-0577
3. Five Lessons in Five Years
Mr. William Paisley, Executive Vice President of Knowledge
Access International, related the lessons of five years
experience in developing and distributing CD-ROMs. The book will
be the model for CD-ROM publishing in the future. Designing a
CD-ROM, which is analogous to preparing a draft of a book, is
extremely challenging. Production, on the other hand, is very
uniform. Standard solutions are the secret to developing
inexpensive CD-ROMs.
Publication of a CD-ROM starts with the conversion of
information to machine-readable form. If the information is not
already in machine-readable form, conversion will be the most
expensive part of the publishing process. The next step is to
edit and massage the data to create the indexes, do the hypertext
mark-up, or the editorial mark-up. The Standardized Generalized
Mark-up Language (SGML) is recommended for editorial mark-up. It
is not the least expensive choice, but it is a safe one.
Pre-mastering is two processes in one. First a master image
of the CD-ROM is created, and then it is simulated on magnetic
disk to evaluate performance. Pre-mastering may either be done
in-house or out-of-house. The publisher has better editorial
control over the final product if he chooses the former. CD-ROM
mastering is usually done out-of-house. It includes not only
making the master disc, but also replication and packaging.
Allocate resources for user support. This is part of the product
too.
A publisher's work-station must consist of at least an Intel
80386 class PC, a large hard disk, and a tape unit. The tape
unit is usually a one-half inch industry standard tape, but one-
quarter inch cartridge tape can also be used. The publisher
should produce at least one CD-ROM per month to justify the cost
of the work-station. The CD-ROM user's minimal configuration,
plain vanilla work-station should consist of an IBM PC (or
compatible) with a 10 megabyte hard disk.
Mr. Paisley closed with the following lessons:
1. CD-ROMs are easy to produce, so don't get dazzled by
CD-ROM mystique.
2. A CD-ROM is produced to be used; if it takes six months
to produce it, it is already a failure.
3. CD-ROM design is an art of trade-offs; make it good, not
perfect.
4. Don't overdesign the CD-ROM; cost out the options at each
stage, use standard solutions, and avoid reaching for the
last ten percent of the design.
5. Don't overinvest in publishing capacity.
Mr. Paisley described the KAware Disk Publisher. This is a
software system for the development of information products for
distribution on CD-ROM. It accepts text in all the popular PC
and mainframe file formats; supports field-specific and free-text
search and retrieval; has hypertext capability; and can combine
both text and graphics in the final product. Mr. Paisley made a
special introductory offer of the KAware Disk Publisher/Text-
Image package to SIGCAT members at a price of $399 less an
additional discount of $50 to conference attenders. For further
information call:
Knowledge Access International, Inc.
2685 Marine Way, Suite 1305
Mountain View, CA 94043
Telephone: (415) 969-0606
FAX: (415) 946-2027
4. Optical Publishing vs. Other Database Techniques
Mr. Tom Brown, Project Engineer of American Helix Technology
Corporation, described optical publishing versus other database
techniques, and introduced LASERTEX, a multimedia authoring and
retrieval package. A centralized database, such as a corporate
accounting database, is built by the specialists who use it.
Optical publishing is analogous to a company's annual report. It
contains text, numeric tables, and illustration all in one
package, and is intended to be used by outsiders.
In building an application using LASERTEX, the first step is
to design a storyboard prototype. This is where the user
interface is created. Next, the information is collected,
prepared, and catalogued. Indexes, scripts, and hypertext links
are also created.
Mr. Brown used, as an example, a real estate multiple
listing database. The data consist of textual description of
each property, a photograph, and numeric data (e.g. number of
room, square footage, price). The data fields and record formats
are designed first. Then, the user interface is created with
various scripts, i.e. sequences of user actions. Scripts can be
linear such as browse and search scripts, or branching, as in
hypertext scripts. Next, come the design of interfaces for the
various devices that will be used in the real estate application.
These may include videodisc, modem, fax machine, or printers.
Information collection starts with the text data. It is
converted to machine readable form, and SGML tags are added.
Polaroid pictures of the houses are scanned, and Multiple Listing
Service (MLS) numbers are attached to each of the digitized
images. Next, the information is organized. The MLS numbers
must be added to the text to link images, data, and text
together. Also, indexes must be created. This step is extremely
important, but time consuming. It can be speeded up with the
LASERTEX Index Editor.
American Helix made the following offer, which is only good
until November 10, 1989: for an evaluation fee of $4995, SIGCAT
members may buy the LASERTEX Evaluation Program. This includes
the use of the full LASERTEX software to develop one CD-ROM
application, five days of intensive training on the use of the
software, design support and consultation services, specialized
information preparation services, and free replication of the
master disc. Mastering costs are not included. If you should
decide to buy the LASERTEX program, American Helix will give you
full credit for the evaluation fee toward the purchase price, and
a discount as well. For further information on this offer, or
for a copy the LASERTEX demonstration CD-ROM, call or write:
American Helix Technology Corporation
1857 Colonial Village Lane
Lancaster, PA 17601
Telephone: (800) 525-6575
5. A Macintosh View
Allen Adkins, President of Optical Media International,
spoke on CD-ROM technology from the point of view of the Apple
Macintosh user. First of all, support of the ISO-9660 standard
is crucial to the Macintosh user because it permits CD-ROMs
developed for the PC to be used on the Macintosh, and vice-
versa. Whether you are a Mac or a PC user, it is important to
realize that CD-ROMs are asymmetrical. There is a long
preparation time involved in designing and publishing a CD-ROM
that the user never sees. For the user, CD-ROM is read-only; the
publisher is the one who writes it. This means the CD-ROM is a
fast, low cost, standardized publication and distribution medium.
In contrast WORM (Write Once Read Mostly) technology is
expensive, unstandardized, and unsuited for mass distribution of
information. Prices of CD-ROM hardware and software will
continue to drop as more applications are developed. UNIX
applications software and documentation files will soon be
published on CD-ROM.
Mr. Adkins demonstrated MacTOPiX, a software package for
pre-mastering a CD-ROM on the Apple Macintosh. A PC version is
also available. For further information, call:
Mr. Allen Adkins, President
Optical Media International
Telephone: (408) 395-4332
6. In-House Preparation vs. Service Bureau
Mr. Don Rodgers, Marketing Programs Manager for Reference
Technology Incorporated (RTI), discussed the trade-offs involved
in deciding whether to prepare CD-ROM products in-house, or to
let a contractor do the "dirty work." Mr. Rodgers began by
saying that RTI has had eight years of experience in selling
software and hardware for CD-ROM applications, in project
management for service bureau applications, in developing systems
products for bringing CD-ROM production in house, and has done
over 400 CD-ROM titles for government and private industry.
Also, RTI staff members headed up the High Sierra Group which
created the industry standards for CD-ROM.
The first decision-making rule is to be objective, and cost
out the alternatives at each step of the way. What steps are you
willing to perform, and which ones should be contracted out? The
following table shows the relative magnitude of the tasks
involved in producing a CD-ROM:
Task Time Required
Data Capture 1000 hours
Data Conversion 100 hours
Data Indexing 10 hours
Logical Formatting 1 hour
Data capture, i.e. conversion of data to a machine-readable form,
is generally the most difficult and expensive step of the
process. Even if you elect to do this in-house, turn over a
sample of the data to a service bureau just as a "sanity check."
Data conversion is the conversion of data already in machine-
readable form from one format to another. It is usually
necessary because all of the data that needs to be in the product
is never in one file. Expect lots of mistakes and problems here.
Record formats tend to change over time so that all the records
in a file never have exactly the same format. This can lead to
some real "horror stories."
The principal advantage in allowing a service bureau to do a
task is that it is their staff and their equipment rather than
yours that is being utilized. Their staff usually has some
experience in CD-ROM preparation and so there is a shorter
learning curve to climb before the product can be released.
However, check with previous customers to be sure that the
contractor in question sticks by his schedules. Service bureaus
maintain a battery of data preparation tools, and have peak load
capabilities that make them a cost-effective way to start out in
CD-ROM publishing. For a single product, it is cheaper to
contract it out than to buy the equipment and to hire and train
the staff. A service bureau may be cheaper for on-going jobs as
well.
Migrate to an in-house facility whenever it makes good
business sense. In house, you control the process and the
schedule. On-going costs are lower, and all steps are carried
out in one location, which is crucial if security is a concern.
In-house operations demand lots of disk space, flexibility, and
expandability. For this reason, only buy equipment that has
standard interfaces, and only use software that conforms to the
High Sierra Group (HSG) and ISO-9669 standards as well. In-house
systems should support CD-ROM emulation of the disk images
produced on it. An emulation debugger and optimization tools are
also essential.
Mr. Rodgers closed with the following advice:
1. A variety of systems are available now with different
capabilities.
2. Make decisions based on immediate needs, but plan for the
future too.
3. An open architecture is important for configurability and
growth; don't design it out.
4. Complex applications can benefit from a more complete
development tool kit.
5. In-house development and service bureaus are
complementary; base your decision on what your
organization can and wants to do.
For further information on RTI, their products and services,
call or write:
Reference Technology Reference Technology
5700 Flatiron Parkway 8150 Leesburg Pike
Boulder, CO 80301 Suite 700
Phone: (303) 449-4157 Vienna, VA 22182
Phone: (703) 883-8215
7. SIGCAT Working Groups
After lunch, there were reports from the SIGCAT working
groups.
a. CD-ROM Data Origination Working Group
Lee Brotzman, NASA Goddard Space Flight Center, spoke
for the CD-ROM Data Origination Working Group. The focus of
this group is the CD-ROM pre-mastering work-station that is
commonly used to develop and test CD-ROM products and software.
Membership is open to all federal managers and technician who
work with or are planning to purchase a pre-mastering work-
station. Some topics of current interest are: a) extended
attribute records for describing MS-DOS files, b) text editors
for large files, and c) addition of 8mm. tapes to work-
stations. For more information about the Data Origination
Working Group, call:
Lee Brotzman
NASA Goddard Space Flight Center
Code 630.3
Greenbelt, MD 20771
Phone: (301) 286-6953
b. ISO 9660 Working Group
Barbara Rose reported on the ISO 9660 Working Group
which is concerned with the promotion and use of the ISO 9660
standard for CD-ROM. At a recent meeting, there was a
discussion of problems related to the VAX/VMS system. Data
originated on a VAX and then transferred to a CD-Publisher for
premastering, could not be read back on the VAX. The question
is why? For more information on this working group, call:
Barbara Rose
Phone: (202) 355-3029
c. Common Indexing Standards Working Group
Laurie Amichetti reported on the Common Indexing
Standards Working Group which is trying to develop a standard
indexing format for CD-ROM text applications. A common
indexing standard would separate the retrieval software from
the pre-mastering software. At present, the two are linked
together. If there were a common indexing standard, CD-ROM
publishers would only have to distribute a minimal retrieval
program with their product or, no software at all. The CD-ROM
user could develop or buy retrieval software to suit their own
needs. They would no longer need a separate retrieval program
for each product. The Air Force is working on the requirements
analysis for such a standard. If you are interested, or would
like to take part, call or write:
Lt. Shenkin or Ms Laurie Amichetti
ESD/AVSI
Bldg. 1704, Room 206
Hanscom AFB, MA 01731-5000
Telephone: (617) 377-2105
d. SIGLIT
Carol Bursick reported for the Special Interest Group on
Library and Information Technology (SIGLIT). This is a three-
year old, ad hoc group of Federal librarians who are interested
in CD-ROM and related technology for libraries. They meet each
month, usually at the Library of Congress in downtown D.C., to
exchange information on CD-ROM. Field trips are a regular part
of the agenda for this working group. If you enjoy travel, and
would like to come along, call:
Ms. Susan David
Library of Congress
Telephone: (202) 707-6447
8. CD-ROM Drives and Their Components
Mr. Carl Beckman, Washington Metropolitan Area
representative for the Hitachi Sales Corporation of America,
described CD-ROM drive components, and several of the drive that
are currently manufactured by Hitachi. Components are a) an
interface card, b) a cable, c) a CD-ROM caddie, and d) Microsoft
Extensions. The interface card is a printed circuit board that
goes in a slot in the back of the PC. The 5A interface card is
used in the PC/XT/AT class of machines and is relatively
inexpensive. The interface card for the PS/2 with its
Microchannel Architecture (MCA) costs about three times as much.
The cable connects the external CD-ROM drive to the interface
card, and should be about five feet long. The CD-ROM caddie
developed by Sony is already a de facto standard and will
eventually become an industry standard. There are strong reasons
for this. CD-ROMs need protection, the caddies make a handy
storage device, and they can also be used as mailers.
Microsoft Extensions are driver routines that allow MS-DOS
to access a CD-ROM drive. The Extensions are make and model
specific, and will only work with the CD-ROM drive for which they
were designed. When ordering the Extensions, state the make and
model of the CD-ROM drive, whether you want them on 5.25 or 3.5
inch diskettes, and lastly, whether your PC has Industry Standard
Architecture (ISA) or MSA.
Hitachi manufactures a variety of CD-ROM drive, both
internal and external, single and multiple units. All Hitachi
drives are HSG and ISO 9660 compatible, are front loading, and
can be daisy-chained. The newer drives have additional features
such as double doors to keep out dust, a power-off ejection
button, and automatic lens cleaning. Double and quadruple disc
drives are popular with libraries.
Mr. Beckman explained why a CD-ROM drive costs two or three
times as much as a CD audio drive. About 30 million PCs have
been sold in the United States alone, while only 250,000 CD-ROM
drives have been sold world-wide. Manufacturers have not yet
recovered their capital investment in CD-ROM manufacturing plant
and equipment. Furthermore, CD-ROM drives require much more
hand-holding that do CD audio drives. As more and more drives
are sold, the price will come down.
Mr. Beckman made two offers to the audience. The first was
a limited time offer to sell the Hitachi Model 1503 CD-ROM drive
to any SIGCAT member for $449. To qualify, it must say SIGCAT on
the purchase order. The second offer was to provide a new
version of Microsoft Extensions to anyone sending in the floppy
disk with the old version. It must be the original diskette sold
by Microsoft, or they cannot make the swap. Send the diskette
to:
Hitachi Sales Corporation of America
31 Water Mill Lane
Great Neck, NY 11021
ATTN: Carl Beckman
Telephone: (202) 249-8266
9. Quality Assurance in CD-ROM
Mr. John Sands, of Nimbus Information Systems, told what
measures can be taken to assure the manufacture and delivery of a
highly reliable CD-ROMs. CD-ROM technology is based on the
"Yellow Book" CD audio standard which is accepted by all CD
manufactures (e.g. Phillips, Sony, Hitachi) world wide. Because
of this, manufacture of CD-ROM is similar to CD audio with its
advantages of low cost, durability, and high capacity. Most of
the cost of a CD-ROM product is in its design; pre-mastering,
mastering, and replication are less than five percent of the
total cost.
A master copy of a CD-ROM disc can last through 30,000 to
75,000 pressings. At Nimbus Information Systems, three
"stampers" are made from each master, and all replication is done
with the "stampers." Quality control is maintained by checking
the block error rate of each copy. Block error rate is the
number of blocks per disc that contain errors. If the block
error rate rises above 15, the "stamper" is changed. This
results in an effective error rate of one bit in 1015 because 12
percent of the storage capacity of each disc is used for a cyclic
redundancy check code. A copy can be made every six seconds, and
each copy can hold up to 700 megabytes of data. Accelerated
ageing tests indicate that CD-ROMs have a life expectancy of
about 100 years in the first world, or about 25 years in the
third world. The life of a CD-ROM can be prolonged by housing
and shipping them in caddies instead of jewel boxes.
Mr. Sands concluded his talk with a photographic tour of
Nimbus CD-ROM mastering facility in Charlottesville, VA. For
further information, call:
Mr. John Sands
Nimbus Information Systems
Telephone: (800) 782-0778
(804) 985-1100