SIGCAT Discourse
Volume 9, Issue 3
May/June 1995
The Professional Journal of the Special Interest Group on
CD-ROM
Applications & Technology
Copyright 1995
The SIGCAT Foundation
P.O. Box 3706
Reston, VA 22090
President: E. J. (Jerry) McFaul
U.S. Geological Survey
(703) 648-7126
(703) 648-6536 (fax)
Internet: JMCFAUL@GCCMAIL.CR.USGS.GOV
Executive Vice President: Duane Marquis
Department of Commerce
(202) 482-3055
(202) 482-4595 (fax)
Vice President: Mike Rubinfeld
National Institute of Standards and Technology
(301) 975-3064
(301) 590-0932 (fax)
Internet: MIKER@MML.NCSL.NIST.GOV
Vice President: Reenie Prettyman
National Institutes of Health
(301) 496-1936
(301) 480-6183 (fax)
Treasurer: Carol Cini
U.S. Government Printing Office
(202) 512-1283
(202) 512-1255 (fax)
Secretary: Joan McKean
National Oceanic and Atmospheric Administration
(301) 713-1170
(301) 713-1179 (fax)
Editor: Judith Lamont
(703) 790-0677(703) 435-5553 (fax)
Technical Editor:
Kathie Fraser
The Special Interest Group on CD-ROM Applications & Technology
(SIGCAT)
is a network of people who share a vision that CD-ROM will
profoundly
change the way in which they deal with information. Begun in 1986
and
sponsored by the U.S. Geological Survey, SIGCAT provides an open
forum
for people from all walks of life to investigate CD-ROM
technology and
its myriad applications. SIGCAT now comprises a wide spectrum of
individuals from Government, industry, academia, and the
interested
public. With its global network of over 8,000 members
representing more
than 300 organizations, SIGCAT has grown into the world's largest
CD-ROM user group.
SIGCAT members benefit from the ongoing exchange of ideas,
information,
and experiences through SIGCAT meetings and the bimonthly SIGCAT
DISCourse newsletter. Members receive substantial discounts on
CD-ROM
drives, media, software, and publications, plus free passes to
technical expositions. Members can also participate in any of 15
working groups that address specific special interests.
Membership in SIGCAT is free. Subscription rates for the SIGCAT
DISCourse newsletter are $40 per year.
To register as a new member, please call TheSIGCAT Foundation at (703) 435-5200, and leave your full name (please spell out), title, organization, address, and phone and fax numbers.
To receive the SIGCAT DISCourse newsletter, please register as a SIGCAT subscribing member by filling out the subscription form on page 15 of this newsletter and sending it with your $40 check payable to SIGCAT Foundation, P.O. Box 3706, Reston, VA 22090.
SIGCAT is pleased to introduce a column which will offer DISCourse subscribers a new perspective on CD-ROM developments. This column will provide insights on CD-ROM technology developments outside the United States.
Cannes on the French Riviera, already known as "The Image Mecca" for hosting the annual Cannes Film Festival, is expanding its role into electronic media. The Marche International de l'Edition et des Nouveaux Media (MILIA), held annually in January, is making its mark as a key multimedia exhibition. The preponderance of European companies at MILIA '95 indicates that the show is not yet well known in the United States, but, once word gets out, U.S. manufacturers may find this attractive site hard to resist.
Three types of companies were the major representatives at this year's show:
Regardless of the sector represented, CD-ROM technology was the star of the show. Many European companies, sensing the vast potential of this market, are preparing strategies to exploit the upcoming boom in CD-ROMs. Relative to the United States, Europe is in a "catch-up" mode, because only 17% of homes in France have personal computers versus a third of American homes. In England, more than 20% of the homes have PCs, but many are of very early vintage. With the cost of both computers and CD-ROM drives nearly in free fall, the European market is expected to blossom.
Another category of company represented at MILIA was newspaper and magazine publishers who, with decades of experience, have a certain "savoir-faire" in the art and manner of illustrating, in an attractive and instructive fashion, various themes of public interest. It is relatively easy for them to rework text, designs, and photos into an interactive multimedia product. However, this process requires the use of information technology with which publishers are not totally familiar. Despite the obstacles, many have taken the multimedia plunge and have produced titles in cooking, sports, sailing, and tour guides to the great cities of the world. The first French CD-ROM encyclopedia was produced during the 1994 Christmas season, and others are expected to follow soon. Two television stations_one German and one French_were offering "The Year in Review" on CD-ROM, which presented an assortment of reporting and video clips of important national events.
With the new technology comes a new vocabulary, and words describing several new job categories made their appearance at MILIA. The "infographist" is anyone who applies design skills in tandem with sophisticated design programs such as Photoshop, Painter, or Director. Typically, these individuals apply their skills to the development of children's games or educational software, for which the demand has been very strong lately. The multimedia "conceptor" is the talented individual who conceives the idea, the market, the subject content, its sequence, and its display_and then produces it on CD-ROM with the assistance of numerous technically inclined staff.
A source of considerable controversy at MILIA was the apparent disregard for copyright royalties exhibited by some distributors of images and video clips. To date, no major legal actions have been taken, but, as sales build, individuals at each level_authors, photographers, artists, and producers_will want to share in the revenue stream. Obviously, the onus is on CD-ROM producers to verify their copyrights before rushing off to market their products around the world. A warning: the registration of a trademark in the United States does not guarantee automatic worldwide coverage of that mark.
Those interested in participating in MILIA should contact the Reed Midem Organization in Paris at (33) 44 34 44 44. Other European events of interest are:
Tony Perla is Marketing Director for Media Archival and Retrieval Systems (MARS), the European distributor for re:Search International's authoring software products, re:Search and re:Text. In addition, his company provides electronic document management (EDM) consulting and multimedia composition services. He can be contacted at 145 Route de St. Mathieu, 01630 Grasse, France. Telephone: (33) 93 09 20 49. FAX (33) 93 40 42.
SIGCAT '95 surpassed even our most optimistic expectations in attendance, quality of presentations, and overall enthusiasm. We will provide a brief look at the conference now and more details in the next issue of the DISCourse. Over 1,000 registrants, vendors, and speakers enjoyed a dynamic and full program in our most attractive setting so far.
We are happy to have had such a good turnout for the conference, but numbers don't tell the whole story. Many comments were made about the quality of presentations and attendees. Whether they were experienced users or new to the field, attendees showed great interest and sense of purpose. The size of the conference was still small enough to allow interaction among attendees, presenters, vendors, and SIGCAT staff.
Some of the topics of greatest interest were multimedia, authoring software products, and the Internet_these topics and others had overflow audiences. Sessions on CD-Recordables and premastering software were also very well attended. Some tutorials were repeated a second time to accommodate last-minute attendees.
Our special exhibits were also very popular. The CD-ROM Classroom, Federal Showcase, and CD-ROM Information Resources Center highlighted the many advantages and uses of CD-ROM technology.
Awards were made to the Department of Transportation and the Environmental Protection Agency for outstanding discs and to Meridian Data Inc., IAC/Computer Select, 3M Corporation, and Kubik Enterprises for product innovations. We also made 11 awards to companies for educational products, including A.D.A.M. Software and Benjamin/Cummings Publishing Company, Dorling Kindersley, Softkey, and Broderbund.
Gold Disk provided copies of Astound, which was used as the official presentation software at the conference. It is an impressive product, and we appreciate the technical support provided to our speakers by Gold Disk.
Corel contributed a copy of CD-Creator, its premastering package, and four discs of photographic images. These items were awarded in a drawing held the last day of the conference.
Conference attendees received two discs, the Conference Disc and the Showcase Disc. The Conference Disc contains exhibitor and speaker lists, session listings, and public domain software tools. The Showcase Disc has demos of 15 retrieval engines, all accessing the CD-ROMCompendium, a database of Federal Government CD-ROM discs. If you were not able to attend the conference, you can obtain each of these discs for $30. Audio tapes were made of each conference session and can be obtained from National Cassette Services at (703) 635-4181.
Special mention should be made of the many volunteers who supported our effort by assisting speakers in the rooms, helping with set-ups, and resolving difficulties of all kinds. Thank you!
The DISCourse is the journal of the SIGCAT Foundation, published six times a year to keep members up to date on important developments in CD-ROM technology and products. In the past several issues, for example, you could have gained insights into the complex pricing of CD-ROM authoring software; learned how the Patent and Trademark Office provided public access to its data via CD-ROM; read a technical report on the lifetime of writable CDs and Photo CDs; and seen how the use of CD-ROM technology in networks is growing as advances are made in jukebox management software.
We also report on the activities of our working groups, the small group environments that allow SIGCAT members to focus their interest in specialized areas. In this issue, SIGLIT (Library Information Technology) and SIGSOFT (Authoring and Retrieval Software) report on meetings that they held. A complete listing of the working groups and contact numbers is provided on page 15.
In this issue, we are launching a new column, "European Viewpoint," which describes developments in CD-ROM technology in Europe. Tony Perla, Marketing Director for Media Archive and Retrieval Systems in Paris, will be writing this column. We welcome contributions from SIGCAT's other international members as well.
We would like to thank former editor Yvonne Kidd for her role over the past 2 years in establishing and maintaining a high quality of writing in our publication. Her contributions are much appreciated.
CD-ROM discs are becoming increasingly popular for use as mass storage devices. The combination of low media cost, improved recording technology, and a uniform file standard has resulted in a wide acceptance of CD-ROMs for WORM technology in mass storage applications.
CD-ROM jukebox systems from numerous manufacturers contain a wide range of disc capacities as well as number and type of drives. The high storage capacity makes these jukeboxes extremely attractive in terms of size, cost, and performance.
This increasing popularity creates a need to standardize the method of evaluating jukebox performance as mass storage devices in order to allow comparison between competitive products. At present, there are a number of "specifications" for CD-ROM jukeboxes. Many of these are actually "drive" specifications and not a true measure of jukebox performance. Table 1 lists the most common specifications quoted for CD-ROM jukeboxes.
These specifications alone do not provide adequate information to allow prospective purchasers to determine which jukebox model or configuration will deliver the best performance in their working environments. In order to evaluate the potential performance of a jukebox, remember that, although it is possible to configure a jukebox with multiple drives, these multiple drives are serviced (loaded) by a single robotics element.
The practical concerns of a prospective purchaser of a CD-ROM jukebox include:
When attempting to calculate the time to deliver data (TDD) of a jukebox, consider that only three different functions exist inside a jukebox. These are:
Media exchange time as defined specifically does not include any disc-loading actions or other processes performed while the robotics is free to begin loading another disc.
Next, two different multiple-drive jukeboxes are compared to determine which will better service the data storage needs in an active multi-user environment. It is assumed that there is a constant demand among users for image files of a consistent size and that these files are large (6 MB) high-resolution color photo images. The characteristics of Jukebox A and Jukebox B are given in table 2 below.
Starting at the zero point on the time line, both jukeboxes begin to load discs and process data transfer. Because a constant data demand is present, both jukeboxes proceed to the next "load" function immediately after completing the previous "load." This process is then continuous.
Figure 1 illustrates the actions along a time line as Jukebox A loads its first disc in 12 seconds and cue time takes 2 seconds more. With a data transfer rate of 600 kb/sec (.6 MB/sec), the data transfer time is 10 seconds. The resulting TDD for the first file to the first user is 24 seconds. Twelve seconds after loading the first disc, or 24 seconds after point zero, the second disc is loaded onto the second drive, and 2 seconds later, transfer of the second file begins. Again, data transfer time is 10 seconds, so the total TDD for the second file is 36 seconds. Repeating the process two more times to fill the remaining drives results in TDD times of 48 and 60 seconds for the third and fourth files, respectively. At this time, the first drive has already completed its file transfer and is ready to accept another disc. The entire process, and therefore the graphs, becomes repetitive.
For Jukebox B, media exchange time is only 4 seconds (faster robotics) and data transfer time increases to 20 seconds (slower drives). It takes Jukebox B 4 seconds to load the first disc, cue time is 2 seconds, and data transfer time is 20 seconds. The TDD for the first file transfer is 26 seconds. The second requested disc is loaded 4 seconds after the first, or 8 seconds from the point zero. This second drive takes 2 seconds to cue and 20 seconds for data transfer, so that the total time to deliver the second file is 30 seconds. As in the first example, this process becomes repetitive, and the TDDs are 34 and 38 seconds, respectively for the third and fourth files. However, at the point in time where the fourth disc has been loaded into the reader, the first drive has not yet completed its file transfer. The robotics must then wait the additional time needed by the first drive to complete its file transfer before it can service a fifth user. Therefore, the TDD of the fifth file becomes 52 seconds, and then the pattern again becomes repetitive for the next successive three users.
Table 3 summarizes these results. Jukebox B can deliver considerably better TDD performance in servicing an ongoing multi-user demand for data, despite the fact that it contains slower drives and the file size is relatively large.
It is important to note that, as the size of the files being accessed decreases, the TDD performance of Jukebox B becomes increasingly better than Jukebox A. Additionally, the difference in TDD between Jukebox A and Jukebox B increases with the number of files requested from the jukebox. This fact is clearly illustrated in figure 2.
A closer examination of figure 1 reveals that the TDD performance of Jukebox A is not improved by having more than two drives inside the jukebox. Because the first drive has already completed its file transfer by the time the second disc is loaded into the second drive, this first drive is available to accept the disc for the third user. As a result, there is no advantage in installing additional drives in Jukebox A for this application.
Because there is a dynamic relationship between robotics time, data transfer time, cue time, and number of drives, there is a file size value for any given jukebox configuration at which the system performs most efficiently in terms of TDD. The optimum file size (OFS) of a jukebox is defined as the size of the file that allows the jukebox to achieve maximum operating efficiency. Because jukeboxes rely upon a single robotics assembly to deliver the discs to the drives, it becomes most important to maximize the efficiency of this action. This efficiency is highest when the first drive in the system has completed its data transfer at the precise moment that the robotics has completed the mounting of a disc into the last drive in the system. The OFS of a jukebox is calculated by using the following formula:
OFS = [ ( (N _ 1) x M ) _ C ] x D where:
N = number of drives
M = media exchange time
C = cue time
D = data transfer rate (MB/sec)
Multiplying (N _ 1) and M results in the time (in seconds) it takes the robotics to load the remaining drives. We then subtract from this figure the cue time (C) where no data are being transferred. The result is the time remaining to transfer data, multiplied by the data transfer rate (D) to determine the OFS.
In a given multiple-drive jukebox, when file size is smaller than the OFS, the drives must wait for the robotics to become free to service them. Conversely, when the file size is larger than the OFS, the robotics must wait for the drives to complete their data transfer.
Applying the above formula to the two jukeboxes used in the sample, we come up with the following OFS figures:
Jukebox A = [ ( (4 _ 1) x 12 sec) _ 2 sec] x .6 MB/sec = 20.4
MB
Jukebox B = [ ( (4 _ 1) x 4 sec) _ 2 sec ] x .3 MB/sec = 3.0
MB
Jukebox A has an OFS that is larger than the 6-MB files of the sample application, and Jukebox B has an OFS that is smaller than these 6-MB files.
Figure 2 illustrates that Jukebox A has a linear TDD curve the slope of which is determined by the media exchange time. In a jukebox system operating with files smaller than its OFS, the TDD curve is no longer linear owing to the fact that the robotics must wait for the first drive to complete data transfer. Therefore, TDD calculations for this application must include the waiting time, which will occur every time the machine completes loading all drives.
It was documented earlier that Jukebox A showed no improvement in performance as a result of having four drives. This fact is verified by using the OFS calculation for Jukebox A reconfigured with only two drives and identified as Jukebox C:
Jukebox C = [ ( (2 _ 1) x 12 sec ) _ 2 sec ] x .6 MB/sec = 6.0 MB
As can be seen, Jukebox A would have been equally suited to handling the 6.0-MB image files of the sample application if it were configured with only two drives, because the applications file size equals the OFS of Jukebox C.
Current jukebox specifications fail to adequately define the TDD performance of a jukebox. It is necessary that manufacturers consistently measure the three functions that directly affect the performance of a jukebox_media exchange time, cue time, and data transfer time_to evaluate and compare performance in a meaningful way.
It has also been demonstrated that in jukeboxes in general, and in multiple drive jukeboxes specifically, the use of a single robotics assembly makes the media exchange time the most significant performance measurement in determining the TDD in an active multi-user environment. The TDD for each successive user of a jukebox operating with files equal to or smaller than the OFS of the jukebox will be directly affected by the media exchange time.
In a mass storage environment with files being stored in and accessed from a jukebox, the OFS must be used instead of the ratio of discs to drives to determine the most efficient configuration of drives in a jukebox. The assumption that an ideal ratio exists for all applications is incorrect. The storage and access requirements of the users should determine the capacity of the jukebox. Kevin Byrne is President of Euroson America, a marketing and sales agent for NSM. He can be reached at (800) 423-1122. Table 1. Definitions for the more common specifications Media exchange time (secs.)The time required to remove a disc from, and load a disc onto a drive. Spin up (seconds)The time required after load for the drive to spin up and identify the disc. Access time (seconds)The time required to position the laser to the beginning of any file. Cue timeThe time required for disc spin up and access after freeing the robotics. Data transfer rate (kb/sec)The speed at which the drive reads data off the discs. Data transfer timeThe time required for transfer of a given file. Number of drivesThe number of installed drives Number of discs The disc capacity of the jukebox Discs/driveExpressed as a ratio,i.e., 500 discs/4 drives = 125
Table 2. Comparison of multiple drive jukeboxesJukebox A Jukebox B Media exchange time (seconds) 12 (average) 4 (average) Cue time (seconds) 2 2 Data transfer rate (kb/sec) 600 300 (4x drive) (2x drive) Number of drives 4 4 Table 3. Comparison of TDDs Time to deliver data (TDD) (seconds) File 1 File 2 File 3 File 4 File 5 File 6 File 7 File 8 Jukebox A 24 36 48 60 72 84 96 108 Jukebox B 26 30 34 38 52 56 60 64
At the March 29 SIGCAT regional meeting, attendees learned how jukeboxes, storage management software, and CD-Recordable technology are expanding the use of CD-ROM technology. The value of CD-ROM technology in handling vast amounts of data_well beyond the 650-MB capacity of a single disc_is being more fully exploited by use of a variety of devices that handle multiple CD-ROMs, including jukeboxes and CD-ROM towers. Handling this complexity requires specially designed software that can effectively access the data. Finally, the increasing ease of use and lower prices of CD-Recordables is placing these devices within the reach of a larger group of users. Contact information for each of the presenters is given on page 13. Jukeboxes and Jukebox Management Software
Kathleen Dell'Orto, U.S. Patent and Trademark Office (PTO), discussed some of the challenges of developing a new system to provide electronic access to foreign patent and trademark information. The option of using CD-ROMs was clearly the best choice_especially considering the disadvantages of paper and microfilm as storage and distribution media_but the system needed skillful design in order to handle more than a terabyte of data and be easy enough to use so that the general public could access it readily.
Dell'Orto emphasized the importance of careful planning in the successful implementation of a system. She advised that developers consider the number and capabilities of the users, speed and access requirements, and the projected system size. In order to be as user friendly as possible and accommodate future growth, the Foreign Patent Access System (FPAS PLUS) was designed with these features:
The hardware requirements were stringent, including an easy-to-open case (because new discs are added frequently) and easily accessible components; modular replacement capability, particularly of CD-ROM drives; large capacity; and efficient use of space. These considerations were important in view of the large volume of data involved_the index alone is an impressive 4 gigabytes, and the system incorporates 1.2 terabytes of storage.
The resulting system, which has been in place since March 1994, has eight workstations running on 486/33 machines with a Pentium server and both jukeboxes and CD-ROM towers to hold the discs. Users can perform searches, view documents, and print copies of patents. The system stores images of the patents, which allows users to view or print any drawings in the patent. FPAS has been used by over 1,000 people since it opened.
Kirti Vashee, Smart Storage, Inc., discussed features to look for in storage management software, a key element of effective jukebox use. The software should allow for easily increasing the volume of data handled without requiring that the system be redone. It should also be able to integrate a number of jukeboxes within a single file structure to allow for configuration flexibility. It is desirable to have a single software interface to all devices.
Vashee said that other important features include:
The network interface should support multiple retrievals, provide for cross-platform access, support across-the-network operating system, network administration tools, and minimize the load to network traffic. In some cases, multiple networks may access the same jukebox, and the jukebox management software needs to support all of them.
Ideally, storage management software should maintain transparent access to information in systems containing many discs. (The largest system installed by Smart Storage has 2200 discs.) Storage management software is especially useful in legal environments, where quick access to data is essential, and in image archives, which present special challenges because of their large size. When images are indexed, the index is contained on the file server and the images themselves are on CD-ROMs that may be contained in a variety of jukeboxes.
Mike Clark, Eastman Kodak Company, described Kodak's Professional Image Library 30 (PIL 30), which operates on PCs under Windows or on the Macintosh. It consists of a jukebox, Kodak's Shoebox image catalog software, and jukebox manager software. The image management software can search for images by attributes and captions and can search for any type of file, including video and documents. It can search 10,000 indexed thumbnail images in less than 10 seconds.
Kodak is planning several new releases, including the next version of the Jukebox Manager software, which will allow automatic daisy chaining (connecting a series of devices) and multi-user access_currently the system is designed for a single user. Also, on-line help for the server and client software will be available. The software will support Windows for Workgroup, Windows NT, OS/2, and Novell.
Kodak partners with NSM, which provides the jukeboxes, and adds its traditional support, training, and warranties. Other partnerships with Kodak will be announced shortly.
Ron Cannizzaro, Pioneer New Media Technologies, described Pioneer's 500-disc jukebox, which is among the largest in the industry. It holds 330 GB of data and can contain up to four CD-ROM readers or CD-Recordables. The removable magazines, that hold 100 discs each, facilitate changing discs and allow continuous operation of the other magazines in the jukebox during the changing process. Discs are in two columns rather than one, an advantageous design feature which minimizes the distance between the disc readers and the most often used discs. There are many software management options. Interfaces include Novell, PC Networks, and UNIX. The suggested retail price of the jukebox with four readers is $22,000. The installed base includes government facilities and libraries that require large storage capacity.
Carl Beckman, Alan Safarty & Associates, discussed factors to consider in the selection of jukeboxes: data and application, storage requirements, standalone vs. network, and access time required. An enhancement in jukebox design that will be available in the summer of 1995 is a gear-driven picker as opposed to a belt drive, which should be more reliable than the older belt drive. A new path for the robotic arm reduces access time by having the arm move up and over simultaneously, rather than moving through two dimensions sequentially. A design feature that facilitates maintenance and repair is the ability to remove drives without any tools. The jukebox uses Toshiba's 4X drives. The company runs end-user focus groups for beta products, and Safarty invited attendees of the SIGCAT meeting to participate in this process.
Kevin Byrne, Euroson America, represents NSM, the largest manufacturer of commercial jukeboxes. Originally involved in musical CDs through the karaoke business, NSM adapted its first CD-ROM jukeboxes from this musical application. The new Mercury series was designed from the beginning for data storage rather than being an adaptation of an audio jukebox. Some key questions to consider when selecting a jukebox are:
A detailed analysis of jukebox performance can be found in the article that begins on page 1. This article demonstrates that, surprisingly, more read drives do not necessarily improve jukebox performance.
Robert van Eijk, Laser Magnetic Storage/Philips, predicted that 12 million CD-Rs will be installed in the North American market alone by 1999, less than 5 years from now. Primary applications for CD-Recordables are title creation, data exchange, and archiving. Van Eijk identified several key forces in this market: u Decreasing prices. When the price of CD-Recordables approaches $500, they will begin to replace CD-ROMs in the small office/home office (SOHO) market. At this price, the CD-R would be about the price of a CD-ROM and tape backup combined and would begin to serve both purposes.
Philips' market analysis for 1994 indicates that it has captured a 50% share, with Yamaha at 15%, Sony at 10%, JVC and Ricoh with 5% each, and others accounting for 15%.
Philips now has on the market the CDD 522, an external double-speed read/write CD-ROM drive with a 2-MB buffer expandable to 32 MB. It has a SCSI-2 interface for data and audio transfer. The company is choosing to stick with the double-speed recorder now because of concerns about standards. By June, Philips plans to introduce the CDD 2000, a half-height CD-Recordable, which has a double-speed write and quad-speed read drive, an incremental write capability up to 100 sessions, and a 1-MB buffer that will handle all current CD formats. An external version is scheduled for introduction in the third quarter of 1995. Several software options are available with this package, including Win-On-CD to Go, which comes as a basic package, or one that records CD-I and Photo-CD. Alchemy Basic is offered as an archiver, along with VideoPak.
For the future, van Eijk forecasted the availability of a high-density CD-ROM by 1996, high-density erasable by 1998, and a two-layer CD-ROM by the year 2000.
Chris Manzione, Sony Corporation, described CD-Constructor_, an easy-to-use, low-cost CD-Recordable archiving formatter that creates a virtual ISO file structure and works in a drag-and-drop mode from the Windows File Manager or from within an application. It can operate at single, double or quad speed, is multisession and handles all types of files, including WAV and AIFF audio files. The Macintosh version ($199) is available now, and the Windows version ($149) is expected in June. CD-Constructor_ supports Sony, Philips, Yamaha, Ricoh, and Kodak CD-Recordable products. Hybrid Formatter ($695), a separate Sony product for the Macintosh platform, allows production of a disc playable on both the PC and the Macintosh.
Sony's new double-speed CD-Recordable, the CDU 920S, supports all existing CD formats and can record either single session ("disc-at-once") or multisession ("track-at-once"). It has a caddy load drive and can write a CD-ROM in Mode 1 and 2, CD-ROM XA audio, CD-I, CD-Plus and audio combined. The CDU 920S is priced at $1625 bundled with CD-Constructor_ for Windows. Laptops and Labels
Jan O'Hara, Toshiba America Information Systems Inc., demonstrated the new T2150CDT notebook computer with a double-speed CD-ROM drive. Standard equipment for the $3,400 system includes a 500-MB hard drive, sound card with MIDI and .WAV sound support, and a passive color monitor (active matrix color is also available). Competitive features are a battery life of over 2 hours and a weight of under 7 pounds. The laptop was on display at Toshiba's table at the SIGCAT meeting.
Gretchen Tobel, Avery Dennison, offered meeting attendees beta test packs for their new CD-ROM labeling system. Avery, well known for labels of all types, has developed a simple and cost-effective means of labeling one-offs. The labels are printed two to a sheet on a laser or ink jet printer. An alignment device called a "tophat," which is a part of the label, is used to place the label on the disc. After the label is secured, the tophat tears off. The labels will be offered with both permanent and temporary adhesive.
The ABCs of evaluating CD-ROM products was the subject of a presentation for SIGLIT members in January at the U.S. Court of Appeals. Georgia Higley, Serial Reference Specialist in the Serial and Government Publication Division of the Library of Congress, spoke before a large audience of special librarians. Ms. Higley discussed and distributed a list of standards that the Library of Congress applies in deciding which CD-ROM products to acquire for its collection. If you would like a copy of these standards, please call Felice Sacks, chairperson of SIGLIT, at (202) 663-6763. In addition, Felice is asking those members of SIGLIT who have Internet mailing addresses to contact her at fsacks@capcon.net or felice@access.digex.net so that she can develop a SIGLIT network in cyberspace. SIGSOFT
The long-awaited listing of CD-ROM authoring software is now available and will be distributed on the SIGCAT '95 Showcase and Resource Disc. This matrix lists 21 of the leading authoring products and has a total of 75 fields describing the products' features. Participants in this group include chairperson Tom Tiedeman (Quick Source), Tom Haskell (DynCorp), Mark Lunsford (dataDisc), Bill Reilly (GPO), and Greg Velott (Navy Tactical Support Activity). We developed a questionnaire, sent it out, and compiled the answers given by the producers of the software. The product can be viewed and browsed from the main menu of the Showcase Disc and can also be searched using GSSEARCH. The entire database can be exported from GSSEARCH as a dbf file or as ascii text; individual records may also be printed or saved to a file.
The following products are included in the authoring software matrix:
If you did not attend the conference and would like a copy of the Showcase Disc, please contact the SIGCAT office at (703) 435-5200. SIGMULTEC
SIGMULTEC, SIGCAT's Multimedia Working Group, met in a "standing room only" format at SIGCAT '95. One of SIGCAT's newer working groups, SIGMULTEC is expected to have lively participation over the next year. At the meeting, Mr. Frank Garber of United States Video Corp. (Vienna, Virginia) discussed the process of repurposing media elements into interactive multimedia for CD-ROM. He showed how materials such as audio cassettes and slides used in training programs developed for the National Technical Information Service (NTIS) were incorporated into a new interactive product.
In addition, he demonstrated two techniques for presenting interactive video: 1) the hardware-based MPEG-1; and, 2) software-based Indeo. In another product, U.S. Video Corp is synchronizing text of the Congressional Record with the C-SPAN delivery of these statements. Using authoring software developed in-house, the product presents video and text side-by-side on the screen.
Mr. John Graves, a multimedia developer based in San Diego, discussed the SIGCAT Home Page he recently developed for the World Wide Web. He is also developing information on SIGMULTEC which will be connected to this page.
One address or Uniform Resource Locator (URL) for the SIGCAT Web Page is:
http://www.learned.com/browse/jgraves/orgs/SIGCAT/sigcat.htm
. We hope
to have a simplified address soon.
SIGMULTEC will meet again at the next bimonthly SIGCAT meeting, date to be announced. Please join us!
"CD-ROM _ A transitional medium, soon to be eclipsed," or so say some pundits in our industry. Are they right? Is CD-ROM already a mature technology destined to be replaced by better ways of storing and distributing information?
To publishers, charged with the responsibility of devising information delivery strategies, these aren't just theoretical questions. They involve issues that are central to deciding how and when to invest in new electronic publishing technologies. The stakes_in terms of time, money and human resources_are immense.
Fortunately, the marketplace_not pundits_will determine the extent of CD-ROM's acceptance and its ultimate fate. Thus far, the marketplace has resoundingly endorsed CD-ROM as the most cost-effective way to deliver information that isn't highly time-sensitive. Other distribution media, including online services and the Internet, offer advantages, but cost isn't one of them. Moreover, CD-ROM's current cost advantage over online distribution won't disappear any time soon. On the contrary, it will almost certainly increase in the next few years, due to advances in CD technology.
Before selecting an electronic publishing medium, publishers need to consider the nature of their information and their markets. Is the information time-sensitive? Will its value to customers diminish if it's more than a day or a week old? Will customers pay extra for frequent updating and for instant access to information? If the answer to these questions is no, then CD-ROM is the better choice because it costs less and is simpler to use.
Those looking for a horse race between the two distribution media will be disappointed. The most likely scenario for the future is for CD-ROM and online services to coexist peacefully as they evolve, and even work together in some cases. Comparisons of the two are instructive and inevitable. There is much to be learned in observing how they stack up at present and exploring how technological advances will affect the costs of both CD-ROM and online distribution. CD-ROM Technology Compared to Online Costs: Advantages and Disadvantages
One of the main selling points of online services is instant access to vast amounts of data that can be updated daily, hourly or even more frequently. The high cost associated with frequent updating among other factors, results in relatively high price for data. Less time-sensitive material can usually be provided more cheaply by CD-ROM, with its relatively high-capacity and cost-efficient data storage capabilities.
Delivering and accessing information by CD-ROM involves a variety of development, manufacturing and hardware costs. The cost of data preparation, a major aspect of development, ranges from a few thousand dollars for a small project to several hundred thousand dollars for a large one. Manufacturing a standard CD-ROM currently costs about $1 per unit in pressing quantities of 1000, with mastering fees adding approximately $700 to the total. Manufacturing costs are unaffected by the amount of data contained on the disc, which currently tops out around 700 megabytes without compression. Thus, excluding fulfillment and postage charges but including mastering fees, the current unit cost of delivering 700 megabytes of data by CD-ROM is well under $2, and drops quickly as pressing quantities increase. The hardware component is the CD-ROM drive, which now adds less than $100 to the cost of a new computer. Other varied but generally minor costs may include rights payments, licensing fees for authoring and search software, packaging and documentation.
The chart on the bottom of page 6 shows the different costs of providing the same amount of information by online service, print, diskette, and CD-ROM. The online cost shown is the charge for downloading a one-megabyte file from The Federal Bulletin Board at 9600 bps plus an estimated telecommunications charge. The print equivalent of one megabyte is represented by the unit paper, printing and binding cost of a 500-page book in a 1,000 copy printing. The cost of a diskette includes duplication charges for the data, along with the physical medium itself. For CD-ROM, the per-megabyte cost equals 1/700 the manufacturing cost of the disc and a pro-rata share of mastering fees. While each medium is cost-effective in its proper context, the basic point that CD-ROM offers the lowest cost storage and delivery of data is irrefutable.
Comparing CD-ROM and online further, both require development and data preparation. Both require customers to have specialized hardware_drives for CD-ROM and modems for online. Even though drives are somewhat more expensive than modems, the difference is not great enough to affect the cost comparison.
CD-ROM's huge storage capacity allows text, images (both still and video) and audio files to be combined, thereby providing the basis for interactive multimedia products. While it is technically possible to transmit video images and multimedia presentations by means of standard phone lines, the time needed to do so makes it impractical.
CD-ROM gives publishers total control over how their information is presented and sold to customers. Unless a publisher is willing to develop its own online service_a costly proposition_offerings must be made by licensing information to commercial online services, which use their interfaces as the presentation vehicle. With CD-ROM, publishers maintain a direct relationship with their customers, thus controlling sales and P&L results. If the choice of media is available, distributing data by CD-ROM offers publishers far greater profit potential than online.
The basic costs incurred in transmitting information online are telecommunications charges, which cover data transfers over local lines and packet networks, and connect-time charges, which cover the costs of network servers, maintaining databases and royalty payments to providers. A modem is needed to link the customer's workstation with the central computer. As with CD readers, modem prices have dropped to the point where they are no longer a significant cost factor. Customers must still pay their local phone company monthly service charges for an extra line, however, or use an existing phone line.
Business and professional online services, which still account for over 90% of the $11 billion U.S. online industry, typically separate telecommunications charges from connect-time charges. As with long-distance phone service, telecommunications rates vary by time of day. Typical charges are $8 to $12 per hour for daytime use. The variance in connect-time charges is much greater, ranging from a few dollars per hour for a Government-produced database to $300 per hour for a highly specialized scientific database. The average for searching commercial scientific and business databases is around $100 per hour, a figure which has remained stable for several years. Charges for access time are usually calculated in minutes at the applicable hourly rates.
In the small, fast-growing and highly competitive consumer segment, pricing of online services is less predictable. While the major services do offer specialized data at premium rates, pricing policies are designed to encourage high-volume usage of basic services. America Online's customers, for instance, use the service's communications features_e-mail and "chat"_more than its information retrieval options. Monthly charges of $10 or less cover a limited_or unlimited, in the case of at least one service_amount of basic services usage. For additional use or for special services, customers pay hourly rates of a few dollars per hour, which cover both telecommunications and connect-time rates.
Higher transmission speeds affect transaction costs positively, since the faster data is transferred, the shorter the connect-time needed (although the time a customer spends online in entering and refining a search strategy is unaffected by transmission speed). Although this practice seems to be fading, many commercial online services traditionally lessened this apparent advantage by charging users more for sending data at higher speeds, typically doubling connect-time charges for data sent at 9600 bps versus 2400 bps. But as long as customers accord comparable value to the offering, the prices are regarded as fair. For example, brokers are quite willing to pay for up-to-the-minute stock quotes, which can be provided efficiently and economically online.
Due to the interactive nature of online searches and the need for low error rates, speeds below 9600 bps are still widely used. Figure 1 demonstrates that online transmission is cost-effective for down loading only small amounts by calculating the time needed to transfer the contents of a standard CD via modem, which calculates the time needed to transfer the contents of a standard CD via modem.
In order to serve multiple users accessing the same database simultaneously, online services require large, costly computer systems. The systems must be operated and maintained 24 hours a day, 365 days a year, with enough capacity to handle service peaks comfortably. As with PCs, large system costs have been declining. Still, the cost of operating and maintaining a data center must be passed along to customers.
One of the driving forces behind the acceptance and success of commercial online services has been their ability to price data for occasional use. Because the services control and monitor each customer's access to databases, they are able to bill customers for connect-time and other uses such as printing. Though charges are still predominantly based on access time measured in minutes or hours, some services now base charges on search results, which benefits less efficient searchers by reducing some of the time pressure. One service, for instance, has recently replaced connect-time charges with a pricing scheme that charges $1.50 per 1,000 characters (including spaces and returns) viewed. Another now separates royalty payments from connect-time charges, which flattens the rates.
Usage, or pay-as-you-go, pricing offers advantages to both service provider and customer. Online services can offset their high costs by employing the pricing model that offers the greatest revenue-generating potential. Customers benefit by being able to purchase small amounts of high-value data as needed. For these reasons, online providers serving business and professional markets will continue to rely on time-based charges as the primary way of selling their information. Turning CD-ROM Disadvantages into Advantages
We mentioned earlier that the main advantage of online compared to CD-ROMs is its ability to provide current information or highly specialized information economically. Two enhancements allow CD-ROM users to reap the same benefits as their online counterparts: hybrid systems, which provide timely data transmission by allowing the two to work in tandem; and usage-pricing structures implemented through encryption technology to provide and exceed in cost-efficiency the pay-as-you-go convenience of some online systems.
Hybrid systems provide the capability for doing basic research on CD-ROM and updating information by means of online searches. In such a hybrid system the cost of using an online service would be a fraction of a usual search, since telecommunications and connect-time charges are kept to a minimum. This approach combines the advantages of both online and CD systems.
For example, Knight-Ridder Information, Inc., a leader in business and professional markets, offers many databases in both CD-ROM and online formats. Its DialogLink communications software allows customers to transfer searches from CD to online databases in order to obtain the most current data. In another hybrid application developed by the consumer online services, CD-ROMs are used to deliver graphics and related multimedia enhancements to online offerings.
Further advances in technology will allow publishers to provide customers with updates of CD-ROM databases without involving a conventional online service. The system developed by Teleshuttle Corporation, for example, uses a communication module on the CD to transfer up-to-the minute information from a central server to the user's hard disk by means of a modem link. This approach permits publishers to integrate frequent updates with CD databases while maintaining control over the user interface and terms of sale. Encryption and Metering Technology Makes Usage Pricing of CD Information Feasible
Combining encryption technology with usage-tracking software allows CD information to be sold on a usage basis. [See "CD-ROM Pricing: The Pay-As-You-Go Option," CD-ROM Professional, September/October, 1994] This advance radically enhances CD-ROM's ability to meet the demands of the marketplace and compete even more effectively with online distribution. Since telecommunications and connect-time charges are not cost factors in delivering information on CD, usage pricing of CD information can be based on measures other than time_a point that will make customers happy. A search that might cost $25 from a commercial online service could be priced at half that amount or less on CD, and still allow a substantial profit for the information provider.
In addition to alleviating time pressure, usage pricing of CD information promises much lower base fees, which will allow customers to acquire many more titles. Because of the inherently low costs of delivering and accessing data on CD, this new method of charging for information gives providers great latitude in deciding how to make their pricing flexible, predictable and market-sensitive
One CD-ROM-oriented usage pricing system available from CD-MAX, Inc., combines proprietary encryption and usage-tracking software with communications and database programs. In this approach, CD information may be decrypted, or unlocked, only by registered customers, whose searches are then recorded in a special "watch" file. To gain continuing access to the CD, customers must allow security keys installed on the hard disk to be updated periodically. The transmission of new keys, which may be accomplished by either modem or diskette, also allows usage information to be retrieved by the company. Customers are then billed for usage as directed by the publisher.
A practical usage pricing mechanism for CD information allows publishers to charge LAN users on a pay-as-you-go basis, thereby eliminating or lessening dependence on inflexible (and ineffective, from a security standpoint) multiple-user licenses. Encryption and metering technology also enable the distributions of CD-based information on wide-area networks, by securing copyrights and permitting publishers to charge for actual use. What's in Store for CD-ROM and Online:The Possibilities and the Costs
CD-ROM is still in its first generation; the technology is by no means mature. Compared to the evolution of computers, current CD technology can be likened to the technology of Apple II days, when 32-bit processors were only theoretical possibilities. Those who would dismiss CD-ROM as a "transitional" medium can only do so by ignoring the ongoing advances work in optical publishing technology.
In the years ahead, improvements in reader performance will offer greater convenience and better value for customers. Double-speed readers, already the industry norm, are being replaced by quadruple-speed readers as their prices decrease. Prices of multi-disc changers, which further improve access speed and convenience, are bound to drop as a result of the large manufacturing base established for CD audio equipment. Lastly, prices of CD-R drives will continue to fall, which in turn will lower development costs and broaden the market for desk-top CD publishing applications.
Other exciting improvements lie ahead in disc capacity and performance. Philips is likely to introduce high-density video CD-ROM in late 1995 or early 1996 to satisfy the market's demand for a way to put full-length movies on a less costly medium than videotape. High-density CD-ROM will quickly follow. In both cases, the new format increases disc capacity from the current ISO standard of 658 MB to 3.3 gigabytes_a fivefold increase. Once the new format has an established hardware and manufacturing base, per-megabyte information delivery costs will decrease dramatically. Data that costs less than one-quarter of a cent to store on CD-ROM today will cost just a few hundredths of a cent in the near future.
High-density CD is made possible by increasing the frequency of the laser used to read data, in this case from infrared to red, which allows more efficient use of disc space. Further increases in disc density are possible by raising the laser frequency still higher. Frequency-doubled lasers, which shift the wavelength from red to blue-green, are reportedly in prototype stages. Another approach nearing commercialization uses multi-layered discs to increase data density. In fact, the forthcoming new standard for high-density CDs includes optional specifications for double-layer CDs_a powerful one-two punch indeed. Improvements in local and wide-area networking technology will make CD databases even more economical and accessible. Dramatic Cost Reductions Unlikely In Online Delivery
Online cost factors have remained relatively stable for years. In contrast to CD-ROM, no dramatic cost reductions are likely to come about from technological advances in the near future. Rates for telecommunications charges, for example, have already been reduced as a result of deregulation and the widespread use of fiber optic technology for long-distance calls. Overhead and royalty costs, the basic factor represented in connect-time charges or output-based equivalents, are even less responsive to technological advances.
Video-on-demand, which for many is the apotheosis of the information superhighway, requires huge improvements in bandwidth pipelines. Fiber optic technology, which allows the necessary increase in bandwidth, is cost-effective for long-distance trunk lines but not for individual home connections, even if implemented on the massive scale required to build a large consumer base. The establishment of digital networks using fiber optic technology may be a decade or more away.
The Internet seems to offer publishers an online alternative, though early uses deliver free marketing_not commercial_information. Even after the technical, marketing, security, and pricing issues involved in delivering for-sale information on Internet are worked out, the same question of cost remains: Will customers pay extra for online access if the same information is available in a lower-cost format?
For the foreseeable future, online services of all kinds will depend on large computer installations, whose costs must be passed along to customers.
Will the National Information Infrastructure (NII) as envisioned by the Clinton administration offer a less costly and more powerful alternative to the commercial technologies in use today? Someday, perhaps, but many obstacles must be overcome before the NII can be realized_especially in its more visionary aspects. A welter of regulatory, political and technical issues must be untangled. And cost issues are equally daunting.
Building the NII will require the replacement of an existing information delivery system at a huge cost. In economic terms, new technologies make sense when the cost of maintaining the performance of an existing system is greater than the total cost of installing and maintaining a new system, plus the value of the extra services provided by the new system. Thanks in large measure to the dynamism of CD-ROM technology, the cost/performance ratio of the existing system is actually improving, which pushes farther out into the future the day when the NII will be feasible. The information delivery system we have today is practical, cheap and rapidly improving_so, long live the "transition!"