College and Research Libraries Research Notes Library of Congress Staff Test Optical Disk System Victoria Ann Reich and Melissa Ann Betcher The Library of Congress has four optical disk reader stations that allow patrons to view and print document images at a resolution of 150 by 300 lines per inch. This phase of the evaluation project involved the staff during the implemen- tation of this technology. It measured satisfac- tion with training, documentation, equip- ment, and printing; system reliability; and materials to be scanned. Staff found no major flaws with the system and saw its usefulness primarily as an access tool. Some problems were experienced viewing nontextual materi- als, and care in choosing materials for inclusion onto the system is recommended. BACKGROUND The future provides many challenges to the library and business worlds. In a soci- ety that generates masses of paper and de- mands quick and easy access to informa- tion, new options must be realized. Optical disk technology is one method be- ing explored by a variety of institutions. The promises held out by optical disk technology are just now being realized: the ability to store large quantities of infor- mation in a compact space, the nearly in- stantaneous retrieval and display of infor- mation, the prospects for telecommunica- tion, and the preservation of the original item when desired by diminishing the need for physical handling. For the library world, the Library of Congress' Optical Disk Print Pilot Program is on the leading edge of experimentation with this new technology. The pilot program, begun in late 1982, is a four-year program designed to assess the applicability of digital optical disk technology to library services and preservation. The Library of Congress em- barked on this project as a means of pre- serving and providing rapid access to a va- riety of collections of high-use and rare and previously unavailable materials. Much of the library's equipment and software has been specially prepared un- der contract to Sony Corporation of Amer- ica and to a California firm called Inte- grated Automation. Two systems are in use: a videodisc system using off-the-shelf Sony videodisc equipment and an optical digital disk system using twelve-inch disks, with a target capacity of ten thou- sand to fifteen thousand pages per side of a disk. In the digital system, approxi- mately eight million bits are scanned and captured to represent one page. Selected journal article cited in BIBL, one of the li- brary's online bibliographic files, are available on the optical disk system. Pa- Victoria Ann Reich is head of acquisitions at the National Agricultural Library, Beltsville, Maryland 20705, and Melissa Ann Betcher is planning assistant at the Office of Planning and Development, Library of Congress, Washington, D. C. 20540. No copyright is claimed on this artcile, which the authors wrote as part of their official duties as employees of the U.S. government. 385 386 College & Research Libraries trons may ,search the BlliL file by author, title, or subject. When the citations from a particular search are displayed, those hav- ing the text available online are noted. If the text is available, the reader can then call up images of the document on a black- and-white high resolution screen. The res- olution is 150 by 300 lines per inch, dis- playing each image with approximately 4.7 million bits which, by using an "or- dered dithering" process, approximates grey scale. Reader stations for the digital system are located in four of the library's reading rooms: Law, Main, Newspaper and Cur- rent Periodical, and Science. A fifth sta- tion is in the Congressional Research Ser- vice (CRS). Users have the option of reading the document on the terminal or printing it. The Newspaper and Current Periodical reading room has a Xerox 2700 laser "convenience" printer, permitting the printing of the displayed page. Pa- trons in all stations can command a laser printer Xerox 5700 centrally located in the computer center. The resolution of the two printers is identical, and printer reso- lution exceeds that of the terminal since it prints three hundred lines per inch both vertically and horizontally. The videodisc systems are not con- nected to the library's online systems but instead are "stand alone." These systems are available in the Prints and Photo- graphs Reading Room and the Motion Pic- ture, Broadcasting, and Recorded Sound Reading Room. The video images are es- sentially television images with the same low resolution as broadcast television. Consequently, they are not suitable for printed text but are satisfactory for photo- graphs. They are in color and in black- and-white, and the user can either photo- graph the screen or print out a black-and-white image if a copy is needed. OBJECTIVES A significant portion of the pilot pro- gram will be the evaluation process. For the digital system, data transaction tapes containing data captured at each terminal session will be analyzed, a~ interviews will be conducted with patrons. The first phase of the evaluation process is a test of July 1986 the system using reference staff who will be assisting patrons when the system is re- leased to the public. The goals of the staff evaluation phase are to involve a wide representation of Li- brary of Congress personnel during the implementation of this new technology and to collect data on the following: satis- faction with tr!:lining, documentation, equipment, printing, and system reliabil- ity; materials to be scanned; and general satisfaction. METHODOLOGY In order to meet these objectives, staff opinions on the system were solicited through both voluntary and formal meth- ods. Following installation of the optical disk terminals in late March 1985, each staff member received approximately a half hour of individual training. At that time logs were placed near the terminals, and a message center phone was insti- tuted. During March and April, informal opinions of the staff were collected as they began to use the system. When staff mem- bers signed onto the system, they were asked to note the time and date on the log as well as any impressions of the system. In May, the staff participated in a more formal evaluation of the system by com- pleting three questionnaires. The ques- tionnaires, entitled "Training and Docu- mentation,'' ''Equipment and Software,'' and ''Printing,'' were designed to address the objectives of the pilot program. Each questionnaire included a log. People were asked each time they accessed the optical disk system to note the system's status. The questionnaires were pretested, re- vised, and distributed to all professional staff from the Science and Newspaper and Current Periodical reading rooms. All staff from the Main Reading Room, Te- lephone/ Correspondence/Bibliography, and Automation and Reference Collection Sections of the General Reading Rooms Division were also queried. Staff in the CRS participated voluntarily. System trainers in the reading rooms distributed and collected the surveys. Respondents' identities were anonymous to the evalua- tion team. The distribution and return rates for the questionnaires are given in ta- Research Notes 387 TABLE 1 QUESTIONNAIRE DISTRIBUTION AND RETURN RATES Distributed Training & Dotumenta- lion Equipment & Software Returne~rin~~eturned De£artment (Each Survel) Returned % Returned Returned % Returned Main Readidf Room 25 25 100% 25 100% 24 96% Science Rea ng Room 20 18 90% 18 90% 18 90% Newspdter an Current Perio teal Reading Room 17 17 100% 17 100% 17 100% Total 62 60 97% 60 97% 59 95% Congressional Research Service* 20 9 45% 6 30% 4 20% *Staff in CRS participated voluntarily . Consequently, the totals for this table are given separately, and CRS results are not included in other tabulations. ble 1. As a follow-up, interview sessions were held to allow staff to express their views directly to the evaluation team. In- terview participation was voluntary, with 18 percent attending. RESULTS All of the questionnaires and test prints were analyzed by the authors. Many an- swers gave new insight into perceived and actual drawbacks to the optical disk sys- tem itself or to its current environment and in the library. Ninety-two percent of the staff was pleased with their optical disk training. Of those desiring more in- struction, there was an even split between those interested in specific optical disk manipulation and those needed assist- ance in using the retrieval software. A half hour of individual training appears to be the minimum time needed to integrate a new system into the work flow. More than half of the respondents felt that the pub- lic's previous knowledge of Library of Congress databases would affect how suc- cessfully they could conduct a search on the optical disk system. When comparing the manual and the online instructions, eight out of fifteen respondents who vol- unteered comments said that a hard copy version is more useful, particularly for the naive user. At this point there has been no experi- ence reading text from the screen for long periods of time. Sixty percent of the re- spondents have only read text from the screen for one to five minutes, with an- other 32 percent reading from for six to fif- teen minutes (see table 2). However, when asked how long they felt it would be possible to read comfortably from the screen a wide range of answers was re- ceived. A small percentage appeared at each end of the range (15 percent for one to five minutes and 13 percent for more than sixty minutes), with a more even spread in the middle range (34 percent for six to fifteen minutes and 30 percent for sixteen to thirty minutes). However, with 79 percent of respondents believing that thirty minutes is the maximum amount one can comfortably read from the screen, it becomes apparent that the librarians perceive the optical disk primarily as an access tool. Of those people citing problems while reading from the screen, 57 percent en- countered difficulties with text, and 56 percent had trouble with graphic images. Four people complained of lines through TABLE2 READING TEXT FROM THE SCREEN Time Period Longest Time Longest Time (Min .) Actually Used* Possible to Use+ N 48 47 1-5 60% 15% 6-15 32% 34% 16-30 8% 30% 31-60 0 8% More than 60 0 13% *What is the longest approximate time you have read text from the screen? +How long do you think you would be able to read text comfortably from the screen? the text, and three stated there was a loss of print or black sections on the screen (ta- ble 3). Another four staff members sug- gested that the screen be tiltable in order 388 College & Research Libraries July 1986 TABLE3 PROBLEMS READING FROM THE SCREEN Text* Graphicst No o/o No o/o Respondents citing no problems 20 43 20 44 Res~ondents citing one or more problems 26 57 25 56 otal 46 100 45 100 Problems Citiedt. 26 25 Screen flickers 38 16 Surface glare 46 32 Blurry image 27 28 Images too bright 23 Ima~es too dim 4 Una le to read small print 35 Fine detail missing 44 Need color to interpret display 24 Halftones indistinct 40 Other 42 8 *Have you encountered problems while trying to read text from the image screen? tHave you encountered problems while looking at graphic or picture images? +Respondents saying they had difficulties were asked to check all appropriate items . to eliminate some of the difficulties in reading. A tilting screen might have as- sisted those respondents who felt that the angle and glare on the screen contributed to the difficulties in reading small print. Glare was cited as the primary difficulty by 46 percent of those responding that they experienced problems. Staff had dif- ficulties with viewing color images, with 40 percent of those answering stating that the half-tones were insufficient to repre- sent the colors, thus making interpreta- tion difficult. Print size and the lack of a zoom feature caused difficulties with both text and pictorial materials, with 35 per- cent saying they were unable to read fine print and 44 percent saying fine detail was missing in pictures or graphs. Similar comments were received on questions dealing with print resolution (see table 4). Although 68 percent were satisfied with the overall quality of the printed copy, 42 percent still felt that fine details were illegible. Those items that tended to be illegible included bar graphs, some photographs, characters on multi- color images, and fine details such as cur- sive writing on old documents. As a means of testing printer reliability, ''test'' printers of a specified article were ordered by respondents and sent to the evaluation team. Twenty-nine prints were received, and their clarity, darkness, and legibility · were identical. The print resolution was equal to or better than the screen resolu- tion. It should be noted, however, that the article was straight text. In another part of the evaluation, staff were asked to send the evaluation team an offline print of their choosing. Twenty-seven were re- ceived . Staff cited several problems with . the clarity and resolution of these images, although printer and screen resolution were supposedly identical. The evalua- tion team compared the prints with the images on the screen and discovered that print and screen images were indeed iden- tical and that those items identified as be- ing illegible in the print were also hard to discern on the screen. Variations noted in copy contrast could be attributed to the mechanics of the photocopy machines. Staff members in the Newspaper and Current Periodical Reading Room were also asked to comment on the conve- nience printer . Percentages on overall quality and fine detail legibility were com- parable for convenience printing and off- · line printing. Staff often stated a prefer- ence for one type of print over the other. Impressions about which copy was of bet- ter quality varied. Since the printers are identical, except for speed, these com- ments raised new questions for the re- searchers. The evaluation team tested this by printing five articles offline and, at the same time, doing a convenience copy. A comparison revealed that print resolution was identical. However, those items printed offline have a darker copy con- Research Notes 389 TABLE4 PRINTER RESOLUTION Satisfied with overall quality* Yes No Reasons for dissatisfaction+ Text blurred Black lines Too dark Too light Too cfi.opped off Bottom cfi.opped off Poor contrast Other Fine details legiblet. Yes No Not applicable Copies legible§ Yes No Comparison with screen imagell More legible Less legible Same * Are you satisfied with the overall quality of the image? Offline No 35 24 11 11 36 Convenience % No % 13 69 9 69 31 4 31 4 27 75 45 25 18 0 9 25 45 0 18 0 73 0 18 0 12 53 58 42 42 5 0 13 92 8 11 18 18 64 t Respondents saying they were dissatisfied were asked to check all appropriate items . t To the best of your knowledge, are the fine details legible? § In your experience are the copies made by the convenience printer legible? n How does the printout compare to the image on the screen? trast, making some details more legible and others less so than the convenience copies. This could explain staff percep- tions of a difference in the quality of prints. DISCUSSION An essential part of the evaluation was to test the system's reliability, an aspect attempted through the logs located by the terminals and those in the questionnaires. Logs from March and April indicate that half of the recorded attempts to access the system were successful. Figure 1 shows the data recorded in May as people ap- proached the system to complete their questionnaires. This data includes infor- mation from the Congressional Research Service not included elsewhere. During the twenty days of testing in May, 111 at- tempts were made to access the system. These data are hard to substantiate since the system is a series of complicated com- ponents interrelated with other library equipment and software. If any one com- ponent is not available, the patron caimot access the optical disk system and can only guess which component is not avail- able. Use of both the terminal and the questionnaire logs was sporadic, and any findings are inconclusive. Scheduled stress tests and data from the question- naire logs indicate that the system works well when multiple users (up to five, the current maximum) are signed on. Engi- neers in the Automated Systems Office, responsible for system maintenance, esti- mate that the system's availability is com- parable to that of other online systems. A further part of the pilot project is to help the library determine how to use the optical disk technology optimally. Staff felt that the optical disk was not suited for high-use current periodicals unless many more terminals were installed (twenty in the Newspaper and Current Periodical Reading Room, for example). One alterna- tive was to scan specialized high-use items such as the Physicians Desk Reference and other items that are hard to control, such 390 NO. OF ATTEMPTS 16 14 12 10 College & Research Libraries BIBL and Optical Disk System up Optical Disk System down BIBL down July 1986 TIME FIGURE 1 System Reliability-May (Total Number of Attempts = 111) as the Congressional Record. Staff did not uncover any major flaws in the technology, and their feedback gave the library useful information on how to change the environment to enhance use of the technology. A large percentage of peo- ple experienced problems viewing docu- ments on the screen. The causes of these difficulties are often unclear, although dif- ficulties usually concerned graphic or pic- torial materials, not straight text. Some of these problems may be alleviated by alter- ing the work station, as by installing tilt- able screens. Further, staff offered many specific suggestions for improving the printed and online documentation. Until the technology can affordably pro- vide color or enlargement of details on the screen, the library may wish to use discre- tion in choosing color or small print mate- rials for scanning. The problems experi- enced with printer quality may also be alleviated somewhat by judicious choice of materials to be scanned. Current proce- dures require that users of the remote print facility receive printouts the next day. The impact of this limitation will be strongly affected by the materials chosen for inclusion on the disk-for example, are copies available more quickly elsewhere in the library? The staff's opinions about good candidates for inclusion were mate- rials that are hard to obtain elsewhere in the library. This would include specialized and hard to control items (mentioned pre- viously) as well as items locked up for their own protection (such as the comic book collection and the pulp fiction collec- tion), pamphlets, association newsletters, and foreign language technical materials, among others. Staff members look for- ward to the time, not too far from now, when Library of Congress patrons can sit at a terminal and search and access the li- brary's resources easily. SELECTED BIBLIOGRAPHY ON OPTICAL DISK TECHNOLOGY ~ Criswell, Lela Beth. "Serials on Optical Disk." Library Hi Tech (Winter 1983), 17-21. Hahn, Ellen Z. "A Report on Print Project Activities." Library of Congress Information Bulletin 42:37 4-76 (Oct. 13! 1983). Research Notes 391 Information Systems Consultants. Videodisc and Optical Digital Disk Technologies and Their Applications in Libraries, A Report to the Council on Library Resources. Washington, D.C.: Council on Library Re- sources, 1985. Krayeski, Felix. ''Transition of an Image System-From Paper to Microfiche to Optical Disk.'' Journal of Imaging Technology 10:161-62 (1984). Lunin, Lois, ed . "Perspectives on Videodisc and Optical Disk: Technology, Research, and Applica- tions.'' Journal of the American Society of Information Science (Nov. 1983), 406-40. Nugent, William. ''Applications of Digital Optical Disks in Library Preservation and Reference.'' Pro- ceedings of the American Federation of Information Processing Societies (AFIPS), 52:771-75 (1983). Nugent, William. ''Optical Disk Technology.'' In Minutes of the One Hundredth Meeting of the Association of Research Libraries, Scottsdale, Ariz., August 6-7, 1982 (1982), 71-79. Nugent, William, and Jessica Harding. "Optical Storage of Page Images and Pictorial Data- Opportunities and Needed Advances in Information Retrieval." Proceedings of the 1983 Annual Con- ference of the Association for Computing Machinery, New York, October 24-26, 1983 (1983), 79-82 . Price, Joseph. "Optical Disk Pilot Program at the Library of Congress." Videodisc and Optical Disk, 4:424-32 (Dec. 1984). Thompson, Mark. "High-Tech Library." American Way (Sept. 1984), 36-40. The Effect of Service A ware ness on Survey Response Ruth A. Page II and Edward J. Lusk The authors had examined the relative effect of university or library sponsorship in a 1984 sur- vey of computer utilization of Wharton School M.B.A. 'sand faculty. In the intervening year, many new services have been offered to the M.B.A.'s. The survey was redistributed. The response rate to the library was significantly higher for students who had used the new ser- vices than for those who had not. This result suggests that the interaction of sponsorship and the respondent group, not sponsorship it- self, is the critical variable. A possible sex bias between the male professor and female librarian was found not to be significant. In 1984 the authors conducted a survey of students and faculty of the Wharton School to determine the extent to which personal computers were used for online searching. In developing the survey, sponsorship became an issue. Previous re- search on sponsorship suggested that if there is a perceived prestige difference be- tween sponsoring groups, there may be a significant difference in the response rates. In general, the more prestigious sponsor elicits the higher response rate. Sponsorship as a factor affecting ques- tionnaire return rate has been examined in numerous studies in the literature. Doob and Peterson both found that university sponsorship increased response rate over business firm sponsorship by 10 and 13 percent, respectively . 1' 2 Peterson sug- gested that sponsorship was the domi- nant factor influencing returns. 3 Jones and Linda also found that university sponsor- ship significantly increased response rates. 4 Several other studies are cited by W. Jack Duncan in his review article. 5 Fur- ther, Jones, based on previous studies by Scott and by Sudman and Ferber, sug- Ruth A. Pagell is head of public services at the Lippincott Library of the Wharton School, University of Pennsyl- vania, Philadelphia, Pennsylvania 19104. She is also an associate professor at Drexel College of Information Stud- ies and instructor of decision sciences at Wharton. Edward J. Lusk is associate professor at the Wharton School, University of Pennsylvania, Philadelphia, Pennsylvania 19104, and director of biostatistics for the Cancer Cen- ter, Hospital of the University of Pennsylvania.