COST COMPARISON OF COMPUTER VERSUS MANUAL CATALOG MAINTENANCE 159 John C. KOUNTZ : County of Orange Public Library, Orange, California Is a computer assisted catalog system less expensive than . its manual counterpart? A method for comparing the two was developed and ap- plied to historical data from the Orange County Public Library. Com- parative costs obtained were $ .89 per entry for computer assisted catalog maintenance versus $1.71 for manual maintenance. INTRODUCTION Since November 1965, the County of Orange Public Library has per- formed all acquisitions by means of a computer assisted system. As a by- product of this continuing operation, records for over 30,000 titles are now available in machine readable form on magnetic tape. The next logi- cal step to realize the Library's goal of mechanizing a major portion of its many nonprofessional functions is the production of a comprehensive multi-access list of its holdings suitable for both Library and patron use; in short, a Book Catalog. The 30,000 captive entries, however, comprise only a quarter of the Library's total holdings of 120,000 titles. Before the envisioned Book Catalog can be produced, approximately 90,000 titles remain to be captured, and subsequent file handling and data printout operations must be developed. An undertaking of this magnitude naturally prompted a review of the literature. Initially, Hayes and Shoffner's work for the Stanford Univer- sity Undergraduate Library ( 1) would appear adequate. On closer exam- 160 Journal of Library Autouwtion Vol. 1/ 3 September, 1968 Fig. 1. Manual Card Catalog System. CEM~\.\UO 0~'\lii..,.IOtl~ (C~ ... "'"'""'C.T\OH) Cost Comparisonj KOUNTZ 161 Fig. 2. Proposed Computer Assisted Book Catalog System. 162 Journal of Library Automation Vol. 1/ 3 September, 1968 ination, however, their approach did not optimize the cycle for supple- ment production or catalog reprint; nor was particular attention given this problem in the Institute of Library Research Report to the California State Library ( 2). The Cartwright and Shoffner Study for the California State Library ( 3) paid close attention to cycle length, but the system therein described differed extensively from the system proposed for Orange County. Further, though the costing of data capture has been well documented and continues to appear in the literature ( 4,5,6), there is little concerning the cost of maintaining data once on file. In brief, neither a method nor basic information was available which could be ap- plied generally, although several specific approaches and results had been presented (1,7,8,9,10,11), and an approach to the analysis of manual op- erations established ( 12). When it became apparent that more than article reading .would be required, cost analysis of the existing manual operation and the proposed computer assisted Book Catalog program was performed. In addition, a method was designed to discern what cost benefit, if any, was implied in a computer maintained file before a massive keying effort and systems development should be undertaken. It is important to note that the analy- sis gives no consideration to increased level of service, esthetics, practi- cality, or the subsidiary products of a computerized system. Nor is the capital investment represented by existing card catalogs considered, as those units are assumed to have been paid for in the course of their creation. MANUAL CARD CATALOG SYSTEM The manual system to be replaced consists of individual card catalogs and shelf lists in each of the Library's service units, comprising 25 branches and a separate Bookmobile base. This system, depicted in Figure 1, consists of: 1) centralized card pro- duction, and; 2) branch catalog maintenance. In the centralized opera- tions, offset masters are created from worksheets prepared by the cata- loging section and used for two-up card production. These cards are col- lated into sets, the sets merged with their corresponding books, and the completed packages sent to the ordering branches. When book and card packages are received by the branch, shelf list and catalog cards are sorted and merged with their respective files. Withdrawal of a book (dis- carded or lost) from a branch collection triggers a reversal of this process, and all cards for the withdrawn item are purged from the files. PROPOSED COMPUTER ASSISTED BOOK CATALOG SYSTEM The computer assisted system (Figure 2) consists of tlll'ee phases of computer operation and catalog printing. In the first phase the computer receives as input magnetic tapes produced by the Library's ongoing book acquisition system and/ or the output of a device providing a direct key- Cost Comparisonj KOUNTZ 163 board to tape capability, processes the input data into updated records, and merges the updated records with the Master File of Library holdings. The first phase will build the initial Master File through capture of the Library's remaining 90,000 titles via the keyboard-to-tape device in- dicated above, and will also form the main avenue for communicating revision (update) data to the Master File. In the second phase the computer extracts two print tapes from the Master File: the first is the Biblio File, consisting of all the bibliographic data and the record number for each Master File entry in alphabetical sequence (author-title mix); the second, or Locate File, contains location codes and copy counts for each record number in numeric sequence. In the third and final phase, the Biblio and Locate Files are processed. From tl1e Biblio File are produced keylines (camera-ready copy) for the Book Catalog and periodic cumulative supplements of new entries. Out of the Locate File are generated three numeric listings: 1) a Locate List containing all entries, 2) periodic cumulative Locate Supplements, 3) Branch Inventories. In production of the Book Catalog, the computer produced keylines are used to create offset masters for printing. The end product of the printing process is 400 bound copies of the Book Catalog. FACTORS IN COST COMPARISON Following is an examination of the principal factors which must be equal or identical to permit comparative analysis of the two systems. Unit of Comparison (ENTRY) To facilitate the cost analysis between manual and computer assisted file maintenance systems, a unit of comparison was established which would be compatible to both. This unit is called the ENT}\Y, and in the analysis which follows is the basis for all cost comparisons. For the manual system, ENTRY means creation, distribution, filing and, ultimately, purging of the complete set of cards (Figure 3) pertaining to a specific book; while for the computer assisted counterpart, an entry is a record (Figure 4) in machine readable form which has been captured, sorted, listed and updated. Frequency of Transactions Either system, in addition to creating and posting new records to a file, must periodically update both entire records and the elements of those records. The number of these updates can be determined for a given period of time, and for our purposes we call this figure the fre- quency of transactions. With regard to the systems under analysis, the frequency of transactions is identical, and includes two elements: titles added and withdrawn; and volumes added and withdrawn (including re-assignments) as shown in Table 1. 164 Journal of Library Automation Vol. 1/ 3 September, 1968 Don baa 940.5472 940.5472 Sandulescu, Jacques Donbas. McKay, 1968. 217p $4.95 ESCAPES Sandulescu, Jacques Donbas. McKay, 1968, 217p $4.95 WORLD WAR, 1939-1945 - PRISONERS AND PRISONS, RUSSIAN - PERSONAL NARRATIVES 940.5472 ., 940.5472 Sandulescu, Jacques Donbas. McKay, 1968. 217p $4.95 Sandulescu, Jacques Donbas. McKay, 1968, 217p $4.95 5217S3 940,5472 940.5472 Sandulescu, Jacques · Donbaa. McKay, 1968, 217p $4.95 Sandulescu, Jacques Donbas. McKay, 1968. 217p $4.95 521763 1. World War, 1939-1945 - Prisoners and prisons, Russian - Personal narratives (WO 63866) 2. Escapes (ES14042) I, T 0 68-14127 Fig. 3. Set of Catalog Cards. 1- 1-- I. ~ M~~fR. ~ Rl::c.olitc z ~ ... 0 "' ...J "' a: "' • "' :> z 0 c: ... 0 "' ...J "' a: "' .. 2 ::> z 0 ~ <> "' .J "' 2. ::; NA.!'l!: ; - :> Sv~!:>)E<.T ~ REc.oRO:: 0 Fig. 4. "' .J "' DEP ARTM ENT ~' . . . MULTIPLE LAYOUT FORM FOR ELECTRIC ACCOUNT ING MAC HINE CARDS INTERPRETER SPACING N'-MI:: / SvaJE.CT C.ODES LC./OC. NVM8t:R • . ~ IU...F-. o-n l.tNnnc."n•• 0 c.o~ coac C.ODt c.oot. c.oo~ c.oot:. P1tlrt• NV"at:" ~ ~ ~ ~ ~ ~ ~ S' 0 0 0 0 0 0 EFFECTIVE DATE----- FILING- TITLE. v tt « « tt. II It 9 9 919 919 9 9 9 9 919 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 919 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 I 2 3 4 S 6 1 I S 10 11 I Jl 14 15 1l 11 II II 10 21 2 23 24 25 2l 27 2t 1'! 30 )I 32 ll 34 35 3C l1 31 Jt 41 41 4~43 44 4S 46 47 41 0 $0 51 52 S3 S4 55 ~ 51 S. 5t SO 61 ' 2 6J 64 5$ "51 61 " 70 71 11 1314 75 11 7171 7t to FIUNIPTITI..E (C.ON'r.) SVS·T\Tl.E: 9 9 9 9 9 9 9 9 9 9 9 919 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 1 ~ 3 4 5 I 7 I t \ 0 II 12 ll 14 15 " 11 II It )(! 21 n ll 24 2'5 K 1l 21 2! 30 ll l2 33 :W l S 3l 31 31: ll 40 41 42 4J 44 '54& 41 U 49 SO 51 52 5J S4 5.5 ~ 57 5I St 50 61 CJ 63 S4 65 6' 51 M tt 70 71 ll 7l 74 15 71 71 1111 • 5UB·TtTL.E (.c.• .. T - ) D"-TE Jfvelfi$4.-FieS A~ Pte /SONS 1'?1):.~ •1\N - Pt"IU~N~ t N~tttt'A11VGS(Wll ".39,6) ;). • Fsc"PJ:s (Es 11./oi.l~) STRIPS: 6 PRICE New Ti tle 0 Recol. [=:J End Papers New Ed. D New Set c=J Bind Have D l abels I ,._., Re info rce F850-66 . 2 Bind ing: TRADE Closs: 9 01 02 1 20 21 32 53 54 55 56 61 Fig. 5. Sub Purchase Order. ANF . 1 63 71 72 73 Cost Comparisonj KOUNTZ 171 ADULT TITLES ON OROER BRANCH 11 LAGUNA BEACH 04-30-67 PAGE 3 01933.5 ARCO 03-67 $6.50 SCORING HIGH BN READING TESTS 019413 ARCO 03-67 1 S4.00 VOCABULARY, SPElliNG, GRA1<4MAR 022015 ARMSTRONG, CHARLOTTE 05-67 1 $4.95 THE CIH SHOP 022483 ASHLEY ~ONTAGU, M. F. 05-67 1 $5.95 AMERICAN WAY OF LIFE 017896 ASHUY-~ONTAGU S4.50 ON BEING HIJMAN 021535 ASI"'OV, ISAAC 05-67 1 S3.75 WHLSPRINGS OF LIFF 022599 ATTWOOD, . WIL~IA~ 05-67 1 $5.95 THE RF.DS & THF BLACKS 020713 AUCfiiNCLOSS, LOUIS 04-67 1 $4.95 TALES OF ~ANHATTAN 019960 AUER, ALFONS 03-67 1 S5.95 OPEN TO THE WORLD 022460 AUSTEN, JANE 05-67 S5.95 PRIDE & PREJUDICE 018680 AUSTEN, JANE 1 S3.95 EM'4A 021536 BAK F.R, GEOFFREY 05..;67 1 Sl5.00 MOTFLS 021666 BALOGH. THAS 05-67 $7.95 ECONOMICS OF POVERTY 018116 BANNISTER, MARGARET $4.95 6UP.N TfiE LITTLE LAMP 022093 BARING-GOULO, wiLLIAM 05-67 ss.oo THE LURE OF TH£ LIMERICK 018392 BARLOW, JAMES 1 s5.95 ONE MAN IN THE WORLD 021188 BARNETT, A •. OOAK 04-67 1 $6.00 CHINA AFTEI-t MAO Fig. 6. On Order Listing. 172 Journal of Library Automation Vol. 1/ 3 September, 1968 is taken from historical costs for the operation of a report generator doing this in the acquisitions system. Similarly, the reformatted entries in "printable" form must also be : sequenced alphabetically (single ·author- title mix) before they can be printed, and again the $0.00034 cost is taken from historical data. Finally, the sorted, reformatted entries are printed (upper case) at a cost of $0.027 each ( 2 lines). The total cost for these operations becomes $0.036 per entry, as shown in Table 6. Table 6. Keyline Production Cost Computer Reformat Master File Entries Sort Reformatted Enh·ies Print (Offline) TOTAL Entry Cost $00.009 00.00034 00.027 $00.036 For catalog printing, the computer generated keylines will be reduced photographically to 60 percent, and the reductions assembled for 16-up reproduction with approximately 100 entries per sheet (both sides). Ini- tial book catalog production will be 400 copies of approximately 1,800 sheets. There are slightly more than 61,000 author entries which will re- ceive full bibliographic data, call and LC order numbers, while the 120,000 title entries will present only author data and call number. The estimated total cost of printing is given in Table 7. The resultant. cost per entry is $00.186, regardless of the number of lines required. Table 7. Catalog Printing Cost Set-up Plates Run Time Gather/ Collate Paper Cover / Perfect Bind TOTAL: $ 3,000 7,000 6,000 2,000 4,000 350 $22,350 As the printed and bound Book Catalog will not present the locations of the materials it lists, an off-line Locate List will be produced concurrent with catalog creation. This list will contain 120,000 numeric entries ( LC order number, coded locations and price), and will be generated for Library use only. The cost of offline printing of this list ( 25 copies) is based on a historical print cost of $0.014 per one-line entry extended to the number of entries, or $1,680.00. Cost Comparisonj KOUNTZ 173 Summary of Computer Assisted Catalog Production Costs The grand total cost per entry for all operations leading to the initial book catalog (based on initial data capture through file construction above) is given in Table 8. As shown in this table, the computer cost per Table 8. Composite Cost Per Entry of a Computer Produced Book Catalog Operation File Construction Keyline Production Book Production Locate List Production TOTAL Cost $0.286 0.036 0.186 0.014 $0.52 entry for the first 'edition' of the book catalog is $0.52 .. This figure is com- parable to the manual system figure of $0.99 per entry derived earlier. However, the cost per entry figure for computer assisted file maintenance must also be derived before comparison with the total manual figure of $1.71 per entry is possible. Table 9. Cost of Posting and Printing Catalog Update Data Operation Unit Cost Total Locate Update Input Print Locate List (offline) Subtotal TOTAL for 133,000 actions Biblio Update Reformat Master File Entry Sort Reformatted Entry Print Biblio List (offline) Subtotal ( rounded) TOTAL for 9,000 actions GRAND TOTAL ( 142,000 actions) $00.007 00.014 00.021 00.009 00.00034 00.027 00.036 $2,793.00 324.00 $3,117.00 174 Journal of Library Automation Vol. 1/ 3 September, 1968 Computer File Maintenance The figures developed in Table 9 establish a cost per entry for file maintenance, from keyboarding corrected data to the production of an offline printout of Biblio and Locate supplements. To understand their derivation, let us review the frequency of transactions. In Table 1, it can be noted that the 15,000 titles added to the collection annually will ne- cessitate Master File location update for the volumes they represent. The locations for withdrawals will also require update. In combination, addi- tions and withdrawals mean a total of 129,000 actions, plus 4,000 last copy withdrawals, or 133,000 updates yearly to keep the Master Locate File current. In contrast, only the 9,000 new titles will require bibliographic listing. The Locate File update input cost is identical to that used for the entry of error correction data (Table 5). This is possible since approximately the same number of characters must be keyed to address an entry and enter updated location data ( 9 characters for record and card code, an action digit, and 2 numeric location characters for a total of 12 characters versus the 14 characters required for entry correction) . Similarly, the off- line print cost for Locate data remains the same as that indicated for the initial Locate List printout: $0.014 per entry. The Biblio File update costs are the computer keyline production fig- ures presented in Table 6. To derive the cost per entry, both Locate and Biblio figures are extended to reflect the proportion of the final figure they represent, and reduced to a single cost per entry. In summation, $0.0242 is the cost per entry for file maintenance for one year. However, this figure is of limited value without reference to either the frequency of supplemental production or total catalog reprint period. Therefore, the optimum frequency of supplement production and the period of mainte- nance are discussed below to bring this raw $0.0242 per entry into per- spective. Optimum Frequency of Supplement Production and Catalog Reprints The optimum frequency of bibliographic supplement production is based on the most timely reporting of new title disposition at the least cost. That is, a determination of the number of cumulative listings of new titles in concert with all location changes which can be produced before their production cost equals or exceeds the cost of total catalog reprint. The most economical approach to reporting revised, new, or deleted bibliographic and location entries would be through listing only those entries which have been changed. The summary figures presented in Table 10 reflect only the cost per entry developed in Table 9 for the pro- duction of cumulative exception listings, assuming an equal monthly dis- tribution of transactions. In addition, the annual cost per year, excepting the tweHth month, is tabulated to reflect overall cost where total reprint - Cost Comparison/ KOUNTZ 175 would occur instead of last cumulative supplement cycle. A quarterly supplement production cycle is selected, as it best meets the optimum defined earlier (i.e., most timely reporting for the least cost). Table 10. Cumulative Supplement Costs for Various Cycles Computer Runs Per Year 12 6 _..,.. 4 3 2 1 Annual Cost @ $0.0242/Entry 12th Month 12th Month Included Excluded $ 22,335.92 $ 18,899.63 12,027.04 8,590.74 8,590.74 5,154.44 6,872.26 3,436.30 5,154.44 1,718.15 3,436.30 By extending the quarterly supplement production costs shown in Ta- ble 10 to represent recurring annual expenses and cumulating these an- nual expenses for comparison with the total cost of complete book catalog and Locate List product, the number of years between catalog reprints becomes obvious. This calculation is shown in Table 11, where 3 years is the optimum reprint cycle for the qua1terly supple.ment costs selected. Table 11. Years I 2 _..,.. 3 4 Catalog Reprint vs. Supplement Production Costs Supplement Cost Annual Cumulative $ 8,591 17,182 25,773 34,364 12th Month Excluded $ 5,154 13,746 22,337 30,928 (Year's End) Catalog Reprint $ 23,000 24,250 25,500 26,750 COMPARISON AND CONCLUSION To return to the cost per entry for catalog maintenance alone for opti- mum reprint cycle, there is a total outlay of $47,837 for 3 years of cumu- lative supplements and a catalog reprint to report an average of 129,000 titles. From this base can be derived a cost per entry of $0.37 for entry maintenance. This $0.37 can then be summed with the $0.52 cost per entry for the catalog "first edition", for a grand total of $0.89 as the cost per entry for a computer assisted catalog production and maintenance system. Further, this cost per entry is realized in a document equal to 400 card catalogs! In terms of the manual system, maintenance was $0.72 per entry, and some 26 files had to be maintained. Thus, it is possible to extend the single file maintenance cost to a systemwide average of $18.72, plus the $00.99 required for entry preparation, or a grand total of $19.72 per entry, rather than the $1.71 indicated earlier. 176 Journal of Library Automation Vol. 1/3 September, 1968 The lesson implied here is simple: manual cost per entry is dependent upon the number of manual files being maintained. This is of importance since it means a significant increase in outlay for file maintenance with the addition of each new branch; whereas, costing for a computer pro- duced and maintained catalog is relatively independent of the number of service units accommodated. Finally, a word of caution. There is a potential danger lurking in these figures for the small public library which has a limited number of branches. This is the fact that the cost per entry, even for the single shelf list/card-catalog comparison, has been calculated for an operating system serving a relatively large number of branches. The cost-per-entry method used in this paper does not include amortization of the capital outlay for "computerization" which, in this specific case, amounts to al- most $200,000 for design of system, procedures and forms, and for design, coding and debugging of programs. Although savings equal to this amount, or more, would be realized over a period of time because of reduced clerical operations and attendant burden, a large sum would still have to be earmarked for expenditure during a relatively short period with no immediate return. Foreknowledge of this "one-shot" cost and its related cost-per-entry payoff should not be a deterrent. Rather, it should permit the administra- tor of a limited operation to deal effectively with increased clerical costs and to make meaningful decisions relative to service bureau overtures, library board interrogations, or the goals of a new library system. REFERENCES 1. Hayes, Robert M.; Shoffner, Ralph M.; Weber, David C.: "The Eco- nomics of Book Catalog Production," Library Resources and Techni- cal Services, 10 (Winter 1966), 65, 68-82, 87-88. 2. University of California, Institute of Library Research: Report to the California State Library Preliminary Evaluation of the Feasibility of Mechanization (Institute of Library Research, University of Califor- nia, 1966), p . 3-6. 3. Cartwright, Kelly L.; Shoffner, Ralph M.: Catalogs in Book Form: A Research Study of Their Implications for the California State Library and the California Union Catalog, with a Design for Their Implemen- tation (Institute of Library Research, University of California, 1967), p. 58-68. 4. Bourne, Charles : Bibliographic Data Conversion Techniques (Mimeographed tables presented at Oregon Library Mechanization Workshop, June 1968) , Table II. · 5. Chapin, Richard E.; Pretzer, Dale H.: "Comparative Costs of Con- verting Shelf List Records to Machine Readable Form," Journal of Library Automation, 1 (March 1968) , 71. L Cost Comparison j KOUNTZ 177 6. Black, Donald V.: "Creation of Computer Input in an Expanded Character Set," ] ournal of Library Automation, 1 (June 1968), 117- 118. 7. Fasana, Paul J.: "Automating Cataloging Functions in Conventional Libraries," Library Resources and Technical Services, 7 (Summer 1963), 358, 361-365. 8. Robinson, Charles W.: "The Book Catalog: Diving In," Wilson Library Bulletin, 40 (November 1965), 265-268. 9. MacQuarrie, Catherine; Martin, Beryl L.: "The Book Catalog of the Los Angeles County Public Library; How it is Being Made," Library Resources and Technical Services, 4 (Summer 1960), 225-226. 10. Heinritz, Fred: "Book versus Card Catalog Costs," Library Resources and Technical Sm·vices, 7 (Summer 1963), 231-236. 11. Smith, F. R.; Jones, S. 0.: Card Versus Book-Form Printout in a Mechanized Library System, (Douglas Aircraft Company, 1967; Clearing House Document #AD 653 697), p. 7-8. . 12. Wynar, Don: "Cost Analysis in a Technical Services Division:· Li- brary Resources and Technical Services, 7 (Fall 1963 ), 320-326.