The amount of programming required for the merger would determine the need for a new bibliographic loader at ITD; if the records were “clean” enough, they could be uploaded with the regular ongoing OCLC loader routine. With this in mind, ITD estimated a preliminary timeline for integrating the database into Aleph at approximately four months.

At this point, the author was brought into the process. She had been a librarian at BHU for many years and had joined the TU library six months previously. The author had served as BHU’s cataloger and was familiar with the original implementation of BHU’s ILS; she had worked with the vendor when the collection was converted to electronic form. The author had unique insight into the library’s s subcollections, holding codes, local bibliographic customizations, and the origin and significance of existing bibliographic decisions. Her institutional memory became very useful in the ensuing months. Another benefit of her presence was her relationship with her former BHU colleagues. The sharing of information pertinent to the merger of two databases flowed easily, a fact which would prove invaluable as the process progressed. Some of the files and documents that provided crucial information for the cataloging process had been written by the author. In addition, as a specialized Judaic collection, the BHU library contained a large number of materials published in foreign languages, primarily Hebrew and Yiddish. As the only Hebrew language specialist on the TU library staff, the author’s language skill was invaluable for identifying, testing, verifying, and planning for the future addition of these items.

In January 2009, the TU university librarian convened a preliminary committee meeting to discuss the details of the library merger. In light of the expected integration of BHU before the 2009 fall semester, the committee began to gather estimates from outside vendors to address the logistics of the merger. These included companies that handled the physical aspects of the move, such as dismantling and rebuilding shelving, and those that moved entire library collections. It also included those that offered multiple processing services, such as cataloging, authority control, and digitization. BHU was investigating possibilities for funding the renovation of facilities at TU and for the cost of the move. Another outcome of this initial meeting was the selection of a BHU task group at TU, comprised of staff members from content management, public services, circulation, periodicals, technical services, and facilities management. Before the first meeting, the task group began work to prepare the strategic plan for implementing the move.

The practical implications of moving a 70,000-volume collection into the existing space of another library was daunting. The first challenge was the allocation of physical space to hold the collection. The pending agreement specified that the BHU collection would be housed separately from the general TU collection, so a large defined area was required. Fortuitously, a back-burner project that would become essential to the proposed move was already in progress. One entire floor of Cook Library was currently designated for periodicals, with more than 100,000 bound journal volumes shelved on long open stacks. These volumes occupied the majority of the space on the floor, filling the central area with scores of metal shelves. A glassed-in area which housed the periodicals reading room, the periodicals service desk, microform machines, and storage cabinets for library microforms was on one side of this space. The floor also held study tables and carrels, and a small lobby with group study tables (see figure 1). This floor was cleared of the bound periodicals to make space for the BHU collection.

The pending periodicals project addressed a growing trend in academic libraries to interfile bound periodical material by call number with the monographs in the general stacks. This brought together all printed material on one topic and enhanced user accessibility to resources. On a technical level, classifying and interfiling the bound journals required several steps, including reviewing the bibliographic records for all journal titles in the print collection; verifying title changes and production cessations; assigning class numbers to records to be retained in the collection; printing spine labels with the new call numbers; and affixing the labels to 100,000 volumes. Most important in terms of time was be the massive shift in the general stacks to create space and accommodate these bound journal volumes.

Determination that there was sufficient space in the stacks was crucial to the success of this endeavor. Technical services staff managed the projections and measurements to ascertain the possibility of absorbing an influx of materials. Using a formulaic approach to estimate the space required for the periodicals, a detail-oriented library technician determined that the periodical collection could be accommodated on the upper floors by increasing the number of shelves in each range. By removing unneeded shelves from the second floor and rebuilding them on the fourth and fifth floors, the initial shifting could begin. Along with confirmation of space, another issue arose. The layout of the general stacks on the upper floors of the library was not intuitive. The head of circulation, who also managed the re-shelving and shifting, speculated that while the collection was being shifted to make room for the journals, it might be time to shift and re-route the entire collection to correct this directional flow problem. The task group also addressed the facilities aspect of the move, as the spaces on both floors needed reconfiguration. The second floor shelves consisted of two sets of ranges, each many stacks deep, stretching across the room. The new configuration that would hold the BHU collection required removal of one of these sets; the remaining one would be extended by two sections to accommodate the BHU volumes. Additional sections of shelving were required on the upper floors to receive the newly integrated journals. The floor designated to house the BHU collection needed considerable reconfiguration; the number of books shelved there would be reduced by thirty percent. While one half of the floor space would be devoted to the incoming monograph collection, the other half was redesigned as study space (see figure 2).

The library administration investigated the possibility of outsourcing all or some steps of the project. The costs were deemed too steep to accommodate all of the plans, and the eventual outsourced pieces included the moving of the shelves, the shift of the collection, and the re-shelving of the bound journals. In retrospect, there were some serious flaws with the work provided by the company chosen to perform the work. The original conceptual map called for the move of approximately one-half of LC-arranged volumes to the uppermost fifth floor and the rest to the floor below. The contractor, lacking solid knowledge of the collection, encouraged moving more of the collection to the fifth floor. The library had chosen not to hire a project manager for the move and relied on estimations and configurations generated by its own staff. Because the titles could not all be processed for moving in order of their location on the shelves, it became apparent midway through the move that the upper floor would run out of shelf space. This necessitated building additional shelving, some serious reshifting in the middle of the project (involving all members of the library staff), and many overtime hours by the contractor’s workers as they raced against the deadline.

Once the journals were removed from the second floor, renovations began. The stacks area of the floor received new carpeting and paint. While the journal project and the second floor overhaul were proceeding, two other developments were occurring simultaneously. Although unrelated to the merger, the timing of another Cook Library project proved to be fortuitous for the transition from two libraries to one. In 1970, the TU library established the Department of College Archives, which collected historical archives and artifacts from the university, and included the Towson Room, designated for special collections and school archives. The library began plans to renovate and expand the existing special collections space and the Towson Room in 2007. The plans included the creation of a state-of-the-art climate-controlled closed stacks area, space for processing items, reading space for researchers, and a conference room.

This project was scheduled for completion in July 2009, one month before the BHU collection, together with its collection of rare books and artifacts, was scheduled to arrive. The Archives and Special Collections area was completed on time, and the first items from BHU to arrive were the contents of its Rare Book Room. Because of the delicate nature of these items, they were moved directly into the special collections location to unpack and process them there.

The other project that was taking place was the preparation of the BHU collection for its move to its new quarters at TU. This part of the move was under the supervision and sponsorship of BHU, and to avoid storage fees in the face of a deadline for removal from its current location, the timing was coordinated with TU and scheduled for the first week in August. All work was scheduled for completion by July 31, 2009.

The first meeting of the task group in mid-February reflected the work already in progress. Technical Services staff began to review the bound journals to determine which to classify in preparation for integration into the stacks. They analyzed the serial titles for ceased or standing orders and verified holdings of individual titles. They consulted with the reference staff about weeding the stacks in the latter’s liaison areas. With the major goal of collection integration in mind, any items no longer essential to the collection would be removed. Technical Services also worked with the Archives staff to identify bound periodical volumes that should be relocated to Archives due to age, condition, or value. Cataloging staff reviewed each title, taking note of any changes in title or production, importing new records as necessary to conform to upgraded cataloging policies, and assigning call numbers. To accommodate the need for spine labels for 100,000 volumes, the library ordered a new thermal transfer printer.

While plans were developing for the catalog merger, the TU staff member responsible for ordering and maintaining periodical subscriptions began to address this matter. BHU subscribed to approximately 250 current journal or annual titles and owned an additional 150 titles that were no longer active. The initial task was to identify any duplication with TU titles and determine the acquisition details of the subscriptions. Working from a detailed spreadsheet provided by BHU, the staff member compared current titles with those of TU. This step proved problematic for two reasons. First, the process began even before the bibliographic records migrated to Aleph; comparing records within the system would have been far easier than working with the BHU database, which was then under scrutiny for migration preparation and thereby inaccessible. Second, many of the periodical titles were in foreign languages, most notably Hebrew and Yiddish. Because these languages are written in a different script, the titles require transliteration into Roman alphabet script. Although there are standardized rules for such transliteration, the spreadsheet titles did not conform to those rules, and ambiguous spelling was not uncommon. This resulted in frequent consultation with the author and much trial and error regarding the accurate spellings for the titles. Once the titles were identified, the staff member attempted to ascertain the acquisition source for each one. TU’s primary vendor for electronic resources is EBSCO; the plan was to move as many titles as possible to this service, and if possible, to switch print titles to online subscriptions. For those titles not available through EBSCO, the staff member contacted the publisher directly. This was also initially true for the print annual subscriptions. Approximately one year after the arrival of the BHU collection, TU switched from its previous distributor for print and digital books to Yankee Book Peddler (YBP) and moved its annuals subscriptions to this vendor. If possible, titles were moved to online subscriptions.Because BHU was scheduled for closure by the end of June, it was essential to move quickly with regard to the periodicals. Staff needed to know the terms of the current subscription, including financial agreements and renewal dates. Equally important was the request to each vendor to change the address for all current titles to TU’s library to avoid the loss of issues. Some of these steps were hampered by the lack of documentation other than the spreadsheet. Because the database had not yet arrived, arranging for the renewal of some subscriptions became more complex. For example, EBSCO required an order number to process such a request, and the staff member manually input a temporary bibliographic record as a placeholder for the order; she then suppressed the record so that it would not display in the OPAC until it was replaced by an updated record. Once the collection arrived from BHU in August, staff could consult current invoices for more information about the status of the subscriptions.

Some problems still remained. There were manually updated check-in records for the periodicals, but these were not always accurate. Also, because some of the most recent issues of current titles never arrived at TU with the rest of the collection, there was lack of clarity regarding missing issues. Another ongoing problem in the attempt to reconcile subscriptions concerned overseas vendors. Communications from them sometimes arrived in foreign languages only, and time differences complicated contacting them; however, email proved useful for some of these issues. There were also some fulfillment problems because of the numbering system used by some publishers of Hebrew language materials. These differed from the familiar January-December cycle, as these publishers followed the Jewish calendar, which begins in the fall; it was not always easy for the periodicals staff to determine which issues had arrived. Transliterated titles remained a challenge for the staff to decipher. Although labeled a “summer project,” the merging of the periodicals collection took approximately two years to complete.

To prepare for the technical aspects of the database merge, ITD began its work early in February 2009. Specific data was requested: the number of bibliographic records slated for absorption, local customization information, and whether or not to expect the addition of vernacular characters. Hebrew/Yiddish characters are supported in Aleph, and replacing existing mono-lingual records with multi-lingual ones would enhance the discoverability of items in the database. ITD wanted to identify the bibliographic sources of the BHU records. While most of the records were OCLC records, approximately twenty-five percent of them came from other sources, such as TLC’s Bibliofile (an early CD-based bibliographic database), BHU’s originally cataloged records, and on-the-fly records that were created for reserve materials and not intended to become permanent. TU staff designated new sub-library/collection codes for use in Aleph to accommodate the BHU items and determined which BHU holding codes could be mapped to those new sublibrary collection codes. They also identified which codes to use for special collections items. Information that would be overlaid on loading and the mapping destination for all local information were also considerations.The technical services cataloging subcommittee consisted of three staff members, including the author: two catalog librarians and one library technician. Record analysis proceeded by working with comprehensive lists of the BHU collection provided by its ILS vendor, TLC. The subcommittee determined both the BHU holding codes and the record sources, along with the number of titles/items per holding code, and identified the MARC tags for which local notes fields should be retained. They discussed each part of the collection and its future electronic and physical location at TU. The subcommittee decided which sub-library/collection codes were needed, and the actual shelf placement for the items assigned to each of these collections. Some of these decisions could not be made until after an onsite visit to BHU, during which the TU staff could get a cursory look at the materials.

The retention of local, copy-specific bibliographic information was a consideration, and the unique implementation of local notes in the USMAI database requires a footnote in this narrative. Because the database represents the holdings of sixteen separate campuses, with a single bibliographic record shared by all campuses that own copies of the title represented by that record, it was necessary to indicate ownership of local bibliographic information by an individual campus. USMAI’s solution is to provide copy-specific information in a “superholding”—records attached to the shared bibliographic record (see figures 3, 4, 5).

The information from this superholding displays only in the local catalog of the campus that owns that particular item (see figure 6).

USMAI uses an ongoing bibliographic record loader, which processes OCLC export transactions on a daily basis. It loads into Aleph any records produced or updated in OCLC by any of the USMAI campuses. Specifications for the loader include a local tags list, but the BHU records required a tag to indicate TU ownership, and additional local fields needed to be mapped to the superholdings. This affected between four to seven codes.

BHU possessed a significant number of rare, valuable, or fragile items, which were housed in various locations other than the stacks. The bibliographic records for these items contained specific information that did not apply to the rest of the collections. Among these were local notes referring to provenance, donor source, physical condition, or physical attributes of the items. For these items, the subcommittee wanted to retain a larger number of fields (between thirteen to seventeen). Additionally, the total number of titles (more than 51,000) and volumes (more than 71,000) were ascertained, and the subcommittee provided the statistics necessary for ITD to predict costs for matching the records in OCLC and uploading them to Aleph.

Collection overlap was raised early in the merger discussions. One other USMAI campus has a large Jewish studies collection, and concern was raised whether duplicate records would result from the integration of the two catalogs. The search of a random sample of titles revealed some overlap in the general collection, but very little duplication of special or rare items. The plan had been to match duplicates upon batchload through OCLC and to compare the USMAI OCLC control numbers with the BHU control numbers, which provided the OCLC information. After examining the BHU files, ITD staff discovered that the OCLC information recorded was not straightforward. For example, several hundred records had been added to the BHU database from TLC’s BiblioFile Service. Data from this service comes from publishers, vendors, and OCLC, and the records contained fields with an identifying prefix (tlc). These would not match with OCLC records according to the specifications applied by ITD. ITD broadened the criteria for determining a match so that the incoming records would contain enough match points to reduce the number of duplicates to a minimum.

As previously mentioned, an important issue relevant to the BHU records was the prevalence of the local information they contained, such as local notes, gift sources, provenance, or the presence of signatures or inscriptions. Retention of this information was a critical public relations issue for TU, as it often identified people or places significant to BHU, and its importance would have a considerable impact on the choices available for uploading the records.

One option was to load the records directly from TLC into Aleph. There were two main problems with this approach: (1) TLC could not provide protection for any special or local notes, and these fields would be lost in the migration, and (2) uploading the records without first searching them in OCLC would leave them in the transliterated Romanized versions, even when an upgraded or better record with vernacular language existed. Any preprocessing would be prohibitively expensive and time-consuming. A variation of this option was to separate the “Special Collection” records and upload those through OCLC, allowing the bulk of the collection to remain transliterated and bereft of local information. The third option was to send all of the records to OCLC through a batchload and then upload them to Aleph. The latter choice allowed for the exchange of the value-added vernacular records and the preservation of local information, which was dependent on an initial “grooming” of the records by TLC to prepare them for extraction. The TU staff preferred the straightforward OCLC option, as it would maintain the integrity of the records.

In late March, ITD received an estimate for OCLC costs for the BHU load, which included flipping BHU’s OCLC institution symbol to reflect new ownership by TU, uploading the records to Aleph, and upgrading the transliterated records to vernacular when possible. The estimate included the retention of some local notes, but because some of the notes were not recorded in standardized form, those would be lost. Technical services staff decided to edit the records manually and to replace any information that the upload process would remove. This was dependent on either harvesting the information before the BHU database was frozen or having a copy of the BHU records for reference after the load. The information was preserved in both ways. Staff assigned a TU student worker to access various sets of BHU bibliographic records and to cut and paste them into spreadsheets in MARC format. The sets were selected based on the likelihood of a significant number of local notes in the records, such as those for the rare book collection. The spreadsheets were labelled to identify the sets and saved for later reference. Manual searching using a “find” search was necessary to access the information. The staff requested that TLC send a copy of the BHU database in its final form before migration. These records could later be accessed using MarcEdit to locate specific records.

Aleph preparations for the incoming records were underway. With the conceptual plan of integration in mind, TU submitted to ITD a detailed spreadsheet, mapping all BHU holding codes to corresponding Aleph sub-library/collection codes and clarifying which MARC fields should be retained. ITD began loading sets of sample records into the USMAI test database, and by early May, the cataloging staff began testing retrieval and display of these records. The subcommittee wanted to review the search results, location limiters, and display features before the records were loaded into the live database. The cataloging team met with ITD to discuss the details of the data extraction and the load of the BHU records, which was scheduled for later that month. The plan was to establish a timeline and specifications regarding how TLC and OCLC should handle the data.

The team established a freeze date for the BHU catalog and a tentative date for the final extract from TLC. Discussions with TLC included a clarification of data specifications and the retention of a copy of the complete TLC file by ITD. TU initiated steps with OCLC, and together with ITD, requested a batchload order early in June 2009. This included not only the matching and upload of the BHU records, but the symbol flip from BHD to TSC. Based on ITD’s record analysis, some cleanup of records was necessary before the extract. Additional mapping was required for lost, missing, replaced, or out for repair items. ITD completed their discussions with TLC for options/estimates on the data processing and extraction by the middle of July. Given a choice between cleaner data with a later date of completion and potentially inaccurate data with an earlier date, TU chose to delay the completion date by a few weeks. The data was ready for extraction by the end of August, one month after the library collection was scheduled to arrive at TU. TU staff selected BHU records that were representative of various scenarios to review when the programming was completed.

TU catalogers and the ITD project manager tested sample records from TLC and compared results. The first results were promising, as the records loaded properly, but several adjustments regarding holding and item information were required. To ensure proper mapping of data from TLC, the testers reviewed the sublibrary, collection code, item status, and material type that had been assigned in Aleph, and they analyzed the presence or absence of various local fields. One problem detected early in the process was that local notes were not being created as superholdings, and were being lost. Another issue was the inaccurate display of enumeration levels for serial and other multiple volume titles, which would later result in one of many clean-up projects. Because this problem was possibly due to inconsistency in the recording of the information rather than inaccurate display, clean-up was deferred until after the migration.

In early September, after fine-tuning the data, TU gave the final go-ahead to ITD to send the files to OCLC. The request was for a single institution batchload, with a file of the matching records returned to USMAI. The requested fields would be retained, non-matching records would be added to WorldCat as new records, and BHU holdings would be flipped to display as TU holdings (see figure 7).

The files were ready for uploading from OCLC in early October. ITD began loading small samples of records into the Aleph test module for TU staff to review for accuracy of content and OPAC display. The full load of the records into the test system occurred in late October, and the entire staff was encouraged to check access to the records and report any problems. On November 20, 2009, ITD completed the live load of the BHU records.

File cleanup began almost immediately. As previously noted, this was deferred until after migration because of possible inaccuracies in the serials item information. Another reason for delaying clean-up was pressure to complete the migration as quickly as possible. The serials records were of the most concern, as it was imperative to track arrived and expected issues. TU began receiving some periodicals, re-routed from BHU in May. This created a new layer of acquisition, and also impacted the accuracy of the catalog records. Another concern regarding the entire database was that some local holdings did not display properly, and the catalog contained duplicate records. By early December, the local holdings were adjusted, and duplicate records were resolved.

Many of the goals of the library merger have been met. To highlight some of the positive outcomes, one can point to the enhanced research capacity of the unified library institution, its attractiveness to potential student markets, and better facilities and exposure for the BHU collection. To share information with other libraries facing mergers, it is helpful to convey some of the lessons learned from the TU/BHU experience.

On the positive side, an important part of the process was the inclusion of all library departments in the early planning stages of the merger. Staff members from every area of the library attended an initial meeting to discuss all aspects of the merger. Follow-up meetings within some departments occurred regularly, especially in technical services, where the many phases of bibliographic work transpired.

From the very beginning, the head librarian demonstrated her commitment to the success of the merger and freely shared this optimism with the entire staff. She laid out the framework for the libraries integration as early as December 2008, six months before the merger of the universities was formalized, and her determination to complete the job on time motivated the staff to produce results. Another example of the head librarian’s positive approach was her assurance that the library would outsource as much work as possible to lessen the load on in-house staff. She began by gathering estimates from vendors for various stages of the project. These included the reclassification, editing holdings information, spine labels, and reshelving of TU journals; the removal and reassembly of shelving and end panels; and shifting the general collection. While only some of this work was outsourced, the effort to consider all possibilities was helpful to staff morale.

The good working relationship between the two libraries was an exceptionally helpful part of the merger. Although the remaining library staff at BHU was quite small, and most of the merger-related work fell to one person, regular communication and clear requests for information from TU made the process easier. BHU shared all pertinent documentation and facilitated interaction with TLC, and following the database migration, the staff member continued to be available for follow-up questions.

The fact that the merger took place during the summer months was an advantage. There were none of the common semester distractions, such as urgent requests from faculty to fill and process materials. Additionally, the fiscal year was just beginning, and there was no pressure to place orders for the yearly budget fulfillment. Timing is one factor to consider when planning a library merger.

Conversely, despite the fact that the merger was successful, there are things that would have made it smoother. From the beginning, there was not a systematic merger strategy, and preparations and implementation were driven by the need for speedy results. Although the TU staff met to discuss merger requirements, there was no step-by-step plan that would have provided structure for each department. Each department was left to determine the priority of necessary actions.

Although a plan for the database merger process existed, there was no similar plan for other parts of the merger. For example, periodicals management initially had no overall plan other than to remove duplicate subscriptions and to renew the remaining BHU titles. This lack of preparation for the complexity of managing the voluminous work was stressful; a more detailed plan of action would have benefitted the staff.

While the administration was very supportive of its staff, direction from a source experienced in mergers was lacking. Most of the integration was coordinated by technical services staff, none of whom had previously managed a project of this size or scope. It would have been more efficient to hire a consultant to help strategize the project from its inception, and a designated project manager would have been a welcome addition. This would have alleviated the stress placed on inexperienced staff, and it might have anticipated problems before they appeared. For example, it might have obviated the need for all library staff to participate in the physical move of books when the reshelving vendor miscalculated the placement of parts of the collection; a consultant familiar with space-related reshelving projects would have the knowledge to prevent this type of problem. Even though the vendor worked overtime to correct the issue, the library staff helped to meet the looming deadline for completion.

One of the lessons learned highlighted by Swanepoel is that “library mergers at institutions of higher education should not be underestimated in terms of complexity and the volume of work involved.”²¹ Although the staff were involved from the beginning, it was not acknowledged by the administration that the amount and intensity of some of the work may have been overwhelming. The short timeframe mandated by the merger affected technical services’ regular workflow for the entire technical staff, and routine work slowed or ceased. Management could have informed the staff to anticipate duties reassignment and daily workflow interruption. Hiring additional personnel to augment the technical services staff would have been a prudent decision, given the amount of work required.

Another area with room for improvement was the ongoing communication to all library staff as the project progressed. Communication seems to have worked best among the cataloging staff. They held frequent meetings and had a clear plan to follow for merging the databases. Acquisitions staff had less direction, and pertinent information from the administration was spotty and slow to arrive. Although intradepartment meetings in technical services occurred informally every day, staff members did not always have current information about possible services from outside vendors. More communication with the library archivist regarding special collection items would have been helpful; the same is true regarding the research and instruction librarians, relating to the impact of the incoming collection on public services. Despite the less-than-perfect conditions for the integration of the two libraries, there is much to appreciate about the results.²² The research capabilities of Cook Library are now expanded, public services for both collections are operating smoothly, the collections are fully merged, with the ILS accommodating all aspects of both collections, and the BHI collection is housed commodiously and attractively.