RDA: Analysing the Impact of a New Cataloguing Standard on SUNCAT, the UK's Serials Union Catalogue

SUNCAT is the union catalogue of serials in the UK, with just under 100 libraries contributing their serials data and associated holdings. These libraries use a variety of cataloguing standards already (e.g. AACR2, in-house), so the implementation of RDA will have consequences for display, indexing and matching in SUNCAT. To find out about the adoption of RDA in SUNCAT Contributing Libraries, a questionnaire was issued. The results indicated that, although there has been much thought around the topic, some libraries are waiting to see what others are doing before committing themselves to the expense of adapting a new cataloguing standard. Nearly all of SUNCAT Contributing Libraries are looking at moving to RDA, if they have not done so already, for new cataloguing. This means that SUNCAT will adapt processes to cater for RDA, but it is seen as another standard in an already hybrid catalogue.


INTRODUCTION
RDA (Resource, Description & Access) is starting to be used as the cataloguing rules of choice for many libraries, replacing AACR2 (Anglo-American Cataloguing Rules, second edition), and its implementation in library catalogues poses a series of challenges for those union catalogues which are built from the metadata held in local libraries.
SUNCAT (http://www.suncat.ac.uk/) is the Serials Union Catalogue for the UK research community, and contains the serials data and associated holdings information of 100 libraries, and the databases of CONSER (http:// www.loc.gov/acq/conser/), the ISSN Register (http://www.issn.org/) and the Directory of Open Access Journals (http://www.doaj.org/). The SUNCAT project started in 2003, becoming a fully-fledged service in 2006. SUNCAT is funded by Jisc (http://www.jisc.ac.uk/) and is one of the services offered by EDINA (http://edina.ac.uk/) based at the University of Edinburgh. SUNCAT is a physical union catalogue with a deduplicated view, made up of the bibliographic and associated holdings records for the serials held by the 100 SUNCAT Contributing Libraries (with new libraries being added on a regular basis). SUNCAT Contributing Libraries are major research libraries, and as such are a mix of higher education, specialist and national libraries: these include libraries such as the British Library, the National Library of Scotland, Cambridge University Library, Oxford University libraries, through to specialist libraries such as the Wellcome Library, the Natural History Museum and the British Film Institute. There is also one non-UK library in SUNCAT, which is the library of Trinity College, Dublin. CONSER records are searchable through the SUNCAT search interface, but the ISSN records are not, due to the licensing agreement with the ISSN Centre. However, SUNCAT Contributing Libraries are able to download ISSN and CONSER records, along with records from other Contributing Libraries, to upgrade their own catalogues with high quality bibliographic data. SUNCAT also provides a service for the UK Research Reserve (UKRR (http://www.ukrr.ac.uk/)), a system allowing the retention of low-use print journals in the UK Higher Education environment. The service provides a report on the holdings for particular titles of all UKRR members.
As a physical union catalogue, bibliographic records from the Contributing Libraries are sent to SUNCAT, and run through a conversion process to normalize the data. Records for the same title are matched using an algorithm, with the best bibliographic record showing with all the holdings records for all the associated libraries in that set. SUNCAT uses the Aleph 500 library management system, supplied by Ex Libris, as its back-end, with a new, in-house designed discovery layer. SUNCAT is continually updated, with changes reflecting new, altered and deleted records, which may be part of the library's OPAC or other discovery layer. At present, there are over seven million records in the SUNCAT database, catalogued in a variety of standards.
Many of the SUNCAT Contributing Libraries already use AACR2, and are migrating to RDA. Some, however, use other cataloguing rules. This means that SUNCAT is already a hybrid catalogue in a MARC21 format, and the implications of adding another cataloguing standard needed to be considered. To find out what was happening in libraries, it was decided to issue a questionnaire for the SUNCAT Contributing Libraries, which would cover which libraries were moving to RDA, what training was being undertaken by cataloguers, etc. The results were analysed, with the idea of seeing what the Contributing Libraries were doing, and how SUNCAT might need to change indexing, the matching process, and the display fields.

Methodology
The SUNCAT questionnaire contained eighteen questions, and was issued using the Bristol Online Surveys tool. It was vetted by three librarians from Contributing Libraries before being issued. The questionnaire was then posted to the SUNCAT Contributing Libraries mailing list on 4 February 2013, and closed after a period of three weeks. The SUNCAT RDA survey questionnaire had thirty-one replies, a 34% response rate. Of these 31, sev-enteen were Higher Education institutions (54.8%), six specialist libraries (19.4%), one public library (3.2%), one national library (3.2%) and six who did not leave their institution name (19.4%), only one of which filled in the rest of the questionnaire (see Figure 1).

Q1. Will your library be adopting RDA?
Twenty-two of the respondents (71%) said that they were going to adopt RDA, one said that they would not, and four were undecided ( Figure 2). The next part of this question was to ask why the Contributing Library would or would not be adopting RDA. From the answers given, it was clear that record-sharing between libraries and other bibliographic databases (e.g. OCLC, RLUK) was important as a time-saving device. Therefore, if the national libraries across the world, such as the British Library and the Library of Congress, were to adopt RDA, the majority of other libraries would follow suit. This would make record-sharing easier, as less work would have to be done to RDA-compliant records. As was succinctly put in one reply: 'The major suppliers of MARC records will be adopting [RDA]: OCLC, LC, BL. We source approx. 98% of our records externally, mainly via these organizations and feel the decision has already been made for us.' There is a feeling that was noted on more than one response that adopting RDA was an unavoidable option. However, some responses were very positive, saying that they would adopt RDA as a 'natural progression' and to 'make use of FRBR principles once RDA has overcome the obstacle of the MARC21 format'. RDA was seen by some as a means of offering good quality records, which would increase the findability of items.
The view from the specialist libraries was similar -one library said that it would be 'unrealistic' not to adopt RDA. Several libraries indicated that they were waiting until there are a larger number of RDA records in sharing databases, or when the majority of record suppliers send bibliographic records in RDA. One library was waiting until after the implementation of their new LMS before making a decision on adopting RDA. Only one SUNCAT Contributing Library said that they would not be moving to RDA at present, as they were still using UKMARC.
These responses were not unexpected: all bar two respondents download their bibliographic records from outside sources, and it would be inefficient in the extreme to convert records following one standard to another. (The two Contributing Libraries which do not download bibliographic records do all their cataloguing in-house.) This is especially true of batch downloads of records for electronic journals, which may number in the thousands every month. It was also no surprise that some libraries would adopt a 'wait and see' policy, and revisit the use of RDA when there was a larger corpus of records in that standard.
Q2. When will your library be adopting RDA? Two libraries were already using RDA, either for all new records, or for locally-derived records, or for copy-cataloguing. Four libraries were looking to start using RDA in either three months' time; three libraries, in six months' time; seven were looking to adopt RDA some time in 2014, with one looking beyond that. Eight libraries were not sure when they planned to move to RDA, one saying that the only cataloguer had not had training in RDA, and that they did not have access to the RDA Toolkit, but that they were still intending to move to RDA. The one library that said they would not adopt RDA would do so at some time in the future ( Figure 3).
Q3-4. Will you still be accepting AACR2 records from external cataloguing sources after you have adopted RDA? Conversely, if you are not adopting RDA as your cataloguing standard, will you be accepting RDA records from external cataloguing sources?
The next two questions were related, regarding the acceptance of records of an opposing standard into the library catalogue, and if the library would change the records from one standard to another, if necessary (see Figure 4). It was expected that the answers given would explore the depth of the hybrid catalogue, and its prevalence amongst the Contributing Libraries. For those libraries adopting RDA, any records that were in AACR2 would normally be accepted as is, maybe looking to change them to RDA at some point in the future. One library said that it would be 'unrealistic' to consider adapting non-RDA records, although they would still accept them. From a purely pragmatic point of view, the hybrid catalogue would be a necessity, rather than a desirable option. AACR2 records would be accepted only if RDA records were not available; again, this seems to be a pragmatic approach, based on it being easier to download an AACR2 record than create an RDA record from scratch. Thirteen libraries said that they would leave the records in the format in which they were downloaded, with the caveat that AACR2 records would only be chosen if there was not a viable RDA record available. One smaller library, which does not download vast numbers of records, said that they would change any AACR2 records to RDA. In contrast, seven of the larger libraries also said that they would change any AACR2 records to RDA.
One library said that it would continue to catalogue certain formats in AACR2, such as serials, and RDA for other formats. One library was thinking of creating hybrid records, theoretically keeping records as AACR2, but adding RDA fields through bulk editing -but only if there was a definite benefit to users. Another would leave AACR2 records from trusted sources, but convert others to RDA. Overall, most respondents indicated that they would prefer to download RDA records, but would download AACR2 records where no other alternatives were available, and that they would retain those records in the standard in which they were catalogued. There were a few libraries who said that they would change the record to the standard which they used (e.g. to future-proof their catalogue), but most would stay the same as there were not the resources available within the cataloguing department to change every downloaded record.

Q5. How do you see RDA impacting on your cataloguing workflow?
The answers regarding the impact of RDA on workflows were realistic, but encouraging -it was presumed that there would be a reduction in produc-

Figure 4: Changing the standard of downloaded records
tivity whilst cataloguers learnt the new rules, but this would pick up over a relatively short period of time. Most libraries said that workflows would not change, per se. However, it was not clear if the extra typing involved (e.g. spelling out of abbreviations, additional fields) would slow down cataloguing output in the long run. It was presumed that templates and macros would help with this, so that there would not be too much difference in productivity between the two standards. It was noted that the largest impact would be the training of cataloguing staff, especially for original cataloguing. Copy cataloguing was seen as less of an issue, as it could be covered through detailed documentation. Some libraries hoped that there would be no impact on their cataloguing workflow, as the cataloguers would be proficient in both RDA and AACR2. It was noted by one library that BIBFRAME (http://bibframe.org/), the replacement for MARC21, would be more disruptive than changing to RDA. It was clear that this was not a simple question to answer, as it was relying on future forecasting, rather than what had been established.

Q6. How are you training your cataloguers to use RDA?
The answers supplied indicated that a mixture of training, ranging from internal documentation, external seminars and workshops would be adopted. Mention was made of freely-available resources, such as documentation from the Library of Congress (2012a, 2012b) and the CILIP CIG RDA forum held in October 2012 (Cilip Cataloguing and Indexing Group, 2012).
It was also noted that training would be on-going, as the RDA Toolkit would be a continuing resource and updated at regular intervals. Some libraries had not yet started training in RDA, and were planning starting off with some presentations to staff on the general principles (e.g. the FRBR model). There was a comment that any external workshops offered to date were not within a reasonable geographic distance, making it unfeasible to send a group of cataloguers to attend. However, this is being addressed in Scotland, by, for example, the CIGS/BDS workshop 'An Introduction to RDA' (Cunnea, 2013). One library wanted to hire an external trainer to deliver training in-house. This would save the library time in developing its own training materials, but would either have to be general in nature, in which case there would need to be further in-house training to accommodate local practices, or the trainer would have to tailor the training prior to delivery.
However, it was clear that most libraries, even if they had not started training their cataloguers in RDA, had been giving the process some thought. Training would probably be collaborative, with attendees of workshops cascading the experience gained to their colleagues in order to save time and resources.
As an aside, it was interesting to note that none of the respondents raised the issues associated with using the RDA Toolkit which have been raised in various forums, such as the clarity of the text, the financial costs, and so forth, which might have had an impact on training, although one library mentioned that 'getting cataloguers familiar with the RDA Toolkit' was a challenge.
Q7-8. How is your serials cataloguing in RDA going to differ from that in RDA? If there are changes to your serials cataloguing in RDA, what are these? It was felt that there would be changes to serials cataloguing conducted by the Contributing Libraries, but it was not clear for many libraries what those would be. Some felt that it would depend on what their records vendors would supply, so would change accordingly. It was noted that changes that applied to serials records would apply to other formats also. Some respondents were fairly detailed in noting changes, talking about use of the 33X tags, no abbreviations, and giving fuller information in tags such as the 300, use of the 264, etc.
One library did note that there would be no changes; they would carry on cataloguing serials in AACR2, as they create serials records in-house. Another library was most enthusiastic about the opportunities that RDA would offer in being able to clean up serials records. They mentioned that the use of the 336/7/8 tags would help distinguish between different formats for the same title. Looking forward, this library wanted to 'take advantage of the capabilities of RDA, when it is not constrained by MARC21', concentrating on the links between the works, expressions, manifestations and items.
However, it appears as if there is much uncertainty at present, in regards to the changes that will occur in cataloguing workforms. On the plus side, it is also felt that there will be fewer changes to serials records than other formats when cataloguing in RDA, or that vendors such as Serials Solutions would be cataloguing electronic records, so that the impact for serials cataloguing will be minimal.
Q9-10. Have you experienced any issues with serials cataloguing in RDA which are different to those in AACR2? As many libraries were still contemplating how RDA would fit into the cataloguing process, this subject was seen as hypothetical, with very few answers based on practical experience. One library thought that the differences between CONSER and RDA rules would be an issue -presumably, in the same vein, the Library of Congress Policy Statements (formerly, Rule Interpretations) would also pose a challenge until local practice was established. One respondent thought that ISSNs for electronic material would be a problem, wondering if all e-journals would be assigned unique ISSNs. This is less of an RDA issue, and more of a query to take up with the ISSN International Centre.
There were comments that monograph RDA cataloguing is being prioritized over serials cataloguing, so that many libraries have not yet implemented their serials RDA policies. This meant that the respondents felt that they were unable to give a reply based on practical experience, although there were comments regarding the amount of time that the new records would take to complete. This would be mitigated by use of macros for commonly used tags containing similar information, and use of other cataloguing tools, such as standard templates. Fourteen respondents simply said that they had not experienced any issues to date, with the presumption made that they were not looking for them.

Q11. Will you be retrospectively converting any existing records to RDA?
The SUNCAT Contributing Libraries were asked if they would be retrospectively converting existing records to RDA, to provide uniformity across the catalogue. This was seen as a desirable outcome, but, from a practical point of view, one which was unlikely to occur. Some libraries were hopeful that it might happen as a side-effect of any other re-cataloguing projects, as records would then be catalogued using RDA. Other libraries simply said that there were not enough staff resources or time to undertake any retrospective cataloguing. On the other hand, one library in particular said that they would be undertaking retrospective cataloguing, including making changes to the General Material Designation (GMD), and, in particular, to authority records. Other libraries are adopting a hybrid approach; one said that they would possibly retrospectively convert records to RDA where they had a record for a print and electronic title, one record in AACR2 and one in RDA -in this case, the AACR2 record would be altered to become an RDA record. They also stated that items that only 'currently exist in [the] card catalogue may be retrospectively converted to RDA standards'. Another said that, although there would not be a policy for retrospectively converting records to RDA, any global changes applied to the catalogue would result in hybrid records. Along the same lines, yet another said that they would be making global edits, such as spelling out some standard abbreviations, either using MARCEdit or their own Library Management System, to make their records more consistent. Another library is taking a pragmatic approach, and tackling retrospective conversion on a case-by-case basis, 'as and when [they come] across records that could benefit from conversion'. This library also said that, if there were enough staff resources at any time, this approach would be reviewed in favour of a more systematic, comprehensive method, with the overall aim of changing all their records to be RDA-compliant. Another library is also taking a piecemeal approach to altering existing records to RDA. For serials in particular, this would apply to former titles when cataloguing title changes. However, it was made clear that this decision is only in the planning stages at present, and not yet an established policy. Another respondent said that they are considering a similar approach of ad hoc revisions, as and when is deemed necessary. They felt that this would apply to serials in particular, due to title changes and acquisitions of new formats for existing titles.

Q12. Will your vendors of bibliographic records be supplying RDA records?
Sixteen respondents said that their record suppliers would be sending RDA records, but there was general concern across the board that they were not sure when this would be happening. Some libraries suggested that they might already be receiving RDA records from suppliers, but were not sure as the records were batch-loaded in bulk into their Library Management System or other discovery layer. There was also a concern that these batch loads would consist of both AACR2 and RDA records, rather than all records supplied following a single standard. The majority of libraries were in the process of exploring which of their vendors would be supplying RDA records. There were three respondents who said that they did not have any vendor-supplied records at all, so this was not an issue for them. Vendors (or other bibliographic record suppliers) who have indicated that they would be supplying RDA records include: RLUK, OCLC, Library of Congress, Serials Solutions, Talis, Ebsco, BDS, MyiLibrary, Dawsonera and Coutts.

Q13. Will you be changing your OPAC display to account for the new MARC21 codes found in RDA records?
This question was concerned with finding out if the new MARC21 codes, such as the 264 (Production, Publication, Distribution, Manufacture and Copyright Notice), the 336 (Content Type), the 337 (Media Type) and the 338 (Carrier Type), would be used and displayed. Overall, it was generally agreed that the OPAC display for each Contributing Library would have to change, but this was seen as a task for the future, and had not necessarily been considered at this point in time. It has become clear from subsequent communications with the SUNCAT Contributing Libraries that not all Library Management Systems are RDA-compliant, which means that not all libraries would be able to make the appropriate changes even if they wanted to. One supplier has changed this with their latest upgrade (Ex Libris, with their Voyager LMS). The two libraries currently using RDA records have already changed their display, but this was a minority group amongst the questionnaire respondents. Some libraries were looking at their OPAC displays imminently, and others were delaying until more information as to what other libraries were doing became available. For display purposes, the majority view was such that the 264 tag would be shown, but that the 336-8 tags would not, as these were designed to be machine-readable only. Comments were made that there would have to be in-depth communication between departments within the institutions (such as cataloguing and systems), which would indicate that progress might not be as speedy as desired. On the other hand, some libraries hoped that they would be able to display RDA records relatively soon, one hoping that their improved OPAC display would happen in the summer of 2013. Interestingly, more than one library mentioned the creation of a 245$h (GMD) for display based on the 336-8 tags, but this will be explored further below. One library said that they now had enhanced icons to create better display information, such as the copyright symbol ©. One particular library said that they were considering trialling the display of the 336-8 tags, to see if they provided any 'real value' for users before deciding if these tags should be retained or suppressed in display. This appears to be a unique response, but definitely one worth considering; if the tags are shown to be useful, they should be displayed. However, the general consensus is that the 336-8 tags would be of more use in the future, as library systems and discovery layers develop (see Figure 5).

Q14. Do you currently make use of the General Material Designation found in AACR2 records (MARC21 245$h)? If so, will you discontinue using the GMD on implementing RDA?
The next subject was a very specific one, regarding the current use of the General Material Designation (GMD) found in AACR2 records, and in the 245$h MARC21 subfield. This is not used in RDA, as the GMD has become conflated with the Specific Material Designation (SMD), and there appears to be a feeling in the cataloguing community that the descriptor terms as defined in AACR2 (Rule 1.1C) are not appropriate for the web environment. This is why the Content, Media and Carrier Types (MARC21 tags 336-8) were introduced in RDA -to allow for much greater flexibility in the description of the resource type. However, as many libraries are not planning on displaying the 336-8 at present, the information that was provided in the 245$h will not be as easily accessible. There is also much debate regarding the positioning of these tags. At present, the GMD is part of the title, and thus seen immediately after the title proper in many displays. It acts as an early-warning sign as to the nature of the resource about to be consulted. If the 336-8 fields are displayed, maybe as icons based on the content of these fields, then they appear in the descriptive part of the record, before the title. It has been voiced on different cataloguing fora that the RDA record has lost that early-warning sign offered by the 245$h. Therefore, the SUNCAT team were interested in seeing what their Contributing Libraries thought about this issue. For some libraries, this was a non-starter, as they did not However, other libraries, although they had either adopted or were planning to adopt RDA, still retained the GMD and had no intention of not using it. This would create a hybrid record, coded as RDA but retaining some AACR2 elements. Other libraries were considering creating a GMD for display purposes from the data in the 336-8 tags, especially for non-book material. This shows that that GMD supplied a real use for the users of library catalogues. Overall, it was most interesting to note that, out of the fifteen respondents who answered the question, 'Will you discontinue using the GMD on implementing RDA?', seven respondents said that they would not, thus creating hybrid records. Eight of the respondents (so, slightly over half) said that they would discontinue GMD use, with two exploring other ways of displaying GMD-equivalent information. It was not clear from the respondents' answers if the retention of the GMD was due to the presence of AACR2 record in their catalogue, or, more likely, if it was felt that the GMD in the 245$h offered value to end-users which was not replaced by an appropriate rule in RDA (see Figure 6).

Q15. Will you be changing your indexing routines to cater for RDA records?
Indexing routines are linked to display matters, but also include issues with searching, as the indexed information in the new MARC21 tags can be used in filters, limits or straightforward searches. It was interesting to note that this was as much a divisive question as to that pertaining to the GMD. From the

Figure 6: Libraries who said they would continue use of the GMD in RDA records
sixteen respondents who answered this question, seven said that they would not be changing their indexing routines, and nine said that they would. One library said that they did not think that any changes were necessary. Another said that they would not change the indexes at present, but would review the situation once authority records had been fully migrated to RDA. Another library said that it would not index the $e in the 100 tag (Relator term the Personal Name Heading) in authority indexes. This clashes with several libraries who said that they would index this subfield, as it was one of the new MARC21 codes. The main reason given for changing indexing routines was to account for the various changes in MARC21 fields, and to harmonize searching between records coded in AACR2 and RDA. One library made mention of indexing the 264$b (Name of producer, publisher, distributor, manufacturer) so that searching on publisher would bring back results from both the 260 and 264 tags. Another library made mention of indexing the 040$e (Description conventions in the Cataloguing Source tag), presumably to search records coded according to different standards (Figure 7).

Q16. Will you be reconfiguring your cataloguing systems to support RDA?
The last question asked if libraries would be reconfiguring their cataloguing systems to support RDA. This could be done in a variety of ways, such as validating RDA-related MARC coding, adapting cataloguing template and macros, or using a version of MARC Report (http://www.marcofquality. com/soft/softindex.html) configured to check RDA records. Generally, the consensus was that cataloguing systems would, indeed, need to be reconfigured, but that this needed further investigation. From the seventeen respondents to this question, all of them said that they would be reconfiguring their cataloguing systems. From those who had explored this area further, adapting cataloguing templates to cater for RDA was by far and above the top priority. One library was considering the use of macros to fill in the 336-8 tags, to help update records from AACR2 to RDA. Two others made mention of the validation of the new MARC tags, so that they could be loaded into the Library Management System. One library wondered whether some of the tasks mentioned would be completed by the systems vendor, or if such tasks as templates and checking routines would have to be done on a local basis. This would result in the duplication of work across different libraries, but, on the other hand, would mean that the library had a personalized template, for example. Overall, the need for changes to cataloguing systems was recognized, but it was clear that much further investigation was necessary.
The respondents were offered the opportunity to add any comments not covered in the questionnaire. Overall, the opinion was that many libraries were early on in the learning curve for RDA, and had a hope that any changes would not be too vast or disruptive. There was much interest in the outcome of this survey, essentially, to learn what other libraries were doing, which might have an impact on the policies currently undertaken by the reviewing library. Comments were also made that the questions posed were difficult to answer, as some libraries were still investigating RDA, and were in a decision-making phase. A need for training was mentioned, indicating a need for clear documentation, and a lack of self-confidence in establishing own training materials for RDA record creation. It was also noted that plans change and evolve, much in the same way that RDA is currently changing and evolving. It was clear that the onset of RDA has resulted in a position of flux, which was unsettling to those used to the comfort of a well-established standard such as AACR2. One library commented that the questionnaire had proved useful for future planning, as it showed the issues that the library would need to consider. Many libraries felt that RDA was still too new to make any comments, and that its impact could not be measured at the moment. One comment summed up a general feeling regarding RDA, and is reproduced in full: 'The benefits of RDA can only be fully appreciated when the system-vendors rise to the challenge of FRBR. Until then it is hard to justify the time, effort and expense of implementing it. However, we feel it is necessary to do [so] in order to be part of, and contribute to, the international cataloguing community.'

SECTION 2 -SUNCAT AND RDA
The questionnaire for the SUNCAT Contributing Libraries has been most useful in aiding the changes that SUNCAT will make to accommodate RDA.
As has been mentioned previously, SUNCAT is currently a hybrid catalogue, accepting serials records from Contributing Libraries in a variety of standards and formats; RDA is another standard being used in an already hybrid catalogue. Taking one library's serials data may show records following RDA, AACR2, AACR and other legacy standards within one file. Multiply this by 100 (the number of Contributing Libraries in SUNCAT at present), and the hybrid nature of SUNCAT becomes clear. SUNCAT shall adapt its presentation layer to accommodate RDA records alongside records in other standards in a MARC structure. Behind the scenes, indexing routines will be adapted, so that the new data may be searched comprehensively. The SUNCAT matching algorithm will also have to be changed (this matches individual records with the same title into one set, so that there should be one record per title), but that is more of a long-term goal at present. The data specifications, tailored for each library to normalize the data so that it will load into SUNCAT, will also be altered where appropriate.

Display
The search screens of the discovery layer are the most familiar part of SUNCAT. There will be some changes to accommodate RDA, but it is hoped that they will not be obvious. Essentially, the display will be modified to show the new MARC codes, but more work will be done behind the discovery interface. It is most likely that the 264 tags will be displayed along with the 260 tags, under the umbrella heading of 'Publication details'. As can be seen from the results of the SUNCAT RDA questionnaire, it is not expected that the 336-8 tags will display in the full record, but will be visible in the MARC record display. The full record display shows salient bibliographic information, but not the contents of every single MARC tag -those are visible in the MARC record display. The full record display makes for a cleaner, more relevant return for the majority of SUNCAT end-users; the MARC record display offers a more detailed view, including tags such as the Leader and other fixed fields, which are of more use to librarians and those familiar with MARC21. SUNCAT will not try to extrapolate information from the 336-8 tags for displaying a form of general material designator. It is expected that the 100$e / 110$e / 700$e / 710$e (etc.) General Relationship Designators will display, but more work needs to be done on this, exploring its value in a display environment.

Indexing
The 264 tag, and the 100$e / 110$e / 111$j /700$e / 710$e / 711$j (relator terms) subfields will be added to the current indexing routines, but it is unlikely that the 336-8 tags will be included. This means that the 264 and the 100$e will searchable. This will be useful for publishing information and author searches. As a serials catalogue, SUNCAT tends to have more titles as the preferred entry; if the preferred entry is a name, it is more often a corporate body than a personal name. However, there are many corporate bodies and personal names in the catalogue, so having the 100$e and the 110$e indexed will offer useful search results where these occur.

Data conversion routine changes
The first step in adding the initial data from a Contributing Library to SUNCAT is to inspect the incoming records, and write a data specification. This will change some of the data so that it is in a form suitable for loading. SUNCAT uses the Aleph Library Management System as supplied by Ex Libris, but is in the process of adding an in-house discovery layer which is tailored to the serials search environment. The data specifications have to work in both Aleph and Solr, the database behind the new interface, adding an extra layer of complexity to a procedure which is not the most straightforward to begin with. The general principle for writing a SUNCAT data specification is to try to leave the data as intact as possible. Most of the data manipulation occurs with holdings, to ensure that they display in a uniform manner. Therefore, the data specifications for new libraries will be written to take account of RDA, but changes to non-RDA records to make them RDAcompliant will not occur. Essentially, the data specification will remain independent of cataloguing standards, as there is already a mixture of standards within SUNCAT. The one exception to this will be for those libraries who do not supply data in a MARC21 format. In these cases, SUNCAT creates a cross-walk from the library-supplied data to the appropriate MARC21 fields. With any new library files that meet this criterion, the SUNCAT team will write the data specification according to RDA. Existing data specifications will not change.
One particular topic of interest that was raised from the answers in the SUNCAT RDA questionnaire was that of the GMD in the 245$h. SUNCAT will not change records which have been coded as RDA in the 040$e, but still retain a 245$h. SUNCAT is a union catalogue with a deduplicated view of the records contained therein. This deduplicated view shows a single, 'preferred' record for one title, with all associated holdings displaying, rather than showing every individual record that has been contributed. In order to achieve this, SUNCAT uses an algorithm to match records with the same title into one set. This matching occurs above format, so there may be records for print, electronic and other formats in one set. Therefore, SUNCAT does not display the GMD in the full record view, as it may well be misleading to display a record as an electronic resource, for example, when there are holdings for print items in the set. If a GMD has been added in the record, it will show in the MARC record view. The loss of the GMD in the title display in RDA records will therefore not affect the display in SUNCAT.

Matching
The SUNCAT matching algorithm will have to be changed at some point, to cater for the inclusion of the 264 tag (Production, Publication, Distribution, Manufacture, and Copyright Notice). The algorithm is based on a points basis -points are awarded for matching the same field with the same content; partial contents are awarded fewer points. It is presumed that the matching would be similar to that of the 260 tag, that is, on the name of the publisher in the $b. There would be an added level of complexity, with cross matching occurring (but with fewer points) between the 260 (Publication, Distribution, etc. (Imprint)) and the 264, so that records catalogued in AACR2 and RDA for the same title could match.
As part of the regular processing that the library data undergoes before being loaded into SUNCAT, counts are made of important tags, e.g. to see if every record has a title. A new count has recently been added, to count the number of incoming records coded as RDA, as noted in the 040$e. At present, few records are coded thus, but it is expected that that number will rise sharply as more libraries adopt RDA (see Figure 8).

Figure 8: Summary of new MARC21 fields / subfields and their treatment in SUNCAT
will adapt processes to cater for RDA, but it is seen as another standard in an already hybrid catalogue. Individual hybrid records will be treated in the same way as other records. (It was expected that the hybrid records would be AACR2 records with RDA elements, but answers to the questionnaire have indicated that there will also be RDA records with AACR2 elements). It is clear that the SUNCAT RDA questionnaire has asked some pertinent questions; it is equally clear that policies have not yet been established, especially with regards to serials cataloguing. With this is mind, it is intended to repeat the exercise of supplying another questionnaire to SUNCAT Contributing Libraries at some point in 2014 regarding their use of RDA. Then a 'compare and contrast' exercise can be undertaken with the aim of seeing how the new standard has been adopted.
Overall, the SUNCAT RDA questionnaire has proved useful, not only to the SUNCAT team but also to the libraries who responded. The answers supplied have modified the way in which RDA-compliant data will be treated in the SUNCAT database, and some of the respondents have noted that the questions asked have helped them formulate policy. It is clear that the adoption of RDA is in its infancy, but will continue to grow and mature. SUNCAT will reflect this, as it is expected that the majority of Contributing Libraries will adopt RDA over the next few years.
The time and effort spent by the SUNCAT Contributing Libraries on answering the SUNCAT RDA questionnaire is much appreciated.
Natasha Aburrow-Jones, a graduate of Royal Holloway, University of London, is the SUNCAT Bibliographic Project Officer, with the role of overseeing the incoming data to SUNCAT, the Serials Union Catalogue for research libraries in the UK. She is also responsible for data quality in SUNCAT, and solving any issues that may arise due to discrepancies in the data. Her interests lie in following cataloguing standards. She has been working on SUNCAT since 2003. Prior to working in EDINA, part of the University of Edinburgh, where SUNCAT is based, Natasha spent ten years in different Oxford academic libraries, such as the Radcliffe Science Library, the Ashmolean Library and the Bodleian Library, in a variety of cataloguing and bibliographic maintenance roles, specializing in serials cataloguing and retrospective conversion.