NARDAC RDA Update Forum 8.3.20

Editor's Notes

• #21: Record value for one or more of the expressions aggregated. Record value that is common to all expressions aggregated.
• #27:  RDA/PS conversion project update: the LC/PCC policy statements
• #28: PTCP is not only changing LC’s documentation but also the documentation that is used by the members institutions of the Program for Cooperative Cataloging (PCC).
• #29: PTCP first task is to review existing policy statements and create new ones. ALA Publishing asked LC to undertake this task. The Beta Toolkit will be launched on December 15, 2020 but no date for its implementation has been fixed. LC will be consulting with PCC to decide on an implementation date. Although LC is charged with this task by ALA, it is a joint LC/PCC project and LC agreed to keep PCC informed throughout the process. LC provides PCC PoCo with monthly reports and separately with issues that LC and PCC need to discuss.
• #31: The changes in these area are so significant that LC and PCC agreed to set up 4 task groups with co-chairs, one from LC and one from PCC, to examine the changes and make recommendations.
• #32: PTCP is leading the effort but could not do it without assistance from LC catalogers and PCC.
• #33: We hope to test it before the implementation dating drawing on PCC and LC catalogers for the test. Why a test? To make sure the work of the Policy Statement and metadata documentation are correctly aligned when creating metadata for an item.
• #34: One of the first things we realized once we really grappled with the project is that there were actually six phases to it, named here. Each phase is a fairly large project all bit itself!
• #35: The first phase is what we are inside now. We are close to finishing the 1a stage of Phase 1, which is foundational to all the rest. In this stage, we are carefully analyzing the current policy statements, looking at the reorganized beta Toolkit to figure out where stuff now needs to go – which is sometimes in more than one place – and reword the instructions to match the current RDA content. Because this is so monumental, we have set up a process where one cataloger writes something, another cataloger proofreads it and makes comments, and the returns it to the original writer for modification if needed. We have had plenty of interesting discussions as we navigate this change. The next stage of Phase 1, which we are soon to start, we are going to be analyzing the full text of RDA options in order to write new policy statements that didn’t need to exist before, and bring the 1a re-written statements into cohesion with the new text, so that it all works together.
• #36: Phase 2 is the technical part, in which we take the text we’ve created and then input it into the content management system in the DITA XML standard for the text. The technical group is currently studying DITA together and trying to work out the best practices for how we are going to use it in cooperation with the existing guidelines for the RDA content. Phase 3 is for the application profile, which overlaps with the policy statements. We worked out a format for the AP, and how we plan to approach it, but have not yet been able to focus on it. Phases 4, 5, and 6 have not yet started, because all of them are depending on the previous three phases.
• #37: As we have gotten our teeth really into the project, we have found there are some real challenges in getting this work done. One of these is the sheer complexity of the work, in each subproject. Phase 1 is really taking us a while, but I would also note that it is foundation to all the rest. Also, the project has grown in unexpected ways. And of course, the impact of the pandemic is real. In truth, I have been incredibly pleased with how well our group has continued to produce once we adjusted to the challenges of all telework. I am truly grateful for all the hard work that my colleagues in PTCP and in ABA have persisted in their hard work on this project despite the many distractions and challenges we currently face. I am now handing this off to Clara Liao, who came in as a section head to our division in April, and has joined in the work on the project.
• #40: First, a reminder of the context in which we work: Library and Archives Canada is a bilingual national institution, with French and English as the official languages. Our bibliographic description work must therefore reflect that mandate. We create bibliographic and authority records in French for publications in French, English records for publications in English, and records in both languages for bilingual publications.
• #41: Over that past few years, LAC has seen some major changes in the way we work. LAC recently moved its library system from in-house Amicus to OCLC's WorldShare Management Services (WMS). This change provided the opportunity for LAC to transition to the Name Authority Cooperative Program (NACO) to manage its English language authority records. A new module, known as "Canadiana Name Authorities in French", has been developed within WMS to house the French-language authorities.
• #42: The Francophone Name Authority Programme, also known as PFAN (Programme francophone des autorités de noms in French), is a new collaborative programme in which some libraries contribute directly and collectively to the Canadiana file. The programme is the result of collaborative efforts between LAC, Bibliothèque et Archives nationales du Québec, and 15 French-speaking university libraries in Québec. The programme uses the NACO model as the basis for its gouvernance structure, its programme, and its policies and was officially launched in June. Its standards committee is currently working to develop a set of documents to support the work of the PFAN members.
• #43: To recap, since LAC’s transition to OCLC’s WorldShare Management Services, cataloguers are working with two files for authorities (LC/NACO and Canadiana/PFAN) as well as three files for subject headings (Library of Congress Subject Headings, Canadian Subject Headings, Répertoire de vedettes-matières). This large inventory of files will have an impact on the scope of instructions that LAC will have to establish.
• #44: Right now, LAC’s policy statements in the current RDA Toolkit need to be revised. They are available in the Toolkit in both French and English. However, they make reference to the old library system, Amicus, and do not reflect the current environment. It is expected that the current structure will be more or less conserved: LAC practice and BAnQ practice.
• #45: For the policy statements for the new RDA Toolkit, we must follow a few principles. First, the policy statements must be available in both French and English. They must be aligned with LC, PCC, and the PFAN, all while being specific to Canada and LAC where needed. Finally, they must take into account the new work environment that is OCLC WMS as well as our collaboration with BAnQ.
• #46: This is a preliminary list of actions that must be undertaken to develop LAC policy statements in the new RDA Toolkit. Note that this list is not exhaustive.
• #47: A side note on application profiles: LAC has started work to analyze the work environment and to get a vision of the cataloguing tools that will be used by description staff. Furthermore, Elizabeth Sander, acting manager of Description, is a member of the Application Profiles Working Group.
• #48: In conclusion, given the other priorities that we have for 2020-2021 (such as PFAN), the schedule of the French translation of the RDA Toolkit, and LAC’s need to align with policy statements to be produced by LC/PCC, Library and Archives Canada will not be implementing the new RDA Toolkit on December 15. LAC will determine its implementation schedule in the fall and will communicate this information.
• #49: Thank you. Are there any questions?
• #50: In this presentation, I will give an overview of how the 3R Project helped make the translation process more efficient. I will also address some of the issues that we encountered when working on the French translation of the new Toolkit.
• #51: When we began working on the French translation of RDA back in 2010, we started with simple tools, that is, Word files for the chapters, the appendices and the glossary. We were also provided with a Word file of recurring phrases that we translated at the beginning of the process, to ensure consistency between translators. This meant that there was originally a lot of redundant content between instructions and between the definitions and scope notes of elements in the instructions and in the glossary. The process required sustained attention from translators and careful review of their work to achieve consistency.
• #52: Several steps have been taken over the years to automate the production of RDA Toolkit's content, including the translations. One of the objectives of the 3R Project was to implement a process to synchronize RDA translations with RDA Reference data in the RDA Registry so that labels, definitions and scope notes for elements and vocabularies, which appear in instructions, the Glossary and the appendices, could be automatically pulled from the Registry. The synchronization process was first used in the English version in 2016 and was implemented for translations in 2017 in the current Toolkit. For translation teams, this involved providing translations of RDA Reference using templates in CSV format that were later uploaded to the Registry. The slide shows an example of template, in this case the one for the Extension plan vocabulary.
• #53: The use of Registry data was further expanded in the new Toolkit. It is currently used to populate parts of entity and element pages, the glossary and vocabulary encoding schemes. The same will happen with the translations. For translation teams, this involves updating their translation of RDA Reference since the English version was extensively revised and expanded, in part to support the LRM. This is a significant project that a few teams have completed but that others, including the French team, are still working on.
• #54: Registry data is also used in parts of examples. The slide shows a view in context example where the element names in the left column and the term “spoken word” in the right column were pulled from the Registry. This means that translators will not have to translate these terms again when working on the examples. It is also interesting to note that examples are now grouped in separate files, which should give translators more flexibility if they want to replace or delete existing ones, or add new ones.
• #55: During the 3R Project, translators also began working with SDL Trados Studio, which is a computer-assisted translation tool. It includes tools such as a translation memory and terminological databases to speed up the translation process and increase its quality. We are now also using SDL Trados GroupShare which is a cloud-based product that allows teams to share translation memories and terminological databases among themselves.
• #56: As you know, all the element chapters follow the same structure, as outlined on the slide. To avoid having to translate the same section headings hundreds of time, the headings will be translated only once in a table in RDA’s Content Management System. The table will then be used to insert the French headings in the French files that translators will be working on.
• #57: Another objective of the 3R Project was to convert the RDA Toolkit data to the DITA format, which was implemented in 2017. The boilerplate file, as we call it, is an example of how the DITA format improved content management and the translation process. The file contains sentences and paragraphs that are frequently re-used in the RDA text. It is in fact an improved and more granular version of the recurring phrases document that I mentioned earlier. When needed in a specific instruction, boilerplate text is pulled into RDA instructions using content references or content key references. Re-used content is therefore only translated once. If revisions or corrections are needed, they will only need to be done once.
• #58: The boilerplate file contains more than 250 sentences or paragraphs of which the slide shows a very small sample. The fact that conditions and options are formulated as separate sentences in the new Toolkit helps take advantage of the boilerplate technique.
• #59: As of May 26, RDA had a little more than 3000 elements. Instructions for 84% of the elements are composed entirely of boilerplate and will be translated automatically. The element “printer agent of” is one of these elements. This means that only instructions for 481 elements (that is 16%) will have to be translated by a translator. The figures have been provided by RSC Secretary Linda Barnhart.
• #60: One issue that we had to tackle when translating RDA Reference was the removal of parentheses in element names. In the current Toolkit, parentheses are used to qualify similar labels by element range but this practice was found inconsistent and confusing in the English version and was discontinued. In English, it is not difficult to remove the parentheses and produce a satisfactory result without changing the word order. This does not work so easily in French where removing the parentheses requires changing the word order and using an adjective. Appropriate adjectives are not always available and those that are may sound pedantic or inappropriate; changing the word order also means that the most significant word will no longer be the first, which is not an issue in English. It was therefore decided to retain parentheses in some instances for consistency and to keep a more useful order where the entry is under the specific term and not the entity.
• #61: One probably never looks at a text more closely than when one has to translate it and RDA translators have proved to be good proofreaders. During the process of translating RDA Reference, we encountered and reported several editorial problems such as typos and inconsistencies between similar definitions. We also made suggestions to revise the English text whenever a clarification was needed to produce an accurate translation. We also reported more significant problems with some of the definitions. For example, the element coordinates of cartographic content was an attribute in the original Toolkit and became a relationship element in the beta Toolkit with Place as range. The original definition in the beta Toolkit did not reflect the change by defining the element as a place. This was fixed during the 3R Project.
• #62: The word issue and its cognates raise problems. Their meanings are difficult to grasp for a Francophone. For example, does issuing cover only publication? Most reference sources consulted seem to suggest so but its usage in RDA and in terms such as issuing body agent seem to indicate otherwise. What is the distinction between publication, release and issuance in a scope note such as the one for name of publisher, which reads “Publication includes release and issuance”? The distinction seems subtle enough to confuse Google translate which suggested, back in July 2019, a translation that meant “Publication includes publication and publication.” The element issuing body agent and its narrower elements pose their own difficulties. Issuing body used to refer only to a corporate body under AACR2. How to translate issuing body when the term applies to all types of agents? Could body be omitted to avoid confusion and facilitate translation?
• #63: As I mentioned earlier, the French translation team still has some work to do to complete the translation of RDA Reference. Over the past year, we have had to devote time and resources to another major project and this forced us to put the translation on hold for a while. We therefore still have several steps to go through. We set the end of May 2021 as our target date to complete the translation of the new RDA.
• #64: If you want to learn more about the original French translation project, you can refer to the article that Clément Arsenault, Pat Riva and I wrote for Cataloging & Classification Quarterly in 2014.