Joint Steering Committee meeting, November 3, 2015

The Joint Steering Committee for Development of RDA continued its meeting Tuesday morning with a discussion of various papers dealing with modeling.

Before I start, I want to warn you that these discussions are particularly challenging to summarize.  These were mostly discussion papers, they tend to be abstract, and very few specific decisions were made.  Even the decisions about what to do next were deferred until Friday, when the JSC’s program of work for the coming year will be determined.  In the discussions, lots of things get said, and some statements are modified by later statements.  It is difficult to decide which points to report, and to make sure that those points accurately reflect the sense of the discussion as a whole.  I want to give you what I think are the high points of each discussion, the points that seem to form the basis for future development of RDA — most of which will not come to fruition for some time.


Models: transcription and statements

Use of the term “recording”, “record”, and “transcribe” in RDA

This discussion was based on an internal document drafted by Kathy Glennan, the ALA representative.  The issue is the definition and appropriate use of the terms “transcribe” and “record” in RDA instructions.

The first point noted was that transcription only applies to manifestation elements; only manifestations contain transcribable data.  Transcription is a way of capturing the information by which manifestations describe themselves. The practice of transcription supports identification of particular manifestations, and matching manifestations to determine if they are identical.

The sense of the JSC was that guidelines for recording information should be broad and permissive; if information is known and is judged to be significant in supporting user tasks, RDA instructions should allow it to be recorded in an appropriate element.  Guidelines for transcription, on the other hand, should be very closely defined.  It should be very clear what elements should be transcribed, and exactly what that means.  The instructions in 1.4 (Language and script) and 1.7 (Transcription) are the most important in this regard, and the JSC Secretary will prepare draft revisions of these instructions to provide clarification.

6JSC/BL rep/2: Simplification of RDA 2.7-2.10. Follow up

Abstract:  The paper reviews and follows up the discussions on Production, Publication, Distribution and Manufacture Statements (PPDMS) at 2015 JSC Meeting.  The paper develops the arguments around simplification for PPDM Statements into an alternative approach based on elements, relationships and transcription of statements.

The proposal suggests that the various RDA PPDM elements be treated both as separate elements rather than as aggregated statements; that these elements be treated as normalized data – as relationships to the corresponding Place, Corporate Body, and Timespan entities (the latter is a new entity expected to be introduced in FRBR-LRM).  The transcribed data would be treated separately, as an aggregated statement capturing the self-identifying information that appears on the source of information.

The discussion suggested a general model for recording information: (a) generally record any information from any source that supports any user purpose in appropriate RDA elements; (b) for a limited set of elements, capture in some a more precise form the self-identifying information from the resource in a more precise manner (“transcription”).  It was suggested that photomechanical reproduction might be a possible way of capturing this self-identifying information.

As noted above (discussion of these two issues tended to merge), it was decided to draft revisions of RDA 1.4 and 1.7 that would make the guidelines for method (b) more precise; the Secretary will prepare the revisions.


Models: the four-fold path

The “four-fold” path refers to the techniques for recording relationships (specified in RDA Chapter 24): as an identifier, an authorized access point, a structured description, or an unstructured description.  At the 2014 JSC meeting, it was noted that some version of this four-fold path might provide a general model of methods for recording all types of RDA data.  The JSC Technical Working Group developed these ideas, and ALA presented two proposals that dealt with structured descriptions, a critical part of the four-fold path.

6JSC/TechnicalWG/6: RDA accommodation of relationship data

Abstract: This paper discusses the general approaches used by RDA to accommodate data about entities related to the entity being described.  The paper makes some general recommendations for developing RDA to improve its accommodation of relationship data.

The discussion actually began with 6JSC/ALA/41, but it moved quickly to the more general recommendations in the Technical Working Group paper.  Both this paper and the ALA proposal asserted the principle that a structured description should be composed of attributes of the entity being described.  Most obviously, this means that the structured description of a related manifestation should be composed of various attributes of the manifestation; less clearly, a structured description of a work might be constructed using work attributes.

Moving to the specific recommendations in the Technical Working Group paper, the first two called for clarifications in the guidelines for using identifiers to describe relationships.  These recommendations were generally accepted, but specific revision proposals are needed; the LC representative will draft these with the help of the Working Group.

Recommendations 3 and 4 characterize the concept of a structured description; this concept might include both the structured description and the authorized access point, since both are aggregated constructed using appropriate RDA elements.  The discussion of these issues was interesting but inconclusive.  This recommendation will be noted for future work, but no specific actions will be taken.

Recommendation 5 calls for documentation of the guidelines for the four-fold path.  A critical need here is to produce guidelines for the linked-data path (identifiers).  This task will be defined and assigned on Friday.

6JSC/ALA/41: RDA accommodation of relationship data

Abstract: This proposal seeks to accomplish two things: (a) it satisfies the need for explicit instructions for recording contents notes as structured descriptions of the “container of” relationship; and (b) it provides a general structure for describing relationships. This general structure could be extended by adding specific instructions for other types of relationships, as needed.

The discussion returned to the ALA proposal.  The general structure (b above) had been folded into the previous discussion.  The issue of contents notes will be revisited on Friday.

6JSC/ALA/45: Referential relationships: RDA Chapters 24-28 and Appendix J

Abstract:  Revise Chapters 24-28 and add new relationship designators to Appendix J to accommodate
referential relationships between a resource and a description or enumeration of that resource
contained in another resource such as a bibliography or catalog.

The JSC agreed that rare book references to published descriptions were not subject relationships.  They seemed to be relationships between two descriptions — between the description (record) being created and the description of the same resource in an external source.  These are not strictly related resources, but surrogates for those resources.  There was general agreement that these relationships were not appropriate for Appendix J, so the proposal was rejected.

The JSC will be meeting with representatives from the rare books community on Thursday to explore partnership opportunities; they are forming an Archives Working Group to explore partnership opportunities with that community.  This seems like an issue that these groups might be asked to explore on behalf of the JSC.  This will be explored on Thursday.

The ALA proposal also raised the issue of cross-entity relationships.  All the current relationships in Appendix J are defined between entities of the same type (e.g., work to work).  The ALA proposal required relationships between entities of different types (e.g., manifestation to expression); even though the ALA proposal was rejected, there are other use cases that support the need for cross-entity relationships.  This general issue will be referred to the Relationship Designators Working Group as a priority action item.


Models: aggregates and designators

6JSC/AggregatesWG/1: Aggregates and FRBRoo treatment of aggregates

Abstract:  This paper discusses the treatment of aggregate monographs in the FRBRoo model and compares it
with the current RDA treatment.  The paper identifies issues for developing the treatment of aggregates in RDA.

Aggregate monographs are those whose content consists of more or less independent contributions.  The type of aggregates that were examined most closely by the Working Group are those that consist of primary content augmented by other contributions; an example would be Jane Austen’s Emma with an introduction and illustrations.  The question is how to model such resources.

FRBR-LRM will complete the work of the IFLA Working Group on Aggregates; this model is still being developed. The object-oriented version of FRBR (FRBRoo) includes modeling of aggregates.  RDA decided not to deal with aggregates until the IFLA work was completed.  Since FRBRoo already exists, the Aggregates Working Group investigated its treatment of aggregates.

FRBRoo adds features to the FRBRer (the entity-relationship version of the model that is the basis for RDA).  Each of the WEMI entities in FRBRoo may be qualified by type (publication work, individual work, aggregation work; publication expression, self-contained expression; manifestation product type).  The manifestation product type includes the entire content of the manifestation and is the carrier for the publication expression.  This in turn can contain the expressions of one or more works; aggregation of the content takes place at the level of the publication expression (you really need to see the diagrams in the paper for this).  Applying this structure to the Emma example, each of the components (Emma, the introduction, and the illustrations) are separate works aggregated in a single publication expression.

One of the fascinating implications of this is that contributors turn out to be creators.  RDA would currently treat the writer of the introduction and the illustrator as contributors to the expression; in this model, these would both be creators of their respective contributions: author of the introduction, artist/creator of the illustrations.  The relationships of these persons would be to the separate works, not to the expression, and their roles would be creator roles, not contributor roles; if aggregated works were consistently treated in this fashion, there would be no need for contributor roles at all, and even the publisher could be treated as creator of a publication work.

However, the discussion noted that we do not always want to treat all of these separate contributions as works; it is sufficient (and efficient) to treat them as FRBRer and RDA currently do, as contributions to the aggregate expression.  It was also noted in discussion that there are different sorts of components — which FRBR treats as types of parts.  There are separable parts such as chapters, introductions, and illustrations, and there are “organic” parts such as the text and the music of a vocal work.  The questions are (a) what is core? what level of complexity is required in every case? and (b) beyond that core, how do you allow for alternative application of simple and complex models?

These are just some of the issues explored in this fascinating discussion.  There were no specific follow-up actions decided, but it was suggested that the most useful FRBRoo types might be publication expressions (for this sort of aggregated monographs) and self-contained expressions (for unique items such as unpublished resources, which will be of critical importance to the archives and museum communities).

6JSC/ALA/43: Revision and expansion of RDA Appendix K

Abstract:  Modify the structure of and expand Appendix K to specify reciprocal relationship designators
and allow catalogers to use a greater number of relationship designators.

This is ALA’s second attempt to provide a relatively complete set of relationship designators for relationships between agents (persons, families, and corporate bodies).  In the discussion, the view that was most strongly stated was that the nature and status of agent-to-agent relationships was currently unclear.  As was noted yesterday, FRBR-LRM may make a distinction between primary entities for bibliographic resources (WEMI) and secondary entities (everything else, including agents); secondary entities would have functional requirements limited to discovery of resources.  What I should have added is that there is some discomfort with this distinction; some argue that all entities should be given equal treatment, in which case these agent-to-agent relationships become important.

The discussion also suggested that there is a general need to decide what relationship designators are and what the lists of designators are supposed to be.  In constructing the RDA Registry, it was decided to model relationship designators as elements, refinements of the basic high-level relationships (related work, related person, etc.).  However, it is possible to model them as controlled terms in RDA vocabularies.  Are the lists themselves intended to be an encyclopedia of relationships, a vocabulary of relationship terminology?  Or are they intended to reflect how the resource describes itself?

While these issues are unresolved, the JSC is reluctant to make major changes to Appendix K, and would in fact like to declare a moratorium on proposals for any new relationship designators to give the Relationship Designators Working Group and the JSC a chance to resolve the outstanding modeling issues.  However, new relationship designators are the most common requests from users of RDA and a moratorium would be unpopular.

A compromise was agreed upon; ALA will review the responses to the proposal and submit fast-track proposals for those on which there was agreement.  The JSC will consider declaring a moratorium on proposals for all the relationship designator appendices in 2016.  All this will need to be confirmed on Friday.

6JSC/BL/27: Appendix I relationships for works issued over time

The British Library proposed two new relationship designators for Appendix I; both were approved.

The designator “editorial director” will be added, using the definition proposed by LC, with some additions from the CCC response.

The designator “founder” (of a work) will be added.

6JSC/TechnicalWG/5: RDA models for authority data

Abstract:  This paper discusses the models used by RDA to accommodate authority data.  The paper makes some general recommendations for developing RDA to improve its accommodation of authority data.

This paper deals with the question of what place authority data (control of the textual names by which entities are identified) has in RDA.  It takes off from the fact that FRBR-LRM is likely to treat Name as an entity (Nomen) rather than an attribute.

The first recommendation was to define a high level relationship has appellation and to define specific types of Nomen (preferred name, variant name, identifier, authorized access point, etc.) as refinements of this general relationship.  The second recommendation was to develop appropriate sub-elements and ranges for the existing RDA elements.  There was agreement on both of these recommendation.

Recommendation #3 was to define sub-elements of Name of the Person (e.g., Family name, Given name) and to add these to the Registry.  It was pointed out that such sub-elements are diversely defined in different cultures and that this task will not be straight-forward.  The Technical Working Group will be asked to work on the definitions, and the RDA Development Team to work on adding the sub-elements to the Registry.

Recommendation #4 was to investigate (a) whether the distinction between preferred and variant names should remain in RDA or be dealt with in application profiles, and (b) whether/how RDA should accommodate Nomen-to-Nomen relationships.  Since the recommendation was only to investigate the issue, it was agreed that the Working Group will do so.

Recommendation #5 was to generalize the instructions for constructing authorized and variant access points as part of general guidelines for creating Nomen data, and possibly to move the detailed guidelines on access points to application profiles.  All of this cannot be done until FRBR-LRM is final; in the meantime, work will continue on the development of a core RDA application profile.

Recommendation #6 was to consult with international cultural heritage communities in developing RDA guidelines and instructions.  This was also agreed.

6JSC/DNB/Discussion/1: Discussion paper: First issue v. latest (current) issue

This is a piece of unfinished business from the Deutsche National Bibliothek that relates to aggregates that are serials.  The Anglo-American community, ISBD, and ISSN base the description of a serial on the first issue, treat major title changes as new resources, and make notes on minor changes that occur after the first issue.  Conversely, in Germany, Austria, and probably other countries, the description is based on the latest (current) issue and the serial is redescribed when attributes change.  When this paper was first considered, the JSC agreed that RDA ought to be neutral on this issue, but did not develop proposals to make this happen.  The JSC Chair suggests that modeling the notes involved as structured descriptions of the resource would solve the problem.  This will be discussed further on Thursday and will likely be referred to the Aggregates Working Group.


Models: fictitious entities

6JSC/FictitiousWG/1: Fictitious and other entities in RDA and the consolidated FR models

The Working Group considered how fictitious, pseudonymous and non-human entities could be incorporated into FRBR-LRM.

In the discussion it was noted that the FRBR consolidation group has already taken a strong position that only human persons can be agents.  This is because the concept of creation is central to the relationship between agents and resources.  Agents create works; contributors create expressions; publishers create manifestations.  Only things with minds can create things; fictitious and non-human entities cannot create things.

So what are fictitious and non-human entities?  First, at this time subject entities are out of scope for RDA; we only need to deal with non-humans that purport to be agents.  But they cannot really be agents, and there must be a real person behind the fiction of agency.  So one way to model this is to consider the fictitious or non-human individual as a Nomen for a real person, just as we can consider a pseudonym as a Nomen for a real person.  Potentially, we can record the attributes of a Nomen and we can record the relationship between a Nomen and other Nomens or Persons.  The FR consolidation group is still considering how extensively to develop the Nomen entity: what attributes and relationships to define.  However, this suggests a way to model fictitious and non-human individuals.

One of the problems with this is that it is doubtful that a Nomen can stand on its own; it is the name of some entity.  When you know the identity of the real person, this isn’t a problem.  But what if you don’t know?  You have a Nomen entity that is not related to any particular entity or type of entity; it is the name of … what?  For fictitious persons that purport to be authors, this occurs, but is not typical.  However, for non-humans with names, what is the entity to which that name can be related?  By our definition, we know that this entity cannot be an agent, because agents must be human.  So what sort of thing is it?  Is it sufficient to simply describe and relate Nomens and ignore the entities that they name?

This was a long and complex discussion, and there were many more comments than I can capture here.  The JSC rejected the Working Group recommendations.  The group will be asked to start modeling the Nomen approach to fictitious and non-human entities in order to address the use cases for purported agency.  Further issues may be added when the program of work is discussed on Friday.


Tomorrow’s agenda includes groups of proposals dealing with instructions for manifestation, work, and agent attributes (including a number of proposal relating to musical works), and (finally) some proposals relating to internationalization of RDA.  In contrast to the very abstract proposals dealt with today, most of tomorrow’s proposals are for specific revisions to RDA instructions.  This is a change of pace that I will welcome; writing these summaries has been a challenge.

Be warned that the summaries of today’s discussions are not complete and may contain inadvertent misrepresentations of the discussion.  I hope that you find them interesting and useful.  If nothing else, I hope you get a sense of the extent and complexity of the issues that will need to be resolved in the next few years — and the significance of the changes that will need to be made in RDA.

, , ,

Powered by WordPress. Designed by WooThemes

Skip to toolbar