Proposal: Simplification of RDA 2.7-2.10

6JSC/BL rep/1
15 August 2014

Simplification of RDA 2.7-2.10

British Library (BL)

Print Friendly, PDF & Email

Tagged , . Bookmark the permalink.

17 Responses to Proposal: Simplification of RDA 2.7-2.10

  1. John Myers says:

    This is an interesting alternative solution to the “cascading vortex”. From purely partisan motivations its approval would completely derail our proposal regarding Production Statements. On theoretical grounds though, I fear it does grave injury to the underlying modeling of the relevant statements and their sub-elements. It essentially reinstates AACR2’s aggregate treatment of “Publication, Distribution, Etc., Area”. Lastly, I’m not convinced that “number of pages in MSWord” is adequate justification for this manner of editing a resource expressly formulated for an online, non-linear presentation.

  2. Kathy Glennan says:

    The current PPDMS instructions are core or core if. The suggested use of relationships instead of these current aggregated statements may be fine from a model standpoint, but those relationships aren’t core. (At least they aren’t now.)

  3. Kathy Glennan says:

    I am especially interested in hearing from members & liaisons on the answers to the questions on p. 12-13 of the discussion paper.

  4. Kathy Glennan says:

    In Scenario A, Recording Changes in Publication Statements:

    I believe this scenario encompasses more potential changes than RDA as currently worded, since the revision is about changes in the entire transcribed publication statement.

    I think the situation is the same in Scenario B.

  5. Kathy Glennan says:

    Are there reasons that we no longer want to make explicit distinctions between publication, distribution & manufacture information?

  6. Kathy Glennan says:

    The ISBD presentation of the PPDM elements into aggregated statements is easy to scan/parse. Transcribed statements from the resource are not.

    Would recording relationships adequately replace that function of the ISBD structure?

  7. Kathy Glennan says:

    I don’t understand Appendix C — there are categorizations of 5 types of manifestation dates in the scope statement, but then there are no specific instructions for these.

    I also disagree with the characterization of “copyright date” as a manifestation date. (But that’s really an entirely different can of worms!)

    • Dominique Bourassa says:

      I agree with Kathy. How would one record this type of information?

      Also agree about copyright. If we go the route proposed by this discussion paper, would copyright dates have to be moved to 6.10.

  8. Robert Bratton says:

    I agree with the concerns raised by John and Kathy. I honestly don’t understand the need for these proposed changes. We catalog a fair variety of materials at my shop, and we haven’t had any major problems with RDA 2.7-2.10.

  9. Matthew Haugen says:

    Posting a lengthy response from RBMS (part 1 of 3):

    RBMS thanks the BL representative for this important proposal and offers its enthusiastic support for this proposal, particularly Scenario B.

    Primary benefits:
    1. The changes proposed provide a method for making cleaner separations of transcribed information and recorded information. A simplified transcribed statement, free of supplied information or interpolations such as qualifications, corrections, or normalizations, will better fulfill the RDA principles of representation and differentiation. Other existing elements, along with new elements for Place and Timespan will provide better means to fulfill the principles of relationships and accuracy.
    2. An issuance statement comprised of an unparsed transcription will better allow the data to document the information as it appears in the source without transposition or interpolation. Especially in the context of early printed resources, the distinction between printers, publishers, and booksellers is often indistinct by modern definitions, multiple roles/functions may be combined in a single statement, and grammatical complexity or ambiguity of statements can hinder accurate parsing into distinct elements.
    3. The proposal for treating place as a related entity will better fulfill a user’s ability to find sets of resources by place, which is important to our community. It is already common in rare materials cataloging to record a normalized place name in MARC 752 in addition to the transcribed place in 26x $a, especially for Latin place names and other historical usages, false imprints, and forgeries.
    4. The simplification of guidelines would also be a welcome change.

    Answers to questions posed in the proposal
    1. We prefer Scenario B, a single generalized element for “issuance” transcription, but consider Scenario A to have many of the same benefits.
    2. We support the development of RDA elements for Place & Timespan (option a).
    3. For consistency, Production should be treated in the same way as Publication, Distribution and Manufacture. For Production information, the present proposal and the discussion paper to which RBMS made contributions (6JSC/ALA/Discussion/4) both offer improvements on the current practice prescribed by RDA for unpublished resources. The present proposal would satisfy most if not all of the concerns included in the ALA discussion paper.
    4. Yes, the approach to simplification of PPDM statements should be applied to Edition Statement and Series Statement.
    5. It seems sensible to retain subtypes of dates of manifestation and possible to do so without duplicating instructions, as manifestations may have multiple associated dates.
    6. Issuance isn’t an ideal term, especially for unpublished resources, but we could not readily identify a preferable alternative. The instructions could also retain the PPDM vocabulary, perhaps closer to scenario A, while still combining the instructions.

  10. Matthew Haugen says:

    RBMS Response (part 2 of 3)
    Additional Thoughts
    1. RDA 21.2-21.5: We think it would be sensible to consolidate 21.2-5 into a single element / set of instructions in parallel to the changes in Chapter 2. The guidelines vary little from Producer to Publisher, Distributor, and Manufacturer. If reduced to a single element, use of the appropriate relationship designator would be sufficient to identify the function of the related agent. However, additional relationship designators would need to be added to RDA Appendix I.4.1. Consolidation of these elements would have many of the same benefits (and logic) as the proposal’s Scenario B.

    2. Copyright: We think it may be useful to include RDA 2.11 Copyright Date as part of the present proposal. As with the PPDM statements, why not include the option to transcribe copyright statements here, along with recorded copyright date as a related entity (Timespan). These statements contain information about copyright agents/holders and places/jurisdictions, which could also be recorded as related entities. Both transcribed and recorded copyright information may be useful for finding and identification of published resources, especially where it differs from PPDM data, and if it forms the basis for supplied information in other elements. This, however, would raise another question: Is non-transcribed copyright information an attribute of Expression instead of Manifestation?

    3. Event element: We are uncertain of the full implications of direct linking versus indirect linking (via an Event element). The Event model, however, seems instinctively correct, in the manner that it appears to associate elements that are semantically connected. That is, if a single Manifestation has multiple clusters (Events) concerning publication, it is functionally necessary to associate Publisher1 with Place1 and Date1, Publisher2 with Place2 and Date2 – and the Event entity appears to enable these associations.

    4. The approach applied to other areas of RDA: It is curious that the proposal suggests Series Statement and Edition Statement as additional candidates for this approach, but not Title and Statement of Responsibility. The latter are also aggregate statements. In some respects, they may be considered a single aggregate statement, given the frequency with which one blends into the other. If essential aspects of the Title-SOR Statement, such as Title Proper and creator/contributor (treated in chapters 19 & 20), etc., are additionally recorded in controlled form, it’s unclear what we gain by parsing the transcribed Title-SOR into separate sub-elements. Confusion about what goes where (as noted in LC’s proposal concerning noun phrases) could be alleviated by making the transcription less granular. If the JSC would like to pursue the possibility of treating Title and SOR with this approach, the RBMS community would be willing to put forward the proposal (via ALA).

    5. RDA Core: The proposal suggests that the transcribed PPDM statements would remain Core in RDA (in scenarios A and B); it appears to imply that the data recoded as related entities would not be Core. Clarification on this issue will be useful. In some contexts, such as special collections cataloging, the transcribed form of data will remain essential (perhaps in addition to the controlled form of the data). We wonder, however, about the merits of favoring transcription in general. We suggest thoroughly evaluating the two approaches to discern which better responds to user needs, especially if only one can be Core. At the very least, Production information for unpublished resources is better served by non-transcribed data (as noted in 6JSC/ALA/Discussion/4). Even for published resources, identification of a specific manifestation may rely on attributes that are not–and sometimes cannot be–conveyed in transcription, such as devices/logos, typography, layout, etc., while normalized data also assists with identification in addition to finding. Transcribed PPDM statements are already somewhat resistant to machine parsing, and the proposed deprecation of subelements within the statements would only make them less so, without the recorded or controlled equivalents. On a related note, the example guidelines for Scenario B are worded in a way that would require the cataloger to transcribe Publication, Manufacture, and Distribution information, whereas only the Publication Statement is currently Core.

    • Matthew Haugen says:

      RBMS Response (part 3 of 3)
      6. Normalized dates (Optional Addition): In the example guidelines in the Appendices, the instructions retain an Optional Addition to provide “numerals in the form preferred by the agency creating the data.” It would be better to omit this option, since a separate element (Timespan) would be available to record the date in a controlled form.
      7. Sources of information: Similarly, we would like to see transcribed statements restricted to only that information which is found within the resource. If the source of information is outside the resource (applying instructions at 2.2.4), it would be more logical to record the information in the form of a related entity, providing notes on the source of that information if important and applicable, rather than supplying information in transcribed statements.

      8. Details Relating to Issuance Statement: Many of the notes we record here concern facts instead of the statement as it appears in the source, e.g.: “Actually published in Dublin” or “Ceased publication in 1999.” Such notes can help clarify why the transcribed statement and the corresponding recorded elements differ. However, these two examples more properly concern the related Place and Timespan entitities, not the statement transcribed, so we would need to add appropriate “Details on …” elements to those entities for such notes.

      9. Multiple statement sources: It is unclear how a cataloger would treat simplified PPDM transcriptions if the sources of information come from multiple locations. Is the simplified element intended to be repeatable? Would we record a new instance of the element for every piece of the transcription that is non-consecutive in the source? It is common, for example, for a portion of an imprint statement to appear on the recto of the title page, the remainder on the verso. It is also somewhat common for early resources to have contradictory or earlier/later statements between title page and cover (as in re-issues; cf. RDA 2.2.3.2). In these scenarios, do we need at least two instances of the transcribed statement? And would it be necessary to include notes on the source of each statement as details relating to the issuance statement? Or, would an approach to similar to that of variant titles be useful (e.g. 264 |i Statement from cover: |a Published in Boston by…)?

  11. Dominique Bourassa says:

    When I read this proposal, I had the same concerns as those raised by Kathy and John but I had a feeling RBMS would love the second approach proposed (scenario B). I can see why the rare book community would appreciate such an approach. There is indeed something nice and useful about being able to transcribe such long statements.

    This paper says that at the moment “from the cataloguer’s perspective, this [the total number of instructions] is a lot of content to master” (p. 2) It aims to simplify the current approach. However, one could argue that reducing the number of instructions in 2.7-2.10, but adding new instructions at 2.X and adding other instructions to record attributes or entities will actually make the situation more complex by requiring the cataloger to consult different sections of RDA to find out how to record this information.

    One thing I like from this discussion paper is that it attempts to make the PPDMS more machine actionable and linked data friendly. But because this paper offers more than one option, it is difficult to envision what the end result would look (scenario A with entity option; scenario A with attribute option; scenario B with entity option; scenario B with attribute option).

    Regarding scenario A: It’s difficult to see how this would exactly work because only the instructions for publication statement (2.8) are given. This discussion paper is supposed to help eliminate the vortex of horror. But it’s not clear what will happen with the distribution and manufacture statements if this scenario is chosen. Will they remain “core if” (I doubt that’s what the BL wants, otherwise the problem will remain), become “core,” or not be core at all. This could make a big difference in the number of statements one will have to record.

    If we don’t have a publication statement, I suppose we would not write anything anymore (no more [publisher not identified] statement), and just go into recording a distribution statement instead. That’s not clear in the way scenario A is written.
    The examples in 2.8.1.4 in Scenario A are the same as in scenario B. The last example in 2.8.1.4 shows information that does not belong to publication statement per se (sold by). If we had a long statement that said published by … and distributed by… would the statement be kept together or split into 2 statements?

  12. Kathy Glennan says:

    Comments from Francis Lapka, Catalog Librarian, Yale Center for British Art

    I support the RBMS response and offer these additional thoughts, which do not necessarily represent the views of the RBMS community.

    1. To me, the most important part of the discussion paper is the suggestion to record PPDM place and date as related entities. Even if no changes are made to 2.7-2.10, RDA should be modified to allow for these new relationships. They will be vitally useful.

    2. If RDA gives us this mechanism to record place, agent, and date (for PPDM data) as related entities, I argue that this mechanism is the primary method by which we should record PPDM information. If we’ve already agreed that linked data represents the future of cataloging – and if we really want our data to interact with data from other communities – the inevitable conclusion is that we must record PPDM information as relationships. Transcription of PPDM statements, recorded in parallel, will remain useful in many contexts (for additional keyword hits, sometimes for identification, or as justification for the data recorded in other fields, etc.), but it seems reasonable to me that the transcribed PPDM information could be of secondary importance and given a lower priority in catalog displays (and perhaps not required in some contexts).

    3. John and Kathy suggest that the proposal would lead us back to the days where we did not distinguish between publication, distribution, etc. While it’s true that the distinction might be erased in the transcribed element, it would remain very much encoded where the PPDM data is given as related entities. When considering this proposal, I think it’s very useful to imagine OPAC displays in which our traditional publication (etc.) line(s) are populated from the data entered in controlled form. The controlled form of the data could still be displayed with ISBD syntax, if that form best enables identification or parsing.

    4. Those who’ve tinkered with the sample BIBFRAME editor (http://bibframe.org/tools/editor/) will find that BIBFRAME has already anticipated the changes recommended by the present paper. Go to the editor, create a “New Instance,” and scroll down to find the button for “Provider Entity” (which appears to mimic the semantics of the Event proposed in the present paper). Here the editor gives us a means to establish relationships – with handy look-up tables! – to related places and agents. Even in its somewhat clunky prototype form, this demonstrates the relative ease with which this new method for recording PPDM could be approached. Note that BF also provides a “Provider Statement” element to record transcribed PPDM information.

  13. Steve Kelley says:

    I’ll confess I had a difficult time following this proposal, but I generally agree with John Meyer’s comment that this proposal seems to reinstate the aggregate production statement from AACR2. I prefer the Production Statement revision approach that we have suggested.

  14. Tracey L. Snyder says:

    A colleague in MusLA questions the validity of the Find vs. Identify dichotomy, suspecting that PPDM statements are not used to Find things in significant numbers (with rare materials, etc. being an exception).

    I had noticed, with some unease, what Francis pointed out in Bibframe. We are accustomed to thinking of bringing uniformity to PPDM elements as being outside of our job as catalogers. We just transcribe and leave it at that. But isn’t that antithetical to linked data? I am starting to get used to the idea of deemphasizing transcription for PPDM, since it won’t necessarily help people Find stuff that was published in a certain place, etc. The more I think about it (not that I want catalogers to have to do twice the work), both transcribed and uniformly recorded forms are valuable. Just the other day, I answered a reference question by email from someone looking for scores held here published before 1945 having something to do with Japan. We had a score with 260 $a Edo. This is an illustration of both transcribed and recorded/controlled forms being useful. Someone might decide to search on Edo, figuring it would pull stuff from when Tokyo was called Edo. Or someone might want to facet by publication location of Tokyo, figuring that would pull up everything, no matter what it was called on the resource.

  15. Matthew Haugen says:

    I was glad to read the various constituency responses, and have found myself growing increasingly ambivalent about the best approach to the complex issues raised by this proposal. I did note that some of the responses brought up display issues and citation software as reasons not to support these changes to transcription. I agree, insofar as recorded/controlled data might better lend itself to manipulation for display and citation purposes, but it seems to me those same concerns are also true of the current RDA practice when compared to AACR2 which employed more omissions, abbreviations, and corrections for the sake of clarity within statements.

Leave a Reply