We're dealing with two different aspects of series information: 1) the series statement appearing on the manifestation, and 2) the series part-whole relationship expressed as actionable data (access points, identifiers, etc.). LC (and some others) work with the former, and not with the latter. There's no point of having an identifier for a series statement. Each series statement is a unique thing in itself. It's exactly the same as with the title proper and statement of responsibility for the resource being cataloged. You absolutely do *not* want the string pertaining to this manifestation to end up relating also to some other manifestation; it is peculiar to this single manifestation. I don't believe it's appropriate to have identifiers for any *transcribed* elements in a bibliographic description, no matter how much any of them may end up looking identical. The statement "Series on cataloging conundra" appearing on Manifestation X may *look* the same as the statement "Series on cataloging conundra" appearing on Manifestation Y. But those statements are not the same thing. (I'm having flashbacks to studying Hayakawa back in college freshman English... "Cow 1 is not cow 2...") They are transcriptions of physical manifestations and each is unique, and they must not be identified as being the same thing.
Now, if we're talking about having a system go ahead and in the background take the *series statement* and use it to link up the resource with an identifier for a *series*, then you are getting into the world of authority work. You are either forcing the cataloger to select from established identities (something LC has said they will not do), or you are going ahead and making a selection (or establishing a new identity) without the cataloger's input, which is getting into rather dangerous territory.
I'm quite baffled by the statement that RDA is not an identifier-based cataloging code. RDA has "identifier" as the very first option for recording all relationships!
Kevin M. Randall
Principal Serials Cataloger
Northwestern University Library
[log in to unmask]
Proudly wearing the sensible shoes since 1978!
> -----Original Message-----
> From: Bibliographic Framework Transition Initiative Forum
> [mailto:[log in to unmask]] On Behalf Of Karen Coyle
> Sent: Monday, August 10, 2015 1:41 PM
> To: [log in to unmask]
> Subject: Re: [BIBFRAME] Proposal for treatment of series in BIBFRAME
> Kevin, and that's exactly why I replied as I did to Adam. For the
> purposes of data in RDF, the creation of identifiers is not the same act
> as creating an authority record or even an authoritative form of a name.
> You can create an identifier and 1) change the display string at any
> time 2) declare this identifier to be for the same "thing" as another
> identifier at any time. It is no longer a question of "indexed strings"
> -- those do not have the weight or authority that they have in
> text-based cataloging. It's something entirely different, and it also
> isn't a manual operation.
> Consciously or not (and I suspect it is not) the act of cataloging
> creates machine-actionable data. "Indexes" exist for most words in the
> record, without the cataloger purposely creating those. Indexes also
> exist for any searchable strings or codes. Each index entry has an
> internal identifier. These machine indexes are not the same thing as the
> textual access points that are created in the catalog record. We need to
> move away from thinking of our data as "only" text strings. Text strings
> are not good identifiers, and are definitely not identifiers in a
> machine environment. We need to "identify" separately from human-
> displays, and realize that inside the machine identifiers rule. RDF
> means sharing identifiers, not strings. Our metadata is about identified
> things, not strings.
> Honestly, RDA missed the boat by not producing an identifier-based
> cataloging code. Instead, our "future" cataloging rules were out of date
> with current technology before they were fully formatted as MS Word
> tables. A cataloging code could be identifier-based and still produce
> consistent displays. There's no conflict here, but continuing to focus
> on text strings while ignoring identifiers is totally out of sync with
> how post-analog technology works.
> On 8/10/15 10:51 AM, Kevin M Randall wrote:
> > I took Adam's point as being that LC is not including *indexed strings*
> (i.e., "series headings" or "access points") for series. They are only
> including the series statement, the transcribed part of the series
> information. LC has abandoned the idea of making explicit links between
> the resource and the series of which it is a part. Recording a series
> identifier is in essence the same thing as recording the series authorized
> access point: you're identifying the particular work (expression,
> manifestation, whatever). And if you have decided that you are *not*
> going to make that identification (as LC has decided), then you will *not*
> want to be trying to match the series statement with any "indexed string"-
> -and you will not want to be having any identifier established when there
> are no matches, either, because that is the essence of authority work. LC
> has deliberately removed itself from the business of explicitly making
> statements about the authoritativeness, or identity, of the related
> resource; it is only transcribing information found on the resource being
> > Kevin M. Randall
> > Principal Serials Cataloger
> > Northwestern University Library
> > [log in to unmask]
> > (847) 491-2939
> > Proudly wearing the sensible shoes since 1978!
> >> -----Original Message-----
> >> From: Bibliographic Framework Transition Initiative Forum
> >> [mailto:[log in to unmask]] On Behalf Of Karen Coyle
> >> Sent: Saturday, August 08, 2015 12:30 PM
> >> To: [log in to unmask]
> >> Subject: Re: [BIBFRAME] Proposal for treatment of series in BIBFRAME
> >> Adam, establishing an identifier is not the same as creating an
> >> authority record. In fact, your system today establishes an identifier
> >> for every record and every indexed string without you having to do
> >> anything. Identifiers serve machine needs and generally are not
> >> by humans. If, rather than typing in a series name you type in the
> >> beginning characters and see a suggest list, that's all you need to do.
> >> There will be an identifier for anything in that list. And if your
> >> series isn't on the list, by creating a new entry in the list an
> >> identifier will be created.
> >> That begs the question of what happens in the future to what we call
> >> "authority records." My guess is that it will be possible to add
> >> information to any identified thing, and that more information about
> >> thing (like alternate labels, beginning dates, publisher names...) will
> >> be added as people either come across that information or find it
> >> useful. Presuming that we share data, this works very much like
> >> authority record sharing today. Undoubtedly, there will need to be
> >> judgments regarding data quality, such that only those descriptions
> >> meet certain standards will be deemed "authoritative", with the rest
> >> being mere "information." But none of this is directly related to
> >> creating identifiers, except that anything we want to "talk about" will
> >> have an identifier.
> >> kc
> >> On 8/7/15 6:34 PM, Adam L. Schiff wrote:
> >>> Many libraries, most importantly LC, are only transcribing series
> >> statements from resources and not tracing them (i.e. not providing a
> >> controlled authorized form as a related series work/expression access
> >> point). Is BIBFRAME going to require those libraries to reverse course
> >> because they must link to a series every time? Will that force them to
> >> have to establish an identifier for each and every series, and possibly a
> >> preferred name/authorized access point for said series?
> >>> Adam L. Schiff
> >>> Principal Cataloger
> >>> University of Washington Libraries
> >>> Box 352900
> >>> Seattle, WA 98195-2900
> >>> [log in to unmask]
> >>> (206) 543-8409
> >>> (206) 685-8782 fax
> Karen Coyle
> [log in to unmask] http://kcoyle.net
> m: +1-510-435-8234
> skype: kcoylenet/+1-510-984-3600