The Hebrew 100 in Appendix D Model A (Zilbershtain, Yitshak ben David Yosef, with dots below some letters) is following ALA-LC romanization.  In this case the systematically romanized form happens to be the authorized form, so this example does not illustrate the HAPY practice we've been talking about.  Ditto for the Arabic example below.  And neither example includes a name with dates, so they're not providing any guidance there either.

Robert.

D. Brooking wrote:
[log in to unmask]" type="cite">
Caroline,

The MARC documentation for 880 and for App. D. both give examples of access fields, not just descriptive fields. Where are you reading that it applies to descriptive fields only?

(This is very hard over email. Are we going to meet in Chicago? Clearly we have too much documentation! Or maybe not enough documentation spread over too many places... )

Caroline, maybe you can look at the Hebrew example in MARC App. D. Is the romanized 100 following ALA transliteration? (I notice ligatures are missing in the Russian example, so it's not clear all the diacritics are represented in the examples, sadly, it might be hard to tell.)



************
Diana Brooking             (206) 685-0389
Cataloging Librarian       (206) 685-8782 fax
Suzzallo Library           [log in to unmask]
University of Washington
Box 352900
Seattle WA  98195-2900

On Wed, 20 May 2009, Miller, Caroline wrote:

All,

Therein lies the problem.  I read through all the MARC documentation cited below and as far as I can tell the assumption is that 880 parallel fields are for direct transcription of non-roman data from the piece and the Romanized fields are direct standard romanizations of those fields **for descriptive fields only** not for fields falling under authority control.  There are no guidelines for those fields.  I believe our charge was to come up with guidelines for non-roman data in authority-controlled fields within the bib record.  Does this make sense?  Peter?

Am I missing something?

Caroline

Caroline R. Miller
Head of Monographic Cataloging
and Authority/Database Maintenance Sections
UCLA Library Cataloging & Metadata Center
BOX 957230
11020 Kinross Ave.
Los Angeles, CA 90095-7230
E-mail:  [log in to unmask]
Phone:  (310) 825-4446
Fax:  (310) 794-9357


-----Original Message-----
From: Deng, Shi [mailto:[log in to unmask]]
Sent: Wednesday, May 20, 2009 5:34 PM
To: D. Brooking; Suzuki, Keiko
Cc: Fletcher, Peter; Miller, Caroline; Avetyan, Nora; Joanna Dyla; [log in to unmask]; Julie Su; Thurston, Patricia; [log in to unmask]; [log in to unmask]; Deng, Shi
Subject: RE: PCC Non Latin task force new member; timeline; update

Dear Diana, and all,

I was talking about MARC standards. When non-Latin data recorded in a bib record, they are in the 880 fields as "Alternate Graphic Representation" (http://www.loc.gov/marc/bibliographic/bd880.html) or "parallel fields" as catalogers often refer to. It is defined: "Fully content-designated representation, in a different script, of another field in the same record." In 880 section, it also refer to MARC21 Bibliographic Appendix D on Multiscript Records (http://www.loc.gov/marc/bibliographic/ecbdmulti.html). It talks about model A vernacular and transliteration. To my understanding, 880 fields used for the non-Latin script data of which we based for transliterating into Romanized data according to ALA/LC Romanization Tables and to be linked to the transliterated field. Of course, if there is no transliterated data, it won't be linked but would be allowed to stay in the OCLC MARC record now.

Because all name headings in bib records, currently established according to AACR2/LCRI, are in Romanized form, if we provide a parallel field (meaning to link together both non-Latin script data to its transliterated representation), it should be the non-Latin script data we based for transliteration. Otherwise it was not called parallel field, but a hybrid field.

If my interpretation was wrong, then the CEAL cataloging community won't have issue with the access points in bib records for author who have names in both non-Latin (i.e. Chinese) and Latin (i.e. English) form that could not display together in access points in bib records, for example, LCCN n  50057442 for ??? = Ye, Weilian, Yip, Wai-lim. Yip, Wai-lim is the heading, and right now it would be called hybrid if it links to ???.

That's why I said yes to Peter's question or statement regarding providing parallel field to name headings.  Actually if you read our charge the last second bullet, it says: "Any resulting documentation must conform to MARC21 Model A: ?vernacular and transliteration: parallel fields (MARC 880 fields) are used when data needs to be duplicated to express it in both the original vernacular script and transliterated in one of more scripts.?*** (http://www.loc.gov/marc/bibliographic/ecbdmulti.html)"

Diana, I think I was not accurate to refer to 880 when I should refer to Appendix D. Sorry to cause the confusion.


Best regards,
--Shi

-----Original Message-----
From: D. Brooking [mailto:[log in to unmask]]
Sent: Wednesday, May 20, 2009 1:02 PM
To: Suzuki, Keiko
Cc: Deng, Shi; Fletcher, Peter; Miller, Caroline; Avetyan, Nora; Joanna Dyla; [log in to unmask]; Julie Su; Thurston, Patricia; [log in to unmask]; [log in to unmask]
Subject: RE: PCC Non Latin task force new member; timeline; update

CONSER is a PCC document, not MARC standards. Shi seemed to think there
was something in the MARC standards that mandated that no non-Latin field
should be supplied if the romanized heading was not ALA-romanization.
Maybe she was thinking of Appendix O? If so, that only applies to CJK, not
Hebrew, etc.


************
Diana Brooking             (206) 685-0389
Cataloging Librarian       (206) 685-8782 fax
Suzzallo Library           [log in to unmask]
University of Washington
Box 352900
Seattle WA  98195-2900

On Wed, 20 May 2009, Suzuki, Keiko wrote:

Diane, Maybe you are talking about "Appendix O.  Creating records with
data in nonroman script for Chinese, Japanese, and Korean serials" in
CONSER Editing Guide? Try "#7. Headings [Optional]". - Keiko

-----Original Message-----
From: D. Brooking [mailto:[log in to unmask]]
Sent: Wednesday, May 20, 2009 1:54 PM
To: Deng, Shi
Cc: Fletcher, Peter; Miller, Caroline; Avetyan, Nora; Joanna Dyla; [log in to unmask]; Julie Su; Suzuki, Keiko; Thurston, Patricia; [log in to unmask]; [log in to unmask]
Subject: RE: PCC Non Latin task force new member; timeline; update

Shi,

Could you tell me where in the MARC standards it specifies exactly the
form of headings that should go in 880 fields and when to leave them out?
I can't find anything that addresses this issue.


The HAPY community has been supplying parallel non-Latin fields for
non-ALA romanized headings for ages. In fact, I think someone told me that
even if a heading is in ALA-romanization, the tendency is to take the
non-Latin form found on the piece and leave off qualifiers and dates and
such. Part of the problem is the whole right-to-left thing, that makes it
hard to mix in qualifiers?? It would be good to know the basis for this
practice before we decide anything about it. I truly don't understand it.



************
Diana Brooking             (206) 685-0389
Cataloging Librarian       (206) 685-8782 fax
Suzzallo Library           [log in to unmask]
University of Washington
Box 352900
Seattle WA  98195-2900

On Wed, 20 May 2009, Deng, Shi wrote:

Peter, I assume that you asking question regarding name headings in 1xx,
6xx, and 7xx.  To answer your question, Yes, it should be the same
practice in CJK that we simply do not provide CJK script equivalent if a
non-ALA/LC romanzied form was established as a heading. This is bound by
MARC standards for 880 application. So it would be out of our hand to
decide something differently, not that as if we haven't thought about
it.

Best,
--Shi

-----Original Message-----
From: Fletcher, Peter [mailto:[log in to unmask]]
Sent: Tuesday, May 19, 2009 5:00 PM
To: Miller, Caroline; D. Brooking; Avetyan, Nora
Cc: Joanna Dyla; [log in to unmask]; Julie Su; [log in to unmask]; [log in to unmask]; [log in to unmask]; Deng, Shi; [log in to unmask]
Subject: RE: PCC Non Latin task force new member; timeline; update

What the best practice would be for parallel non-Latin *heading* fields in *bibliographic* records in such cases--Hebrew and Yiddish found forms and what the cataloging rules tell us to use as the heading--is what we, and especially the Hebraica experts on this task force, can decide (is this also an issue for Arabic and Persian?).

Do we simply not provide the Hebrew equivalent (i.e., no parallel field) if a non-ALC/LC Romanized version is used as the established heading? This is something we need to think about and decide. We need to be careful. Joanna's example was a good example of how sort of free-form it could get. To me consistency is really important in our work for various reasons (consistent headings in the catalog for the user; cataloger time, etc.), and the ALA/LC Romanization tables are essentially our basis for consistency.

Peter

-----Original Message-----
From: Miller, Caroline
Sent: Friday, May 08, 2009 1:07 PM
To: 'D. Brooking'; Avetyan, Nora
Cc: Joanna Dyla; Fletcher, Peter; [log in to unmask]; Julie Su; [log in to unmask]; [log in to unmask]; [log in to unmask]; [log in to unmask]; [log in to unmask]
Subject: RE: PCC Non Latin task force new member; timeline; update

I know that with Hebrew and Yiddish there are different rules.  If there is a Romanized form in the piece, you are required to use that as your heading even if the entire piece is in Hebrew or Yiddish.  More often than not, the found form is not in standard Romanization per ALA-LC.  Most works published in Hebrew do not contain vowels.  There are often multiple ways to Romanize a name with the Israeli phone book being the reference source for modern Israeli authors.  Automated reverse Romanization from the ALA-LC Romanization table for Hebrew would introduce a lot of errors.

I think this topic is critical for the new ALCTS non-roman group now being formed because they will be discussing a lot of policy issues.  I surely hope that OCLC will be represented on that group.  It would also be nice to have someone from this PCC group be a liaison to the ALCTS group.

I believe that the charge of this group is to review existing PCC documentation and, by extension (IMO) LC documentation, for inconsistencies with an eye toward recommending more standardized practices across all the available non-roman scripts and other non-roman scripts as they become available in OCLC.  There is much to discuss about non-roman issues that have been ignored far too long but we can't do it all!

Caroline R. Miller
Head of Monographic Cataloging
and Authority/Database Maintenance Sections
UCLA Library Cataloging & Metadata Center
BOX 957230
11020 Kinross Ave.
Los Angeles, CA 90095-7230
E-mail:  [log in to unmask]
Phone:  (310) 825-4446
Fax:  (310) 794-9357


-----Original Message-----
From: D. Brooking [mailto:[log in to unmask]]
Sent: Friday, May 08, 2009 12:48 PM
To: Avetyan, Nora
Cc: Joanna Dyla; Fletcher, Peter; [log in to unmask]; Julie Su; [log in to unmask]; Miller, Caroline; [log in to unmask]; [log in to unmask]; [log in to unmask]; [log in to unmask]
Subject: RE: PCC Non Latin task force new member; timeline; update

This is complex.

The record Joanna found has a mistake in it. The name heading was
established in its Russian form, so the non-Latin should be Russian as
well. (OK, at least according to the informal practice some Slavic
libraries are using, based roughly on CJK.) So if we as a group want to
characterize the example Joanna found as "wrong," we don't have to account
for it except to explain to people how not to make that mistake.

There is a whole other can of worms about which form of the name to choose
as the established one. And for the former USSR it can be hard, since
Russian no longer dominates in all the republics as it once did. I don't
want to talk about that, though, that's NACO and AACR2 and I think beyond
our scope here.

BUT, it does bring up, what is the purpose of the non-Latin parallel
fields in the bibs? I think I agree with Joanna that the place for
alternate non-Latin headings (like a correctly formulated Ukrainian form)
would be in the authority record. But we don't yet have a policy for
non-Latin cross-references in authority records. And PCC (LC?) didn't want
to designate a preferred non-Latin form (which would have been the obvious
candidate as the required parallel field to headings in bibs, had it
existed).

I think mismatches happen more frequently with HAPY languages than
Cyrillic actually, where catalogers will provide the non-Latin form found
on the piece as a parallel access field, even if it doesn't match the
established romanized heading. Maybe someone can speak to that practice.

************
Diana Brooking             (206) 685-0389
Cataloging Librarian       (206) 685-8782 fax
Suzzallo Library           [log in to unmask]
University of Washington
Box 352900
Seattle WA  98195-2900

On Fri, 8 May 2009, Avetyan, Nora wrote:


Has Golovakha, Evgenii? Ivanovich written anything in Russian? I think the rule calls for establishing the name in the language of the book
that is being cataloged. I encounter the same problem with Armenian books, because during Soviet era they had to write works (especially
the scholarly works) in Russian to be published and to be read by more people. I have many Armenian authors with Russian transliteration
in 1xx and 4xx in Armenian form, even though majority of their works are in Armenian.

I just had a thought, I am not sure.



From: Joanna Dyla [mailto:[log in to unmask]]
Sent: Friday, May 08, 2009 12:13 PM
To: Avetyan, Nora
Cc: Fletcher, Peter; [log in to unmask]; [log in to unmask]; Julie Su; [log in to unmask]; Miller, Caroline;
[log in to unmask]; [log in to unmask]; [log in to unmask]; [log in to unmask]
Subject: Re: PCC Non Latin task force new member; timeline; update



I have been having many of the same observations and concerns that everyone else who just spoke.  But then, I come across a bibliographic
records with parallel fields that raise so may questions.

Please check the following OCLC record no: 39812309, Suspil?stvo, shcho transformui?e?t?si?a? : $b dosvid sot?s?iolohichnoho monitorynhu v Ukrai?ni / $c
I?E?vhen Holovakha.
This title is in Ukrainian (apologies to those of you who do not work with Cyrillic languages) and the record has parallel Cyrillic fields
added.  However, there is a problem with the parallel 1xx.  This author's name has been established in Russian (that also, like Ukrainian,
uses Cyrillic script) and the transliterated name in 1xx is in Russian per n79118651, ARN 350318.  Information transcribed in other fields
in this record reflects the language of this item, which is Ukrainian.  So, while other parallel fields are given correctly in the
original language and script: Ukrainian/Cyrillic, the parallel 1xx heading has been wrongly "de-transliterated" into Ukrainian, even
though it is established and correctly given in 1xx in Russian. I hope that you can see on the OCLC record what I am trying to describe.
The form of the name in the parallel 1xx is "double" wrong: it is wrong as a Cyrillic equivalent of the transliterated Russian form in 1xx
(such form does not exist in any language) and also incorrect as an original Ukrainian name (see the usage in 245 $c).

This is only one example but this problem is frequent enough not to be ignored. Such records in Latin transliteration would also not
benefit from a programmatic (or by using a macro) de-transliteration.  It would be a real challenge to come up and agree on a solution to
handle such situations should we recommend adding parallel fields in bibliographic records for headings under authority control.  In the
above example, by adding a 4xx Cyrillic reference to the NAR, we would have recorded the usage in Ukrainian we found on the piece in hand
and facilitate discovery for the users searching in Ukrainian/Cyrillic.


Have a great weekend, everyone!
--Joanna



--

Joanna K. Dyla

Head, Metadata Development Unit

Metadata Department

Stanford University Libraries

650-723-2529

[log in to unmask]




Avetyan, Nora wrote:

Hello everyone,



I think the problem is that there are many records that do not have original scripts alongside with the transliterated form in the bib.
records, thus OCLC auto-supply won?t be able to ?supply? the original script in the authority records. Also, in some cases, several ?ways? of
transliteration forms are used, based on a person?s passionate conviction, and that creates difficulty to generate original scripts from the
bib. records. So, I think Peter your concerns are still valid.

Nora



From: Fletcher, Peter
Sent: Friday, May 08, 2009 10:26 AM
To: 'Joanna Dyla'
Cc: [log in to unmask]; [log in to unmask]; Julie Su; [log in to unmask]; Miller, Caroline; [log in to unmask];
[log in to unmask]; [log in to unmask]; Avetyan, Nora; [log in to unmask]
Subject: RE: PCC Non Latin task force new member; timeline; update



Joanna, I think the discussion is within the scope, but I think we still need to add parallel fields to access points in bibliographic
records for the foreseeable future since many authority records headings don?t have non-Latin 4xx. I?m certainly open to discussing this
topic; perhaps I my concerns aren?t particularly valid now with the OCLC auto-supply of authority record 4xx (which were extracted from
bibliographic record equivalent fields).



Peter







From: Joanna Dyla [mailto:[log in to unmask]]
Sent: Thursday, May 07, 2009 1:34 PM
To: Fletcher, Peter
Cc: [log in to unmask]; [log in to unmask]; Julie Su; [log in to unmask]; Miller, Caroline; [log in to unmask];
[log in to unmask]; [log in to unmask]; Avetyan, Nora; [log in to unmask]
Subject: Re: PCC Non Latin task force new member; timeline; update



It may be a bit too early to ask specific questions but I will be on vacation from May 10th until May 27th (not sure about Internet
access), so let me ask now...

My questions are about a choice of fields in bibliographic records that get parallel non-Latin script data.  When CONSER practice and
documentation on creating records with data in non-Latin scripts was introduced (2001/2003), an option of adding non-Latin script data to
authority records did not exist.  The BIBCO Core Record Standards: 9. Guidelines for Multiple Character Sets document states: "Whether the
non-Latin equivalents of headings are added as references to the authority record or not, they may be supplied in bibliographic records."
Do we really need to continue adding non-Latin parallel controlled fields in bib records as well as in references on name authority
records (i.e., 4XX fields)?  Could we now rely on authority records with non-Latin references instead of on parallel control fields in
bibliographic records for discovery through non-Latin script searching?  Is this topic/discussion within the scope of this TF?

--Joanna


--

Joanna K. Dyla

Head, Metadata Development Unit

Metadata Department

Stanford University Libraries

650-723-2529

[log in to unmask]


Fletcher, Peter wrote:

PCC non-Latin task force,



We have a new member of the task force: Benjamin Abrahamse

Head, Serials Cataloging Section

Cataloging and Metadata Services

MIT Libraries



His area of expertise is Hebrew and Arabic, so along with Nora Avetyan (Persian) I think we have almost all languages represented. I think
we are   light on Greek experience, but Robert mentioned that he has had some experience with Greek. I tend to think that issues that come
up with that language would be similar to those that might arise with Cyrillic since the two are closely related.



Also, if you all agree, I think we could stick with the current timeline: draft report by ALA Annual; final report by December. The draft
report could simply be a progress report with some kind of draft document (even incomplete if necessary), and I assume I would just be
reporting to CONCER/BIBCO at Large during ALA Annual, with some kind of written report I would give to Joan Schuitema, chair of SCS at
that time. Giving this draft report at this time is good in case the membership thinks we should alter our approach or not, so we don?t
waste too much time going in a wrong direction.



Currently I am trying to go through the current CEG appendices (and the short PCC document) to see how we might consolidate/generalize
certain instructions so they could apply to more than one script (as we already discussed), and to try to grasp how far we can go with
that generalization before getting into specific script/language instruction. I should be able to come up with some kind of introductory
draft (not the whole thing! Just the first part?) before too long that we can discuss, pull apart, add other ideas ? . This should be a good
starting point.



Let me know if you have any ideas about this approach.



Best, Peter





Peter Fletcher

Cyrillic Catalog Librarian and Metadata Specialist

[log in to unmask]

Office: (310) 206-3927

Fax: (310) 794-9357

Cataloging & Metadata Center

11020 Kinross Avenue

Box 957230

Los Angeles, CA 90095-7230