Print

Print


On 1/10/18 2:19 PM, Joseph Kiegel wrote:
<snip>
> 
> 
> - mixed language strings - which language are they given, if any, and why?
> 
> As a working hypothesis, I suggest using “For mixed language strings, apply a language tag for the predominant language, and if a predominant language is not readily apparent, omit a language tag”.  This can be tested against real examples and refined as needed.
> 
> It is important to think of this in native BIBFRAME.  For example, some of the problems with Kelley McGrath’s example fall away when you encode in BIBFRAME properties:
> 
> bf:mainTitle “Le francais pour moi”@fr
> bf:subtitle “learning a second language”@en
> 
> 
> 
> - what is the relationship between language of title and language of the text of the resource, if any? (cf. "Quo Vadis?")
> 
> I’m not sure what the issue is.  RDA has an element for the language of the work.  The language of the work can be, e.g., Polish, while the language tag of the title itself is Latin.

This is where the questions about discovery and display become
important. See below.

> - does this affect display in any way?
> 
> Specific discovery interfaces have not yet been built, so we don’t have details.  But yes, if interfaces will offer users a choice of the language of display, then language tags will be important if not critical for success.  Experiments with the internationalized VIAF show that you can’t just pluck a label out of the pot:  when your audience is English speakers, you probably don’t want Chinese display.

... when your audience is English speakers, you probably don't want
Latin display? Well, with the example of Quo Vadis it wouldn't be very
useful to suppress the display of the title because the title isn't in
English. If you aren't using the language to determine what to display,
then what is the purpose of coding the language? If a user types "quo
vadis" into a search box, do they have to say what language it is they
are searching?

I'm pushing back on the assumption that having each string coded for
language is a necessity because I want a rational reason for adding this
labor to the already complex task of cataloging. This skepticism comes
out of my many years of processing bibliographic data that had
underlying assumptions about needs that weren't borne out in practice.
I'm not against coding for language, but I am against doing so without
thinking through what it means for catalog users.

kc

> 
> 
> - does this affect indexing in any way?
> 
> Indexing is a complex technical issue and someone with expertise in it will have to say.
> 
> 
> - does this affect search in any way?
> 
> An obvious application for language tags is post-search faceting.  Our discovery interface is not accurate because it uses subfields from the 041 rather indiscriminately.  For example, a search for films that is limited to French includes those where French is a language of the subtitles rather the film itself.  Language tags may provide more accuracy with facets.
> 
> 
> 
> 
> 
> -----Original Message-----
> From: Bibliographic Framework Transition Initiative Forum [mailto:[log in to unmask]] On Behalf Of Karen Coyle
> Sent: Saturday, January 06, 2018 10:45 AM
> To: [log in to unmask]
> Subject: Re: [BIBFRAME] CC:AAM Statement in Support of the Internationalization of BIBFRAME
> 
> Thanks, Amanda. This is a good example (your BCP47 example) of the complexity.
> 
>         [ a bf:Title;
> 
>             rdfs:label "西遊記 = Journey to the west"@zh-cmn-hant;
> 
>             bflc:titleSortKey "西遊記 = Journey to the west";
> 
>             bf:mainTitle "西遊記"@zh-cmn-hant;
> 
>             bf:subtitle "Journey to the west"@en-us ] .
> 
> shows that mixed language strings (rdfs:label, bflc:titleSortKey) present problems. (One has a language tag, the other does not -
> intentional?) I would assume that BCP47 would mainly be used by specialist libraries or for special collections.
> 
> Perhaps I should offer my questions in list form, and we can tick them off?
> 
> - 2-char or 3-char codes, or BCP47? Which are used, and under what circumstances?
> - mixed language strings - which language are they given, if any, and why?
> - what is the relationship between language of title and language of the text of the resource, if any? (cf. "Quo Vadis?"[1])
> - who is responsible for the rules that govern decisions? RDA group?
> PCC? LoC BIBFRAME?
> - does this affect display in any way?
> - does this affect indexing in any way?
> - does this affect search in any way?
> 
> I think that's it, along with the obvious need for use cases.
> 
> kc
> [1]
> http://www.worldcat.org/search?qt=worldcat_org_bks&q=quo+vadis&fq=dt%3Abks
> 
> (Also, as a note, one can only see the full OCLC record when accessing from a member organization. The rest of us see only a fairly reduced record and never the MARC fields. In addition, given that RDA is behind a paywall, that also isn't available to those outside of a subscribing institution. This affects not only us renegade retirees, but also many librarians in libraries who cannot afford these services. This "have/have not" is not, IMO, good for the library world in general.
> Silos may be necessary, but they always create a barrier.)
> 
> On 1/5/18 3:23 PM, Xu, Amanda wrote:
>> Hi Karen,
>>
>>  
>>
>> Thank you so much for these wonderful questions.  According to W3C 
>> Recommendation 25 February 2014, RDF 1.1 Concepts and Abstract Syntax, 
>> a literal in an RDF graph consists of two or three elements.  If the 
>> third element is present, a literal is a language-tagged string.
>> Lexical representations of language tags may be converted to lower case.
>> The value space of language tags is always in lower case.  The 
>> language tag must be well-formed according to section 2.2.9 of BCP47, 
>> available from https://tools.ietf.org/html/bcp47.  You can find the 
>> language codes from 
>> https://www.iana.org/assignments/language-subtag-registry/language-sub
>> tag-registry
>>
>>
>>  
>>
>> For compliance, we may consider the adoption of BCP47 language tags. 
>> However, I agree with you that we must build good use cases for the 
>> coding change given the complexity of our data.  I also agree with Joe 
>> Kiegek that in native BIBFRAME and a good user interface, assigning 
>> language tags may not be difficult or time consuming.  In addition, I 
>> am hoping that the next version of MARC2BIBFRAME converter can handle 
>> multiscript record conversion better with the use of BCP47 language 
>> tags if an agreement can be reached by PCC or some such group.
>>
>>  
>>
>> One experiment that I did might be the starting point for us to 
>> collect sample data for use case development.  You can check OCLC#122820377 .
>> It is not a RDA record and relator codes are missing.  But we may list 
>> it as an example for a multiscript record.   The transcribed title and 
>> subtitle are in different language scripts.  Author/title groups, 
>> personal names, TOC, etc. are in different language scripts.
>>
>>  
>>
>> _Paired field for MARC 245 title field in OCLC_:
>>
>>  
>>
>>  
>> imap:[log in to unmask]:143/fetch%3EUID%3E.INBOX.BIBFRAME%3E2922?
>> header=quotebody&part=1.1.2&filename=image001.png
>>
>>  
>>
>> _Titles with language tags using BCP47 in BIBFRAME description_:
>>
>>  
>>
>> <http://example.org/ocn122820377#Work> a bf:MovingImage,
>>
>>         bf:Work;
>>
>>     rdfs:label "Xi you ji"@pinyin;
>>
>>  
>>
>> bf:contribution
>>
>> [ a bf:Contribution;
>>
>>                                 bf:agent 
>> <http://example.org/ocn122820377#Agent700-31>;
>>
>>                                 bf:role 
>> <http://id.loc.gov/vocabulary/relators/ctb> ],
>>
>> [ a bf:Contribution;
>>
>>                                bf:agent 
>> <http://example.org/ocn122820377#Agent880-44>;
>>
>>                                bf:role 
>> <http://id.loc.gov/vocabulary/relators/ctb> ];
>>
>>  
>>
>> bf:title [ a bf:Title;
>>
>>             rdfs:label "Xi you ji"@pinyin;
>>
>>             bflc:titleSortKey "Xi you ji";
>>
>>             bf:mainTitle "Xi you ji"@pinyin ],
>>
>>         [ a bf:Title;
>>
>>             rdfs:label "西遊記 = Journey to the west"@zh-cmn-hant;
>>
>>             bflc:titleSortKey "西遊記 = Journey to the west";
>>
>>             bf:mainTitle "西遊記"@zh-cmn-hant;
>>
>>             bf:subtitle "Journey to the west"@en-us ] .
>>
>>  
>>
>>  
>>
>> <http://example.org/ocn122820377#Agent700-31> a bf:Agent,
>>
>>         bf:Person;
>>
>>     rdfs:label "Yang, Jie"@pinyin;
>>
>>     bflc:name00MarcKey "7001 $6880-04$aYang, Jie";
>>
>>     bflc:name00MatchKey "Yang, Jie" .
>>
>>  
>>
>> <http://example.org/ocn122820377#Agent880-44> a bf:Agent,
>>
>>         bf:Person;
>>
>>     rdfs:label "杨洁"@zh-cmn-hant;
>>
>>     bflc:name00MarcKey "8801 $6700-04/$1$a杨洁";
>>
>>     bflc:name00MatchKey "杨洁" .
>>
>>  
>>
>> Attached is the entire record in .ttl format.  Thanks a lot!
>>
>>  
>>
>> Amanda
>>
>>  
>>
>>  
>>
>> ---
>>
>> Amanda Xu
>>
>> Metadata Analyst Librarian
>>
>> Cataloging and Metadata Department
>>
>> University of Iowa Libraries
>>
>> 100 Main Library (LIB)
>>
>> Iowa City, IA 52242-1420
>>
>>  
>>
>>  
>>
>>  
>>
>> -----Original Message-----
>> From: Bibliographic Framework Transition Initiative Forum 
>> [mailto:[log in to unmask]] On Behalf Of Karen Coyle
>> Sent: Friday, January 05, 2018 2:24 PM
>> To: [log in to unmask]
>> Subject: Re: [BIBFRAME] CC:AAM Statement in Support of the 
>> Internationalization of BIBFRAME
>>
>>  
>>
>> On 1/2/18 9:14 AM, Joseph Kiegel wrote:
>>
>>> There will be edge cases that are difficult, but for the vast 
>>> majority
>> of strings, the language will be obvious to the cataloger.
>>
>>>
>>
>>> In native BIBFRAME and a good user interface, assigning language tags
>> will not be difficult or time consuming.  The language of cataloging 
>> is known and those fields can be tagged automatically.  Templates can 
>> assign tags for catalogers who routinely catalog in a given language.  
>> I have experimented with language tags in a test interface and it was 
>> not hard.
>>
>>  
>>
>> This is the kind of statement which makes me hunger for more detail. 
>> For example, what were the rules for assignment for: transcribed titles?
>>
>> titles with subtitles in different languages? author/title groups (if 
>> they exist in BF - I don't remember the structuring of those)? 
>> personal names? Are there strings with more than one language and how 
>> is that handled? Can a title ever be in a different language than the 
>> language of the text when the text is monolingual?
>>
>>  
>>
>> Also, do we have or is anyone developing rules or guidelines for 
>> cataloging decisions regarding language tagging of individual strings?
>>
>> (This would seem to fall to PCC or some such group?)  Is this covered 
>> in RDA anywhere? What standard are we using? ISO 639-1, -2, or BCP 47?
>>
>>  
>>
>> But above all I have yet to read anything that addresses the use cases 
>> where such encoding facilitates or is essential for user services. We 
>> have long had the separate of subject access by language (O Canada!), 
>> and the selection of language materials by language. But I haven't 
>> seem a non-speculative, practical use for language tagging of strings. 
>> I realize that language tagging of strings is coming to us from RDF, 
>> and is somewhat new, and may in the future be obligatory, but I still 
>> think we need use cases before undertaking coding so that said coding 
>> will provide the desired outcomes, given the complexity of our data.
>>
>>  
>>
>> Perhaps what this amounts to is a knowledge gap between the BF 
>> practitioners and those of us who are on the sidelines. If so, please 
>> point us to the relevant documentation!
>>
>>  
>>
>> thanks,
>>
>> kc
>>
>>  
>>
>>>
>>
>>>
>>
>>>
>>
>>> -----Original Message-----
>>
>>> From: Bibliographic Framework Transition Initiative Forum
>>
>>> [mailto:[log in to unmask]] On Behalf Of Karen Coyle
>>
>>> Sent: Friday, December 22, 2017 6:42 AM
>>
>>> To: [log in to unmask] <mailto:[log in to unmask]>
>>
>>> Subject: Re: [BIBFRAME] CC:AAM Statement in Support of the
>>
>>> Internationalization of BIBFRAME
>>
>>>
>>
>>> Osma, I took all of those examples of 1984 from LoC's catalog. While
>> Wikidata may think they have different titles, we don't know how that 
>> decision was made (there are no cataloging rules for Wikidata). In no 
>> case have I seen "Nineteen Eighty-Four" for the English version 
>> (although it was filed that way in card catalogs as per the ALA Filing 
>> Rules). Your examples all conveniently prove your point, but I still 
>> think that asking catalogers to determine the language of every field 
>> is going to create difficulties. It would be a good idea to take a 
>> sampling of records and try this out. From the cataloger's point of view.
>>
>>>
>>
>>> kc
>>
>>>
>>
>>> On 12/21/17 7:44 AM, Osma Suominen wrote:
>>
>>>>> However, there is a big problem with trying to attribute
>>
>>>>> *language* to fields in bibliographic data. It only takes a few
>>
>>>>> examples to understand why:
>>
>>>>>  
>>
>>>>> Title:
>>
>>>>> 1984 (book in German)
>>
>>>>> 1984 (book in Hebrew)
>>
>>>>> 1984 (book in English)
>>
>>>>  
>>
>>>> I don't think that's a problem at all. In fact this is a great
>>
>>>> example, since the name of Orwell's novel (assuming you meant it)
>>
>>>> actually differs between many languages. According to Wikidata
>>
>>>> (http://www.wikidata.org/entity/Q208460) it is called
>>
>>>>  
>>
>>>> "1984" in German
>>
>>>> "1984" in Hebrew (but rendered with right-to-left alignment!)
>>
>>>> "Nineteen Eighty-Four" in English (not 1984!) "Vuonna 1984" in
>>
>>>> Finnish "নাইন্টিন এইটি-ফোর" in Bengali
>>
>>>
>>
>>> --
>>
>>> Karen Coyle
>>
>>> [log in to unmask] <mailto:[log in to unmask]> http://kcoyle.net
>>
>>> m: +1-510-435-8234
>>
>>> skype: kcoylenet/+1-510-984-3600
>>
>>>
>>
>>  
>>
>> --
>>
>> Karen Coyle
>>
>> [log in to unmask] <mailto:[log in to unmask]> http://kcoyle.net
>>
>> m: +1-510-435-8234
>>
>> skype: kcoylenet/+1-510-984-3600
>>
> 
> --
> Karen Coyle
> [log in to unmask] http://kcoyle.net
> m: +1-510-435-8234
> skype: kcoylenet/+1-510-984-3600
> 

-- 
Karen Coyle
[log in to unmask] http://kcoyle.net
m: +1-510-435-8234
skype: kcoylenet/+1-510-984-3600