Re: Maintenance of the URN namespace registrations

Hello Martin,

In order to give a proper understanding of what manifestation means, I 
need to start a little bit further away.

Library community has develop a model (Functional Requirements for 
Bibliographic Records, FRBR) which has four levels:

1. Work, such as Joyce's Ulysses
2. Expression, for instance Finnish translation of Ulysses
3. Manifestation, e.g. paperback version of the Finnish translation of 
Ulysses, or printed version or Web version of Scientific American.
4. Item, a copy of the aforementioned manifestation of Ulysses or set of 
Scientific American volumes in my book shelf.

Traditional library identifiers such as ISBN and ISSN deal with 
manifestations. Usually every one of them, for instance hardcover book, 
paperback book and e-book should get a different ISBN. Whenever an 
e-book is migrated to guarantee its usability, the new version should 
get a new ISBN. These practices are spesified in the identifier 
standards (but are not always followed properly by those responsible of 
the identifier assignment). URN inherits these community practices from 
the namespace level.

There are standard identifiers for works, such as International Standard 
Text Code (ISTC) for textual works. ISTC and its peers (ISAN, ISWC) do 
not have URN namespaces yet.

 From the national library's point of view, eventually (after 200 years 
of storage, for instance) there will be a lot of digital manifestations 
for each (born digital or digitized) work, created via successive 
migrations within the library's long term preservation system (old 
versions will not be deleted from the system). These manifestations will 
ideally be linked directly to one another and to the work level record, 
if one exists. Some manifestations may still be readable, while some 
others may not / will not be.

Any persistent identifier system to be used in this environment must 
support a broad set of services. For instance, if the URN belongs to a 
work, it must be possible to retrieve bibliographic data related to the 
manifestations of this work, or data related to a specific manifestation 
(for instance the first or the latest) of the work, or bibliographic 
data about items related to given / all manifestations.

Best regards,

Juha

Martin J. Dürst wrote:
> Hello Juha,
> 
> I was almost able to follow your explanations, but it would help me (and 
> maybe others) a lot if you could give some concrete examples of 
> "manifestation of the journal".
> 
> Regards,   Martin.
> 
> On 2010/10/26 21:57, Juha Hakala wrote:
>> Hello all,
>>
>> Library & publisher communities have registered a number of URN
>> namespaces over the years, including namespaces ISBN and ISSN for the
>> well-known international standard identifiers for books and serials.
>>
>> International standards do not change too often in fundamental ways, but
>> by coincidence both the new ISBN used since 2007 and the new ISSN
>> version released in 2007 deviated from their predecessors in such a way
>> that the need to revise RFCs registering their namespaces arose - but in
>> different ways.
>>
>> With ISBN the function of the identifier is still the same, but the
>> syntax of the identifier has changed. The old ISBN had 10 digits, the
>> new one 13, starting with prefix 978 or 979. This has an impact on how
>> the ISBNs are resolved as URNs. The Internet draft written by Maarit
>> Huttunen, Alfred Hoenes and myself (see
>> http://tools.ietf.org/html/draft-hakala-rfc3187bis-isbn-urn-00)
>> describes how to deal with ISBN-13.
>>
>> With ISSN the syntax of the identifier is still the same, but the ISSN
>> functions have changed. Traditional ISSN, which identifies a single
>> manifestation of a journal, such as e-version of Scientific American, is
>> still there. But the standard defines also something new called Linking
>> ISSN. This new identifier will bring together all manifestations of a
>> journal.
>>
>> Syntactically, ISSN-L and ISSN are identical. In practice, ISSN of the
>> first manifestation of the journal has a dual role; it is also used as
>> the ISSN-L.
>>
>> In library systems, ISSN and ISSN-L reside in the same bibliographic
>> record, but they are stored in different data elements. ISSN-L appears
>> in every manifestation of the journal, alongside the ISSN belonging to
>> that particular manifestation.
>>
>> Although ISSN and ISSN-L of the journal are identical, but in library
>> system they can be indexed separately (being stored in different
>> metadata elements). Thus the library system knows that the two identical
>> numbers have different semantics. When searched as ISSN the identifier
>> will only retrieve the eldest manifestation of the serial. When searched
>> as ISSN-L the user should be able to find all manifestations.
>>
>> In URN resolution, something has to make it obvious to the URN resolver
>> whether the user wants to resolve the number as ISSN or ISSN-L. Given
>> the current URN architecture, one way to accomplish this is to register
>> two namespaces, ISSN and ISSN-L. Then, in order to resolve 1234-5678 as
>> ISSN URN:ISSN will do, and to resolve it has ISSN-L, URN:ISSN-L would
>> do. The former query would yield bibliographic metadata about the
>> relevant manifestation or its URL, the latter bibliographic metadata
>> about all manifestations. There can be several of them, since journals
>> increasingly have both printed and e-versions.
>>
>> Revision of RFC 3044 has not yet started, but Persid project has been in
>> contact with the ISSN International Centre on how, when and by whom such
>> revision can be accomplished.
>>
>> Both ISSN and ISBN were developed in 1970s. These recent revisions were
>> the first ones that do require URN namespace registrations. All previous
>> edits were minor and did not change syntax or the functionality of the
>> identifier in a fundamental manner. Thus it would not be right to draw a
>> conclusion that identifier standards are in general volatile and that
>> this will force us into changing namespace registrations every now and
>> then. More likely, the revisions having implications to the namespace
>> registrations will be rare events - unless the practice of resolving the
>> URNs in the namespace in question shows that the namespace registration
>> has to be modified in some manner. This is what happened with the NBN
>> (national bibliography number). When the RFC was written, there were no
>> NBN-based URNs, while now there are millions of them. Practical
>> experienced gained have been used in the preparation of the revised
>> namespace registration, available at
>> http://www.ietf.org/id/draft-hakala-rfc3188bis-nbn-urn-00.txt.
>>
>> Best regards,
>>
>> Juha Hakala
> 

--

-- 

  Juha Hakala
  Senior advisor, standardisation and IT

  The National Library of Finland
  P.O.Box 15 (Unioninkatu 36, room 503), FIN-00014 Helsinki University
  Email juha.hakala <at> helsinki.fi, tel +358 50 382 7678