Re: [urn] Draft Agenda for URNBIS session at IETF 80

[Juha Hakala <[e-mail]>]

Hello,

Peter Saint-Andre wrote:
> <hat type='individual'/>
> 
> On 3/23/11 9:05 AM, Ted Hardie wrote:
>> Hi Juha,
>>
>> Thanks for providing the use case.  What is not clear (to me, at
>> least) is why they cannot provide identifiers for structured portions
>> of larger resources without re-using the fragment identifier syntax.
>>
>> urn:example-nid:large-resource-id:contained-resource-id
>>
>> seems to cover this ground; you can even elaborate it to:
>>
>> urn:example-nid:large-resource-id:container-type:contained-resource-id
>>
>> No one objects to identifying resources  contained within other
>> resources, at least  as far as I know.  The problem is the syntax
>> re-use with the fragment identifier in URIs, which is understood in
>> relation to the media type, rather than more generally.
> 
> Yes, that seems like a reasonable approach. I'm still not quite sure why
> folks think they need to use the fragment identifier syntax here.

I hope Marjan or other Dutch colleagues can explain their situation in 
more details.

Generally, there seems to be at least three approaches:

1. <fragment> is used according to the requirements of URI syntax

2. URNs with no implicit or explicit (URI syntax-based) fragment 
specification are given to component parts of a resource. This is the 
approach chosen by my library when we give identify for instance images 
or chapters within a book using URN:NBN. One (the only?) caveat with 
this is that some standard-based namespaces such as URN:ISBN may not 
allow assignment of identifiers to fragments.

3. URNs with implicit (namespace-based) fragment specification are used. 
The syntax and semantics of this activity should be detailed in the 
namespace registration and based on the identifier standard used. An 
example of this is SICI which identifies e.g. issues and articles in 
journals (see 
http://en.wikipedia.org/wiki/Serial_Item_and_Contribution_Identifier).

SICI is based on ISSN 
(http://en.wikipedia.org/wiki/International_Standard_Serial_Number) but 
but its syntax is a lot more complex since it covers serial "fragments". 
  However, usage of URI syntax <fragment> is not needed and might even 
be counterproductive.

Depending on how the system holding these resources have been built, 
journal articles and issues may or may not be fragments in the URI 
syntax sense of the word. Mapping from URN:SICI to URLs would deal with 
this.

If this approach is acceptable, then RFC2141bis may incorporate 
<fragment> but the RFC should make it clear that a) <fragment> should be 
used only when the media type allows it, b) depending on the URN 
namespace, there may be other approaches for identification of fragments 
which are not limited by the stipulations of the URI syntax.

Best regards,

Juha

> Peter
> 

-- 

  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 [e-mail], tel [phone]