Re: [arks] Inflections vs Linked Open Data

[John A. Kunze]

The implementation of the ? and ?? inflections has been pretty spotty,
and that's one of the reasons for this arks-forum discussion group.

Contributing factors I think include (a) the metadata to return is very
general and a bit ambiguous and (b) the technical barrier of hooking up
these non-standard, vestigial "URL query strings" to metdata databases,
eg, Tomcat servers have a bug that doesn't permit a single '?' ending a
URL to be detected. I'm interested to hear other views as well.

As for Linked Open Data (LOD), I like to think that ARKs promoted the
concept before the LOD acronym existed. Lots of projects produce LOD and
use content-negotiation, and where it makes sense, it can be strategic
for implementors of ARKs to support both, eg, to satisfy the terms of a
grant. While there's lots of momentum and buzz behind LOD, some people
are unclear as to what practical problems it solves. Unlike with content
negotiation, inflections give users a standard way to request metadata
without programmer intervention, and to ask it of RDF-based objects.

Fortunately, however, it doesn't have to be an either/or choice. If you
already support inflections or content negotiation -- the hard part being
to maintain metadata and hook it up to web-based queries -- it's almost
trivial to do both. As an aside, work we are doing with the N2T resolver
stands a good chance of making it easy for any ARK implementor (even with
Tomcat) to support via n2t.net both inflections and content negotiation.
I could see the ARK spec being modified to encourage content negotiation
in addition to (but not replacing) inflections. I also see the number of
inflections being extended to permit optional requests for things like an
object's landing page or change history.

-John

--- On Wed, 24 Oct 2012, pbinkley wrote:
> We're looking at ARKs as a possible improvement to our current local
> identifiers, and I'm trying to get a sense of how they're being used. One
> question is the use of inflections (the urls with ? and ?? at the end to
> fetch ERC records). It appears that many major ARK users haven't implemented
> this. I'd be interested in hearing whether others are seeing use of this
> feature: are these records being harvested or viewed? 
> And is there discussion of developing a new version of the ARK spec that
> would replace inflections with linked open data? There's nothing to prevent
> us offering LOD from ARK urls (as BnF is doing, I believe), but it would
> seem to be a more modern fit for the problem of offering summary machine-
> and human-readable metadata, and in particular to express the maintenance
> commitment in a standard way.
>
> Thanks to anyone who cares to take the time to help educate me on this
> stuff!
>
> Peter
>
> Digital Initiatives Technology Librarian
> University of Alberta

[Sébastien Peyrard]

Hi all,

Sorry for the late answer.

Concerning the ? and ?? question marks, we don't implement them at BnF for the reasons John indicated, but also because it raised a red flag in our minds since question marks are already intensively used  in HTTP, especially for parameters, so it kind of felt like an exception to that practice.
At BnF, we decided to implement them with another ark mechanism: specific qualifiers.
We implemented "??" with a ".policy" qualifier, e.g. the document http://gallica.bnf.fr/ark:/12148/bpt6k70861t
then the policy for the identifier (similarity policy)
http://gallica.bnf.fr/ark:/12148/bpt6k70861t.policy
(apologies, the policy statement is in French :) we intend to precise it at some point in the future)

The "?" qualifier was intended to be replaced with another qualifier, .description, although it has not been implemented yet. Even though nothing is carved in stone yet, we might use our own Dublin Core metadata rather than ERC concerning that (so as not to maintain another mapping).

Hope this helps!

Best wishes,

Sébastien

2012/10/25 John A. Kunze <j...@ucop.edu>

--
You received this message because you are subscribed to the Google Groups "ARKs" group.

To post to this group, send email to arks-...@googlegroups.com.
To unsubscribe from this group, send email to arks-forum+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/arks-forum?hl=en.

[Rick Leir]

Here is my translation of the BNF policy (my French is far from perfect, and Google Translate was a big help.) (HTH, feel free to use my translation):

The BnF (National Library of France) assigns identifiers containing the ARK registered NAAN (Name Assigning Authority Number) 12148 according to the following principles:

* No ARK identifier will be reassigned; that is to say that once a link between an ARK identifier and an object has been published, this link will be considered unique, for an indefinite period.

* ARK identifiers assigned by the BnF do not contain, to the extent possible, any recognizable semantic information; This helps facilitate their use independently of any context of time or place.

* ARK identifiers assigned by the BnF contain a checksum character that guarantees against typographical single character errors and transposition errors.

* The BnF guarantees the perpetual uniqueness of the ARK assigned identifier (but the precise extent of guarantees is being defined). Indeed, the management of a digital object may require appropriate modifications of this object to ensure indefinite persistence. Moreover, the reported level of guarantees will evolve as the standards and sustainability practices become better controlled. The BNF will implement procedures to maintain these guarantees and associate them with objects it manages.

* Some of the digital collections of the BnF duplicate collections from other institutions.

To unsubscribe from this group, send email to arks-forum+...@googlegroups.com.

[John Kunze]

Thanks for sharing this, Rick.  It's nice to see such a clear and concise policy.

On Tue, Mar 31, 2015 at 8:43 AM, Rick Leir <ric...@c7a.ca> wrote:

Here is my translation of the BNF policy (my French is far from perfect, and Google Translate was a big help.) (HTH, feel free to use my translation):

The BnF (National Library of France) assigns identifiers containing the ARK registered NAAN (Name Assigning Authority Number) 12148 according to the following principles:

* No ARK identifier will be reassigned; that is to say that once a link between an ARK identifier and an object has been published, this link will be considered unique, for an indefinite period.

* ARK identifiers assigned by the BnF do not contain, to the extent possible, any recognizable semantic information; This helps facilitate their use independently of any context of time or place.

* ARK identifiers assigned by the BnF contain a checksum character that guarantees against typographical single character errors and transposition errors.

* The BnF guarantees the perpetual uniqueness of the ARK assigned identifier (but the precise extent of guarantees is being defined). Indeed, the management of a digital object may require appropriate modifications of this object to ensure indefinite persistence. Moreover, the reported level of guarantees will evolve as the standards and sustainability practices become better controlled. The BNF will implement procedures to maintain these guarantees and associate them with objects it manages.

* Some of the digital collections of the BnF duplicate collections from other institutions.

On Friday, November 2, 2012 at 11:34:05 PM UTC-4, Sébastien Peyrard wrote:

To unsubscribe from this group, send email to arks-forum+...@googlegroups.com.

[Peter Murray]

I agree -- thanks for the concise bullet points.  Does the last bullet imply that BnF may assign ARKs to objects that are not its own?

Peter

On Mar 31, 2015, at 5:34 PM, John Kunze <j...@ucop.edu> wrote:

Thanks for sharing this, Rick.  It's nice to see such a clear and concise policy.

On Tue, Mar 31, 2015 at 8:43 AM, Rick Leir <ric...@c7a.ca> wrote:

Here is my translation of the BNF policy (my French is far from perfect, and Google Translate was a big help.) (HTH, feel free to use my translation):

The BnF (National Library of France) assigns identifiers containing the ARK registered NAAN (Name Assigning Authority Number) 12148 according to the following principles:

* No ARK identifier will be reassigned; that is to say that once a link between an ARK identifier and an object has been published, this link will be considered unique, for an indefinite period.

* ARK identifiers assigned by the BnF do not contain, to the extent possible, any recognizable semantic information; This helps facilitate their use independently of any context of time or place.

* ARK identifiers assigned by the BnF contain a checksum character that guarantees against typographical single character errors and transposition errors.

* The BnF guarantees the perpetual uniqueness of the ARK assigned identifier (but the precise extent of guarantees is being defined). Indeed, the management of a digital object may require appropriate modifications of this object to ensure indefinite persistence. Moreover, the reported level of guarantees will evolve as the standards and sustainability practices become better controlled. The BNF will implement procedures to maintain these guarantees and associate them with objects it manages.

* Some of the digital collections of the BnF duplicate collections from other institutions.

--
Peter Murray
Assistant Director, Technology Services Development
LYRASIS
Peter....@lyrasis.org
+1 678-235-2955
800.999.8558 x2955

[John Kunze]

* Some of the digital collections of the BnF duplicate collections from other institutions.

I take it to mean that the BnF, like most libraries and aggregators, provide access to collections that did not originate entirely within their library.

I also do not speak for the BnF, but in my library some such collections arrive without globally unique identifiers, and so we assign our own identifiers to those objects.

On Tue, Mar 31, 2015 at 3:16 PM, Peter Murray <peter....@lyrasis.org> wrote:

I agree -- thanks for the concise bullet points.  Does the last bullet imply that BnF may assign ARKs to objects that are not its own?

Peter

On Mar 31, 2015, at 5:34 PM, John Kunze <j...@ucop.edu> wrote:

Thanks for sharing this, Rick.  It's nice to see such a clear and concise policy.

On Tue, Mar 31, 2015 at 8:43 AM, Rick Leir <ric...@c7a.ca> wrote:

Here is my translation of the BNF policy (my French is far from perfect, and Google Translate was a big help.) (HTH, feel free to use my translation):

The BnF (National Library of France) assigns identifiers containing the ARK registered NAAN (Name Assigning Authority Number) 12148 according to the following principles:

* No ARK identifier will be reassigned; that is to say that once a link between an ARK identifier and an object has been published, this link will be considered unique, for an indefinite period.

* ARK identifiers assigned by the BnF do not contain, to the extent possible, any recognizable semantic information; This helps facilitate their use independently of any context of time or place.

* ARK identifiers assigned by the BnF contain a checksum character that guarantees against typographical single character errors and transposition errors.

* The BnF guarantees the perpetual uniqueness of the ARK assigned identifier (but the precise extent of guarantees is being defined). Indeed, the management of a digital object may require appropriate modifications of this object to ensure indefinite persistence. Moreover, the reported level of guarantees will evolve as the standards and sustainability practices become better controlled. The BNF will implement procedures to maintain these guarantees and associate them with objects it manages.

* Some of the digital collections of the BnF duplicate collections from other institutions.

--
Peter Murray
Assistant Director, Technology Services Development
LYRASIS
Peter....@lyrasis.org
+1 678-235-2955
800.999.8558 x2955

--
You received this message because you are subscribed to the Google Groups "ARKs" group.

To unsubscribe from this group and stop receiving emails from it, send an email to arks-forum+...@googlegroups.com.

To post to this group, send email to arks-...@googlegroups.com.

Visit this group at http://groups.google.com/group/arks-forum.
For more options, visit https://groups.google.com/d/optout.

[Sébastien Peyrard]

Dear Rick,

Thank you so much for your interest and for translating the BnF policy!
This is really encouraging to us :-)

Realizing the interest that our policy raises on this list, we will definitely put a translated version alongside the French one further down the road, and your version will be really helpful.

Pete, concerning the question about digital collections, here are some elements:

* In the policy, we were mentioning the case where alternate digitizations are available on the web for the same edition. At BnF, we are identifying with an ARK id the digital item that we own. Other digitizations of the same book (edition) would have a different ARK id identifies our own asset.

* As John mentioned, we do digitize physical items that originate from other French institutions, mainly libraries. In such cases, we are co-owners (with the originating institution) of the digital version of the document. In such cases we have an ARK for the digitized item because we have a persistence policy on it.

* In some cases however, we assign ARKs to digital items that we do not own. This is the case of third party archiving, where we assign an ARK for information packages that we will have to return to the customers when/if the archiving contract is over. However, the corresponding minted ARK ids will be "reserved" and will NOT be re-assigned (in compliance with our policy), even though the digital objects will not be under our stewardship

Hope this helps. Feel free to ask away if you have additional questions or if things are unclear!

Kind regards,

Sébastien Peyrard and Jean-Philippe Tramoni - BnF

[Peter Murray]

Thank you, Sébastien and John!  I appreciate the clarification.

Peter