[Web4lib] Can a DOI identify a web site?

[Grace J. Agnew]

It's important to understand the intent of a permanent identifier, such 
as a DOI.  The identifier is intended to be a permanent surrogate and 
pointer to a resource that may have many volatile aspects--generally the 
physical location is presumed to be volatile, and thus the identifier 
points to metadata in a directory that "resolves" to the current 
location of the resource. That metadata can also provide provenance 
information, indicate what has changed about the resource, etc. It is 
expected that the identifier will be globally unique and so durable that 
it can outlast the resource, and then ideally point to historical 
metadata about the defunct resource.

 Different identifier systems offer different strategies.  DOI uses the 
<indecs>rdd metadata data dictionary and the ontologyX data model to 
provide a reference metadata implementation within the identifier 
directory database.  DOI requires that support for a core metadata 
implementation and encourages DOI users to also map their metadata to 
the DOI metadata schema.  For other identifier systems, which don't 
include a rich metadata strategy but generally just resolution to the 
current physical location of the resource, you might want to assign the 
identifier to a package that bundles the metadata, structure map and 
resource.  Since we use METS, which encapsulates the resource into the 
metadata, we assign our handle to the METS document (metadata, structure 
map, resource datastreams, and behaviors.  Other systems that provide 
this packaging or bundling include MPEG-21 and MXF.

Grace Agnew

Thomas Bennett wrote:
>   A WEB  site
>
>
> On Wednesday 28 March 2007 23:43, christina struik wrote:
>   
>> Hello,
>>
>> I have reading about DOIs and I just cannot figure out if a web site counts
>> as an "object" or if pages within the site must have separate DOIs.
>>     
> From the little I have read on DOI,  "is for identifying content objects in 
> the digital environment".  I would think a WEB site is a container for the 
> objects not an object itself.
>
>
>   
>> Does the (usually changing) web site content make it difficult to link to
>> content?
>>     
> Since their is not a standard on how a WEB server program identifies its 
> content as objects it is difficult.
>
>   
>> Can a web site have a DOI that relies solely on metadata so that when the
>> site changes URL the DOI follows the site?
>>     
> Depends on your server software and settings sort of.  All content in Zope 
> server is in an Object Database with possible exceptions of database adapters 
> to other databases and products that allow direct file system storage.  A 
> zope/plone WEB site has OIDs (Object IDs) for every item: images, folders, 
> documents, scripts, database connectors, etc.  In plone using KUPU (a wysiwyg 
> html editor)  in its configuraiton for the site you can check 
>
> " Link using UIDs 
> Links to objects on this site created by Kupu can use unique object ids so 
> that the links remain valid even if the target object is renamed or moved 
> elsewhere on the site."
>
> Now if there were a DOI to OID table built into Plone, this would work even if 
> the page were moved.  Also would probably need a DOI harvester to keep the 
> table updated for new material.
>   
>> I am doing my degree in library and information studies and I have seen web
>> documents linked by DOI but have not yet come across DOI-linked web sites
>> when they are included in a library catalog.  It would be useful to have
>> more stability in catalog web site links.
>>     
>
> That would be a DOI itself pointing to a container if you mean a catalog entry 
> pointing to a WEB site not a DOI in a WEB site.
>   
>> Thank-you for your time,
>>     
> You're welcome.
>   
>> Christina Struik
>> _______________________________________________
>> Web4lib mailing list
>> Web4lib at webjunction.org
>> http://lists.webjunction.org/web4lib/
>>     
>
> Or have I totally misunderstood the DOI?
>
>