Hi Jonathan,

> So in our marc records, we have these 856 links, the meaning of which is
> basically "some web page related to the entity at hand." You don't
> really know the relation, the granularity is not there.

There is some *minimal* indication of the relationship via the second indicator of the 856 (and subfield $3, for a related resource) [1]:

  Second Indicator - Relationship

	Relationship between the electronic resource at the location specified in field 856 
	and the item described in the record as a whole.

	Used to provide further information about the relationship if it is not a one-to-one relationship.

		# - No information provided

		0 - Resource

			Electronic location in field 856 is for the same resource described by the record as 
			a whole. In this case, the item represented by the bibliographic record is an 
			electronic resource. If the data in field 856 relates to a constituent unit of the 
			resource represented by the record, subfield $3 is used to specify the portion(s) to 
			which the field applies. The display constant Electronic resource: may be generated.
									
		1 - Version of resource

			Location in field 856 is for the same resource described by the record as a whole. In 
			this case, the item represented by the bibliographic record is not electronic but an 
			electronic version is available. If the data in field 856 relates to a constituent 
			unit of the resource represented by the record, subfield $3 is used to specify the 
			portion(s) to which the field applies. The display constant Electronic version: may be 
			generated.
									
		2 - Related resource

			Location in field 856 is for an electronic resource that is related to the bibliographic 
			item described by the record. In this case, the item represented by the bibliographic 
			record is not the electronic resource itself. Subfield $3 can be used to further 
			characterize the relationship between the electronic item identified in field 856 and the 
			item represented by the bibliographic record as a whole. The display constant Related 
			electronic resource: may be generated.
									
		8 - No display constant generated

Of course, subfield $3 values are not any kind of controlled vocabulary, so it's hard to do much with them programmatically. 

-- Michael

[1] From: http://www.loc.gov/marc/holdings/hd856.html

# Michael Doran, Systems Librarian
# University of Texas at Arlington
# 817-272-5326 office
# 817-688-1926 mobile
# [log in to unmask]
# http://rocky.uta.edu/doran/
 

> -----Original Message-----
> From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of Mike
> Taylor
> Sent: Wednesday, July 07, 2010 5:42 PM
> To: [log in to unmask]
> Subject: Re: [CODE4LIB] schema for "some web page"
> 
> Isn't that pretty much what dc:relation is for?  From
> http://dublincore.org/documents/dcmi-terms/#elements-relation
> 
> Label:	Relation
> Definition:	A related resource.
> Comment:	Recommended best practice is to identify the related resource
> by means of a string conforming to a formal identification system.
> 
> 
> 
> On 7 July 2010 23:32, Jonathan Rochkind <[log in to unmask]> wrote:
> > So in our marc records, we have these 856 links, the meaning of which is
> > basically "some web page related to the entity at hand." You don't really
> > know the relation, the granularity is not there.
> >
> > So, fine, data is data, there ought to be some way to model this in standard
> > XML/RDF/DC/whatever, right?
> >
> > It's not dc:identifier, because dc:identifier ends up including all sorts of
> > URIs that are not really "web pages" at all, they are just identifiers of
> > various kinds.  The marc 856s are URI's, it's true, but they really _aren't_
> > URIs given as "identifiers", they do not neccesarily identify the item at
> > hand at all, but they DO neccesarily lead to a web page with some "see also"
> > relationship to the entity at hand.
> >
> > So... how would you include this in, say, a DC set in XML or RDF?  Is there
> > any common way people have done this in the past?
> >
> > Yeah, I _could_ just expose MODS or MARCXML or what have you. But I'm
> > looking for some vocabulary that will handle marc 856s, but also in the
> > future handle other "some kind of see also link" from other formats, when I
> > add other formats into my corpus. Any ideas?
> >
> > Jonathan
> >
> >