Print

Print


Spam: EAD MARC Dublin core pt2 (fwd)
The point of mapping, of course, is to form the basis of some form of transformation of the EAD-encoded data into another metadata scheme or HTML.   For example, I need to convert the data in my EAD data structure into Dublin Core structure for harvesting or I want to generate MARC records for loading into my ILS.
 
So when I convert the data into DC, what do I want to describe?   The DC element Tiitle is instructive of one's choices.  
 
The <titleproper> element in <eadheader>  contains the title of the finding aid, such as "An inventory of the Mary Smith papers".   The <unittitle> element in <archdesc>/<did> contains the  name of the collection, i.e. The Mary Smith Papers.    Those are two very different things.   If I generate a DC record with a Title element that contains the text of the former (An inventory of the Mary Smith papers), I am describing the finding aid (or at least this digital version of it).  
 
If I generate a DC record where the DC:Title element contains the text from <unitttitle>, The Mary Smith Papers, I am describing the collection.   What do you want your potential users to discover- the finding aid or the collection?    The question is not so much how much data goes into the DC record as to what you want to describe.  
 
Unless you want to call attention to the finding aid itself, perhaps as a formally published work, my take is that one probably wants to point the user to the collection.   Others may feel differently.  
 
As I began, encodinganalogs are useful primarily, perhaps exclusively, as directions for some transformation program, such as an XLT stylesheet, as to which element in your <ead> document either goes into a particular element in the DC (or MARC or MODS or METS) record that results from a transformation into those metadata structures or, in the case of transformations into HTML, which elements might go into a <meta> tag in the <head> of the resulting HTML document to encourage retrieval by web search engines.  
 
 
Michael
 
-----Original Message-----
From: Encoded Archival Description List [mailto:[log in to unmask]]On Behalf Of John Bewley
Sent: Wednesday, January 25, 2006 7:42 AM
To: [log in to unmask]
Subject: Spam: EAD MARC Dublin core pt2 (fwd)

Please excuse if this is a duplicate mailing.

> Michael,
>
> Replying with the benefit of reading Dennis Meissner's message, it would
> seem that Dublin Core is more appropriately applied to the finding aid
> itself. Here are some of my related questions:
>
> Is there any consensus in practice regarding mapping or would anybody
> even argue that there's a need for consensus here?
>
> If Dublin Core is used only within <eadheader> the resulting "record"
> seems mighty brief. If this metadata is primarily for web retrieval of
> some sort does the metadata match the needs for this sort of retrieval
> (searching, harvesting).
>
> I have seen some finding aids represented online in HTML rendering using
> the standard frames layout. Some institutions place meta tags mapped to
> Dublin Core within the HTML coding for the home frame page. The set of
> Dublin Core elements are usually more inclusive in these cases, sometimes
> including indexing terms and other metadata outside the scope of
> <eadheader>. In cases where I am able to see the originating XML document
> I don't see Dublin Core mapping so I want to know if the coding is being
> done by hand on the frame.html page or if it is being generated in some
> way from the XML document.
>
> And, on this same issue, is this a short term fix if the intention is to
> be able to mount our XML documents directly on the Web in the future when
> browser capabilities allow?
>
> How do you see Dublin Core mapping within finding aids being utilized?
>
> John
>
> --On Tuesday, January 24, 2006 4:48 PM -0600 "Fox, Michael"
> <[log in to unmask]> wrote:
>
>> The question raised regarding mapping turns on another question.   In
>> your Dublin Core record, are you describing the finding aid (if so map to
>> <eadheader>) or are you describing the collection (if so map to <ead> or,
>> more precisely, to <archdesc> elements)?
>>
>> Michael Fox
>
>