[Metadatalibrarians] Crosswalks

[Cheryl Walters]

Hi Teressa,

I really empathize with your question because crosswalking from one format to another still seems a little like magic (and a lot of hard work) to me.

Depending on your project, you can use commercial applications, shared applications written by a generous colleagues from other institutions, or applications created specifically for your project by in-house programmers.  Below is a scenario for our regional EAD finding aids project which used all three types of applications.

We wanted to import finding aids into CONTENTdm which uses Dublin Core metadata.  We wanted to display the finding aids and make them comprehensibly searchable in a manner compliant with current archival guidelines (i.e. Describing Archives: A Content Standard, DACS and Encoded Archival Description, EAD).   The finding aids were in various formats such as Word documents, html, analog print, etc.

The first step was to encode them all into EAD format which is expressed using XML, a markup language.  So, in other words, each field or element is defined according to the EAD schema and marked up using XML for data manipulation and display.  Our project used programmers to create various tools to assist in transforming the finding aids into XML EAD format.  One such tool was a template, with empty blocks into which you could manually type or paste in the data; once the data was input, the template tool would save it all as an EAD XML file.  All the imputter had to do to create the title information, for example, was to type the actual title (such as Papers of Samuel Adams) into the Title block in the template; when the inputter saved the completely filled in record, the title was regurgitated into an XML EAD formatted title:

<titlestmt>
       <titleproper encodinganalog="title">Papers of Samuel Adams</titleproper> ...

We also used existing applications such as the XML editor Oxygen and Word to help get the old finding aids into EAD XML format using a variety of functions such as Find and Replace, Copy and Paste, Mail Merge, etc. and we hired a vendor to help us with this encoding.

Once the finding aids were encoded into EAD, we needed to ingest them into CONTENTdm where Dublin Core fields were needed for searching purposes.  So another programmer created a tool written in Visual Basic which converted the EAD fields into DC format, saving the DC field data into an Excel spreadsheet that could be uploaded into CONTENTdm along with the original EAD finding aid.  The DC data went into the corresponding DC fields, creating a metadata record during this import process.  The mapping that directed this conversion of EAD title into Dublin Core title formatting looked something like this in the Visual Basic script:

eadToCdm(6) = 5
 ' titleproper type="filing" to title
eadFields(7) = "ead/eadheader/filedesc/titlestmt/titleproper[@type='filing']"
eadToCdm(7) = 0
 ' unittitle to title
eadFields(8) = "ead/archdesc/did/unittitle"

None of this was easy and many of us didn't know either EAD or XML formats.  We shared expertise across a number of institutions.  It has taken us a year to get to where we are just about ready to do a final uploading for release to the public.

Hope this helps make the theory you learned a little more real.


Cheryl D. Walters
Head, Digital Initiatives Dept.
Merrill-Cazier Library
Utah State University
UMC 3000
Logan, UT  84322-300

(435) 797-2623
(435) 797-2880 (fax)
Cheryl.Walters at usu.edu







On 8/14/08 8:18 AM, "Keenan, Teressa" <teressa.keenan at mso.umt.edu> wrote:

Hi everyone;

I am new to metadata librarianship. Though I've been lurking on this list for a while. I officially start my new position as a metadata/digital projects librarian on September 1st.  I am very excited about actually putting my education to use in the real world.  I have a question that I hope isn't too silly and that someone here on this list can point me in the right direction.

We learned a bit about crosswalks in library school.  Basically we either read about or set up tables that show which elements are equivalent to each other, but the one thing we didn't cover was how this actually happens.  So if a person wanted to take MARC records from their local ILS and transform them into DC, how would they do that?  I am assuming that this isn't only done manually with someone physically keying in (or copying and pasting) the information into a form.  Is there some literature out there someplace that explains or describes the process?  Has anyone actually done this in real life and not just in a classroom on paper?

I realize I haven't even officially started my new job yet; but I have a feeling that this is going to be a topic that comes up very soon so I'd like to learn as much as possible.

Thanks in advance for any tips you all can send my way.


Teressa M. Keenan
Library Technician - Bibliographic Management Services (until 08/31/08)
Metadata/Digital Productions Librarian (after 09/1/08)
Maureen and Mike Mansfield Library
The University of Montana
32 Campus Drive (MMLA01)
Missoula, MT 59812-9936
teressa.keenan at umontana.edu , (406) 243-6862

_______________________________________________
Metadatalibrarians mailing list
Metadatalibrarians at lists.monarchos.com
http://lists.monarchos.com/listinfo.cgi/metadatalibrarians-monarchos.com