In his reply to this question, Steve Hensen refers the reader to section
4.3.4 of that indispensible tome, the EAD Application Guidelines, for
further information on this subject.
Let me expand briefly on that text to describe how we are implementing this
conversion here at the Minnesota Historical Society.
1. The MARC catalog record for the collection is downloaded from the OPAC
to a local network drive using a function built into the OPAC- a fairly
standard feature of such software. The record is in the MARC transmission
format. The collection's id number is assigned as the file name, e.g.
P1905 or 1186.
2. The staff member goes to that directory at the DOS command line (both
conversion applications are run from a commnd line) and types "convert
filename", thereby executing a DOS batch file called convert.bat.
3. That batch file executes two operations.
The program marcxml.exe from Logos Research converts the record from the
MARC transmission format to the MARC DTD format as an XML document.
James Clark's XT processor software then uses an XSL stylesheet to transform
the document from the MARC DTD structure to EAD.
The resulting file is output as a fle named Collection_Number.xml, e.g.
p1905.xml or 1186.xml.
4. The resulting xml file is loaded into an XML editor for enhancement,
adding various values required for the <eadheader> section.
The principal ambiguity of this method is that one cannot always predict
what went into a MARC 500 field and therefore reliably map its contents to
the correct EAD element. Any adjustments are made manually in the xml
editing software. There may be other areas in your MARC implementation
where there are similar ambiguities.
As others have reported, success may be had with other tools including the
perl scripts available from the LIbrary of Congress to do the step of
converting the data from the MARC transmission format to XML in the MARC DTD
The process could be reversed- going from EAD to MARC- though I have not
written the stylesheet necessary to do so. However, two additional editign
steps would be requried if took this route. You would have to supply
default values for the MARC fixed fields during the conversion and edit them
afterwards- notably date1, date2 and date type fields. Terms in fields 1XX,
6XX, and 7XX
would have to be mapped to a single subfield $a within the appropriate
fields with the subfield codes and indicator values being edited
subsequently in your MARC editing software. This assumes, of course, that
your OAPC has the facility to import locally created MARC records and not
simply to receive them from a utility.
Head of Processing
Minnesota Historical Society
345 Kellogg Blvd West
St. Paul MN 55102-1906
[log in to unmask]
**NOTE NEW AREA CODE EFFECTIVE JULY 12, 1998**
> From: Stewart Unwin[SMTP:[log in to unmask]]
> Sent: Wednesday, October 27, 1999 11:38 PM
> To: Multiple recipients of list EAD
> Subject: MARC ==> EAD
> Back in June this year there was some discussion on this list of
> MARC records to SGML format using scripts available from the LC MARC SGML
> site http://lcweb.loc.gov/marc/marcsgml.html and then using the
> transformation capabilities of XSL to convert them from MARC DTD to EAD
> Michael Fox mentioned that such techniques would be discussed in the EAD
> Application Guidelines (which I haven't read yet).
> What I need to know is where I can find a document which explicitly maps
> MARC fields/sub-fields to EAD elements. Is there an "official" mapping
> I am overlooking, or are people just developing their own mappings for
> own projects? If the latter, is anyone willing to share the fruits of
> I'd be grateful for any advice.
> Stewart Unwin
> Web Services Branch
> National Library of Australia
> Tel: +61-2 6262 1544
> Fax: +61-2 6273 3648
> E-mail: [log in to unmask]