We also have been looking for somewhere to put the ILS record # into AT and have pretty much decided to put it in the "finding aid note" with a predictable label so that our XSLT stylesheet can find it and tuck it into the place we usually put ILS record #s in EAD.  AT exports the finding aid note into the EAD in <filedesc><notestmt><p>


In our EAD, we've been tucking ILS database record numbers into the eadid identifier attribute value (i.e. <eadid identifier="##########">), which is not exactly kosher, but has a relationship because we deliver the finding aids through the discovery layer and they are connected to the database record identified by that number.  On the other hand, AT  doesn't ingest this attribute or value.


I'm not entirely comfortable with putting a database control number for a foreign system into an EAD <unitid>, even though this is explicitly given as a use in the tag library.  It conflates two different things, leaving the database identifier to be re-defined by attributes.  If you actually use the your ILS record number as the reference code (call number) of your collections then it may be a moot point, but if you don't, it conflates  the identifier of a description of the collection with the identifier of the collection itself.  Then you need to use an attribute that conveys the notion of "Oh, but it's not really a collection identifier, it's really a database record identifier."  I've been around too long and have too seen many instances when such conflations have unintended consequences, but as long as we're careful with attributes, it may be the best option.


We've learned so much that I'm very hopeful that EAD2013 and ArchivesSpace will address a number of issues that were not foreseen when EAD was first imagined.









Kate Bowers

Collections Services Archivist

Harvard University Archives

Cambridge, MA 02138

[log in to unmask]

voice: (617) 384-7787

fax: (617) 495-8011

web: http://nrs.harvard.edu/urn-3:hul.eresource:archives


From: Encoded Archival Description List [mailto:[log in to unmask]] On Behalf Of Noah Huffman
Sent: Tuesday, June 05, 2012 9:30 AM
To: [log in to unmask]
Subject: Re: [EAD] System Control Number in EAD


Hello all,


I'm looking for an appropriate way to encode a bib system number in EAD and found this thread.  I think the multiple <unitid> approach that Michele describes is the best approach--one for the collection number and one for the bib system number--but there is one catch.  As far as I can tell, Archivists' Toolkit only supports the use of one <unitid> element at the collection level.  If you try to import an EAD file with more than one collection level <unitid> element, AT just imports the one without any attributes.


I suppose I could insert the <unitid encodinganalog="001" type="bibsysno"> after exporting EAD from AT, but I'd prefer to find a solution where I could manage both the collection number and bib system number in AT and export them both in EAD without having to post-process the EAD file.


And so I consult the collective wisdom of the EAD list... Any AT users have suggestions?







Noah Huffman

Archivist for Metadata and Encoding

David M. Rubenstein Rare Book & Manuscript Library

Duke University

[log in to unmask]




Visit our blog "The Devil's Tale": http://blogs.library.duke.edu/rubenstein/

Be our fan on Facebook: http://www.facebook.com/rbmscl   






On Fri, Mar 16, 2012 at 11:59 AM, Michele R Combs <[log in to unmask]> wrote:

We map the 001 field (the BIBID) to the collection-level <unitid> (ead/archdesc/did/unitid) .    The definition of unitid seems to support this as the correct choice: “Any alpha-numeric text string that serves as a unique reference point or control number for the described material, such as a lot number, an accession number, a classification number, or an entry number in a bibliography or catalog...the TYPE attribute may be used to indicate the system from which the <unitid> was derived, e.g., accessioning system, record group classification scheme, records retention scheduling system, etc.”


For example:


<unitid encodinganalog="001" type=”opac” label="Identification: " countrycode="US" repositorycode="NSyU">148784</unitid>


<unitid> is of course usable at any c0# level, so you could have different ones for fonds, series, etc. at the appropriate level of the description.




From: Encoded Archival Description List [mailto:[log in to unmask]] On Behalf Of Vicky Phillips
Sent: Thursday, March 15, 2012 6:33 PM
To: [log in to unmask]
Subject: System Control Number in EAD


Hello, We are currently starting on the digitisation of archives and manuscripts and for the first time will be working with EAD within METS. The standard transformation from MARC to EAD does not incorporate the MARC field 001, which contains the System Control Number/Unique identifier within the catalogue, which we depend on heavily in the digitisation workflow. I was wondering if those of you who are already using EAD could let me know if you include this in your EAD and if so which element and attributes do you use? For example the manuscript collection I am starting with has a System Control Number at Fonds, Series and File level. We were thinking on the lines of placing the System Control Number (vtls004593617) for the Fonds in the element within element using identifier and encodinganalog attributes as follows: The above could also contain the handle to the parent object e.g. Then the element within element for FONDS level record will contain the NRA Code (MARC field 099) However, I notice that the LC have done this differently. Placing handle in identifier in eadid and note regarding Catalogue Record (System Control Number) in a note element within filedesc http://hdl.loc.gov/loc.mss/eadmss.ms001004

Catalog Record: http://lccn.loc.gov/mm79011056

For all sub levels Then for all sub levels (Series, File) which have a bibliographic record of placing the System Control Number in the identifier attribute within the element along with encodinganalog attribute which will contain the MARC field number and possibly label attribute e.g. Being relatively new to EAD I was wondering if what we have suggested is ok or whether those of you who have implemented have any other suggestions and better still examples of how you have incorporated LCNs into your EAD. Any guidance / feedback would be greatly appreciated. Thanks, Vicky