1. I am puzzled by the comparison of EAD to ISAD(G) in the last paragraph as they occupy completely different information spaces. ISAD(G)- and DACS, RAD, MAD, AACR, ISAAR(CPF), ISDF, etc.- are all data content and structure standards that have nothing (or at least very little) to do with the how the information is physically stored in some data format, such as a relational database, as an EAD instances in an XML file system, or as an XML file embedded in a relational database or the relative technical performance of such systems.
2. The suggestion that one consider ISAAR(CPF) instead of ISAD(G) for a content standard that might wrap up your indexes is an interesting one. While you probably didn't think of these indexes as authority files of the sort that ISAAR(CPF)was created to support, the data structure of individual records supporting a single entity with a name and relationships to records is certainly a better match in ISAAR that ISAD.
3. As to the best data format standard for storing and manipulating this information, you have several options. Clearly relational databases area choice. In XML, the EAD element <index> was created to support this sort of data but given the size of your indexes this might not be practical, esp. if there is just one big index to multiple collections. EAC, with its close support for ISAAR, might then be a better fit though be warned that, as I understand it, the EAC Working Group is in the process of making significant changes in the next draft of the schema, so don't commit yourself too far to the earlier draft that is currently available.
From: Encoded Archival Description List [mailto:[log in to unmask]] On Behalf Of Gijsbert Kruithof
Sent: Thursday, June 12, 2008 2:07 AM
To: [log in to unmask]
Subject: Re: EAD and finding aid databases
Thanks for your reactions!
In fact I did refer to both situations Kate mentions below. Mostly our indexes refer to a specific recordgroup, but sometimes across a number of record groups and even across a number of inventories! These are really awful in terms of EAD ;-) But maybe the EAC-CPF option Michelle proposes offers a solution ...
I quess we in the Netherlands have a lot of really old archives with indexes made long ago and then later typed over or (partly) digitized.
So this is very similar to the example with the card index.
However the Lowell inventory is a good typical example of the difficulties in EAD. (I would be interested in seeing the actual XML. I imagine it to be quite large and hard to handle ...) Sometimes the size of the EAD XML file is simply getting too big to handle for stylesheets. Especially if the names are all on one level. We have an inventory that is essentially a long list stating all commercial corporations registered by the state in a certain period. This is an XML file with the size of 20 MB. Our pdf-stylesheet can't make a pdf and the browser using the HTML stylesheet can't load the page.
So I think the use of EAD has limitations just because it is a XML file format, whereas ISAD(G) in principle poses no problems in incorporating these large indexes. A solution could be like Elizabeth Perkes describes in Utah: storing and handling the data in a database and only producing EAD XML when needed
Van: Kate Bowers [mailto:[log in to unmask]]
Verzonden: woensdag 11 juni 2008 14:59
Aan: Encoded Archival Description List; Gijsbert Kruithof
CC: [log in to unmask]
Onderwerp: Re: EAD and finding aid databases
My experience is that EAD does not accommodate indexes created prior to ead very well, if at all.
For record groups or collections to which we had a specific index but no inventory, such as:
We managed to derive an inventory of sorts, and add the index text to the document, although such indexes cannot be <index> in EAD because they do not make reference to a specific part of an inventory. I think with Lowell (hua03003) we ended up using <c> even though these are not really components.
We have a very different situation, however, with a card index that indexes persons and topics across a number of record groups of University records up to 1850. I think this latter scenario is the one you are asking about. I have not attempted to encode this. The cards are handwritten and thus do not lend themselves easily to transcription.
Collections Services Archivist
Harvard University Archives
Quoting Gijsbert Kruithof <[log in to unmask]>:
> Apart from our finding aids in EAD the National Archive in the
> Netherlands also has about 300 indexes, cardboxes and databases.
> They contain information about persons and places and other content
> not specifically described in the inventories. Some databases contain
> more than 100.000 items on which information can be found in an
> archive. (See for instance http://databases.tanap.net/vocrecords/)
> Because of their size and complexity these indexes and databases are
> difficult to incorporate in the EAD inventories. Maybe we are
> confronted here with the limitations of EAD as an XML file format ...
> Does anyone has experience in dealing with these 'finding aid
> We plan to create a Database Management System containing all the
> databases. This DBMS will be complementary to the EAD inventories and
> contain links to the inventorynumbers.
> But maybe there are other options. We would be very interested.
> We would especially like to know if someone has experience setting up
> such a system with regard to:
> - functionality for staff and volunteers
> - choice of DBMS and software: MySQL/PHP or SQL-server/ASP or
> something else
> - datamodel
> - linking functionality to EAD inventories
> - digitizing typoscript cards from cardboxes
> We appreciate all suggestions
> Gijsbert Kruithof
> Senior archivist
> Nationaal Archief
> The Netherlands
This message was sent using IMP, the Internet Messaging Program.