[Archive copy mirrored from the URL: http://www.ra.se/nad/datkat2/se_ead.htm; see this canonical version of the document.]
National Archives of Sweden ----- Date: August 22, 1997 ----- Doc.ref.: 60-2855-97
EAD comments
Preliminary comments on the EAD.DTD by a working group at the
National Archives of Sweden
The project group for the Archival Information System (ARKIS) has commissioned a working group
to review the EAD.DTD. The group should especially consider the possibilities of using the EAD for
exchange of data.
Members of the working group are:
Mats Berggren, IT Department, National Archives (mats.berggren@riksarkivet.ra.se)
John Juliusson, Royal Institute of Technology (john@nada.kth.se)
Per-Gunnar Ottosson, National Archives, author of the present report, (per-gunnar.ottosson@riksarkivet.ra.se)
We would like to thank the EAD Working Group of the Society of American Archivists's Committee on
Archival Information Exchange for this opportunity to present our comments. The result of the EAD
project impresses us, and we are sure that it will be of great value for the archival community all over
the world.
The comments presented below are very preliminary and represent only the views of the working
group.
If the EAD Working Group wishes that we elaborate any of our proposals more in detail we will
certainly try to do that.
General comments
The databases of the National and Regional Archives of Sweden have an exchange format which is
now very close to the MARC AMC. The fundamental public information on fonds levels can hence
easily be converted into EAD. The problems are limited to authority records, terminology and
management fields (see below).
Generally there would be no problems to convert existing Swedish finding aids to EAD by manual
tagging. The problems are the same as those that American implementors have already experienced:
Early finding aids ... frequently contain ... descriptive introductory information that mixes
administrative information including acquisitions, provenance, and access and use
restrictions, as well as biographical or historical and brief scope and content information
(The encoded archival description: retrospective conversion guidelines, Berkeley 1997)
It is proposed that Swedish finding aids in the future should be more adopted to the ISAD(G), and in
that development the EAD would become an important tool.
ISAD(G)
For the use of EAD in the international archival community, it is essential that EAD is fully in line
with ISAD(G), and for most elements that is already the case. There might, however, be some doubts
on the intended application of EAD in some cases. We would therefore suggest
- that the EAD Working Group presents an official EAD interpretation of ISAD(G), e.g. by an edition
of ISAD(G) with comments and coded examples
Proposed element changes
- It is essential to introduce a strict method to record the country and repository codes ( ISAD(G)
3.1.1). This could be achieved by required subelements or attributes for country code and
repository code under
<eadid>
. This information is so crucial for the exchange of data that only
typographical conventions cannot be acceptable.
<unitid>
ought to be required
- A direct equivalent to the ISAD(G) element 3.4.6 Finding Aids is missing. The PRO has suggested
a new element
<unpubfindaid>
for that purpose. We support that proposal, but would prefer that
the element be defined to include references to all kinds of finding aids, whether published or not,
that are not available on line: <findaidlist>
.
ISAAR(CPF)
EAD is not yet adopted to include information from authority records according to ISAAR(CPF). This
can probably be achieved by a minor change. We suggest a new wrapper element for all information
pertinent to a corporate body or person as a subject that might be held in an authority record. It is
essential that access points to a person or corporate body as such are distinguished from the access
points of the records.
<nn>
(we have difficulties in suggesting a proper name, authority description?)
- Attributes
authfilenumber
source
type
legal values according to ISAAR(CPF) 1.2
- May contain:
<name>
with attributes to distinguish authority entry, parallel entry and non-preferred terms
- (Is there really a need for separate elements for personal, family and corporate names when they
can be distinguished by a type attribute?)
<date>
type attribute with values such as date of birth, date of death
<geogname>
type attribute with values such as place of birth, place of death, place of
residence, business location etc.
- <
occupation>
including mandates, functions
<organisation>
<subject>
<bioghist>
<note>
In most finding aids there will be only one such "authority description" for the creator under
<origination>
on the highest level of description, but in case of collections there might be several
on different levels. Concerning other references to persons or corporate bodies the present system with
the attribute authfilenumber in the name elements is satisfactory.
This solves, however, only the problem of including information from an authority file in an EAD
document. As intended with ISAAR(CPF) we also have the need to exchange authority records
between different systems. This might be done by allowing an EAD-file to consist only of authority
descriptions or by producing a special DTD for that purpose. We would be happy to take part in a
discussion towards a common solution of this problem.
Such standardization is of international concern when sharing or linking of textual
information is likely to cross national boundaries. ISAAR(CPF) I.4
XML
It would be most valuable if the EAD.DTD is made compatible with the coming XML standard. We
have observed that the following changes are needed for that purpose according to the working draft on
the XML standard 07-Aug-97, Appendix A XML and SGML - http://www.w3.org/TR/WD-xml-lang
- 2. XML and SGML have different syntax for tags declared with value EMPTY. For example the
element linebreak sould be written <LB/> in XML. To make this a valid SGML-document a smaller
modification has to be done in the SGML-declaration.
- 6. A single attlist declaration applying to multiple elements is not allowed, element C01-C12 should
be redeclared.
- 8. CDATA not allowed in content declaration, used by element <EADID>.
- 10. Minimization markings in element-declarations is not allowed in XML. These should be
removed for all elements.
- 11. Mixed content models must be optional repeateable (eg. use *). Element ORIGINATION,
REPOSITORY and TITLE uses the occurence indicator + which is not allowed according to this
rule.
- 13. NAME, NUTOKEN and NUMBER declared values for attributes is not allowed. Used in
element <PTR> and <DENTRY>.
National conventions and archival management
It cannot be expected that EAD adapt to different national conventions, languages or archival
management purposes. We propose instead that EAD should be made more open for local
modifications in about the same way as TEI. We are aware of that this might cause some problems at
the exchange of information. It is, however, probably much better with one EAD that is open for
modifications, than that we get many different EADs around the world as we now have all the national
MARCs. The method we suggest below makes it also very evident for the users which modifications
that have been made.
Examples of local modifications we need to do
- Legal values for several attributes have to be changed to Swedish archival terms
- References must be done to Swedish thesauri or term lists in attributes such as
source
- Subelements or attributes must be created to allow compatibility with certain administrative fields
in the archival databases.
- For our purposes it is necessary to be able to distinguish numeric values and unit of measurements
in elements such as <extent> and <dimensions>. This latter point might be of such general interest
that it should be included in the EAD (cf. the comments from The Bodleian Library).
Proposed changes of the EAD.DTD
Put the following two entitities in the beginning of the file EAD.DTD:
<!ENTITY % EAD.extensions.ent '' >
%EAD.extensions.ent;
In the file associated with this parameter entity, its allowed to redefine entities used in EAD. In this
way we can for instance use Swedish archival terminology as legal attribute values.
<!ENTITY % EAD.extensions.dtd '' >
%EAD.extensions.dtd;
In the file associated with this entity new elements can be declared.
The doctype-subset in document-instance would then look like this when the adoptions to Swedish
archival standards have been made in the files se_mod.ent
and se_mod.dtd
.
<!DOCTYPE EAD PUBLIC "-//Society of American Archivists//DTD ead.dtd
(Encoded Archival Description (EAD))//EN"
[
<!ENTITY % EAD.extensions.ent SYSTEM "se_mod.ent" >
<!ENTITY % EAD.extensions.dtd SYSTEM "se_mod.dtd" >
]>
In the instructions for the EAD.DTD it could be pointed out that such modifications only should be
done by national EAD committes.
Per-Gunnar Ottosson