As a former beginner myself, I recognize that the desire to produce EAD finding aids does not immediately result in their production. Among the members of CRRA, the more prosperous institutions can support computer systems that automatically generate EAD. But even these systems require an understanding of archival theory and how it might be maintained in a way congenial to machines. And other members of CRRA have the more difficult task of producing EAD without sophisticated equipment and IT support.

CRRA developed the EAD Template to help those who have no easy way to produce valid finding aids. While it tries to make EAD as easy as possible, beginners may still have bad luck with it. When I attended the Rare Book School EAD course at the University of Virginia in 1999, I had already produced hundreds of EAD finding aids. Daniel Pitti, the teacher of the course, offered helpful criticism and mentioned several mistaken ways of thinking that he had found among beginners. When CRRA members started using the EAD Template, I found one of these typical mistakes cropping up.

Archivists preserve filing systems that grow organically as routine activities generate records. The context of a document has evidential value; the whole filing system provides evidence about how an organization or individual conducted business, and therefore evidence of values, crises, attacks, defense mechanisms, unexpected events, daily life. Descriptions of such filing systems need to represent the organic form faithfully in all of its meaningful structure and peculiar quirkiness. Since filing systems generally have a hierarchical form, finding aids must also allow for components that contain other components that may contain other components -- boxes within boxes within boxes, or files within files within files.

Thanks to our familiarity with common filing systems on computers, we should have no trouble with this concept. Typically graphical displays of the contents of a hard drive show pictures of folders, and inside them other folders, and after a number of these eventually the individual computer files. One can generally also choose to view the same structure as an outline representing this structure in a slightly different way.

EAD finding aids are like that. But archivists give names to different levels of description, and EAD can assign numbers to individual components indicating their level of indentation in the outline or their place in the hierarchy. The beginner's mistake has to do with the rigid assignment of component numbers to each level of description.

Here are some names of levels of description: record group, subgroup, series, subseries, file, item. The typical mistake goes this way: we will make all of our record groups level 1, all of our subgroups level 2, all of our series level 3, all of our subseries level 4, all of our files level 5, and all of our items level 6.

But any given filing system might not have all of these. Your computer might have a folder called Documents that contains nothing but individual documents. Your filing cabinet might have a series of files without any subseries. And the simple scheme above does not take into account that subseries often have subseries of their own, which might have subseries of their own before the file level and then the item level.

Presently our largest finding aid in the Notre Dame Archives (10 megabytes) describes a correspondence file kept by early Notre Dame presidents. Here are the first few lines of the description of subordinate components:

UPEL - - Notre Dame Presidents' Letters, 1856-1906: Records - 1856-1906

UPEL 1/01 - Document - Binding and Index of Old Book that Contained Folders 2-25.

UPEL 1/02 - Letter - Lucas Brothers: Catholic directory questionnaire - 1856/09/15

UPEL 1/02 - Letter - to: Michigan Central RR - 1856/12/27

 <c01 level="recordgrp"><did><unitid>UPEL</unitid> <unittitle>Notre Dame Presidents' Letters, 1856-1906: Records</unittitle> <unitdate type="inclusive" normal="1856/1906">1856-1906</unitdate></did>

<c02 level="item"><did><unitid><lb/>UPEL 1/01</unitid> <unittitle>Binding and Index of Old Book that Contained Folders 2-25.</unittitle></did></c02>

<c02 level="item"><did><unitid><lb/>UPEL 1/02</unitid> <unittitle>Lucas Brothers: Catholic directory questionnaire</unittitle><unitdate normal="1856-09-15">1856/09/15 </unitdate></did></c02>

<c02 level="item"><did><unitid>UPEL 1/02</unitid>
<unittitle>to: Michigan Central RR</unittitle><unitdate normal="1856-12-27">1856/12/27</unitdate></did></c02>

In this case, the description is nothing more than an item list. (The collection codes and container numbers follow our local practice; other repositories have other conventions.)

Beginners who make the typical mistake might have tried this:

<c01 level="recordgrp"><did><unitid>UPEL</unitid>
<unittitle>Notre Dame Presidents' Letters, 1856-1906: Records</unittitle>
<unitdate type="inclusive" normal="1856/1906">1856-1906</unitdate></did>

<c06 level="item"><did><unitid>UPEL 1/01</unitid>
<unittitle>Binding and Index of Old Book that Contained Folders 2-25.</unittitle></did></c06>

<c06 level="item"><did><unitid>UPEL 1/02</unitid>
<unittitle>Lucas Brothers: Catholic directory questionnaire</unittitle>
<unitdate normal="1856-09-15">1856/09/15</unitdate></did></c06>

<c06 level="item"><did><unitid>UPEL 1/02</unitid>
<unittitle>to: Michigan Central RR</unittitle>
<unitdate normal="1856-12-27">1856/12/27</unitdate></did></c06>

Or, in the format used by the EAD Template:

1 UPEL ~ 1856-1906 ~ recordgrp ~ Notre Dame Presidents' Letters, 1856-1906: Records

6 UPEL 1/01 ~ ~ item ~ Binding and Index of Old Book that Contained Folders 2-25.

6 UPEL 1/02 ~ 1856/09/15 ~ item ~ Lucas Brothers: Catholic directory questionnaire

6 UPEL 1/02 ~ 1856/12/27 ~ item ~ to: Michigan Central RR

This EAD will not work. Inside level 1 we always need level 2. The component numbers must follow the structure of the collection being described.

Nor would it be good practice to fill in the blanks -- for example:

1 UPEL ~ 1856-1906 ~ recordgrp ~ Notre Dame Presidents' Letters, 1856-1906: Records

2 UPEL ~ ~ subgrp ~ No sub-group

3 UPEL ~ ~ series ~ No series

4 UPEL ~ ~ subseries ~ No subseries

5 UPEL 1/01 ~ ~ file ~ Box one, Folder one

6 UPEL 1/01 ~ ~ item ~ Binding and Index of Old Book that Contained Folders 2-25.

5 UPEL 1/02 ~ ~ file ~ Box one, Folder two

6 UPEL 1/02 ~ 1856/09/15 ~ item ~ Lucas Brothers: Catholic directory questionnaire

6 UPEL 1/02 ~ 1856/12/27 ~ item ~ to: Michigan Central RR

This would produce a valid finding aid, but would add little meaningful information.