*** If a file passes all parsers, but refuses to index in OpenText, the cause could be a tag that is not
properly closed - either in the form of

</title

in which the ending bracket is not present,
or in the form of :

<archdesc
level="collection">

in which the line wrap breaks the tag into two sections, which can be understood by parsers, but which confuses OpenText.

TO USE PANORAMA PRO FOR LOCAL TESTING OF FILES - BE SURE TO

1. HAVE YOUR TEST FILES, STYLE SHEET, NAVIGATOR, AND ANY AUXILIARY FILES (INCLUDING IMAGES AND HELPER APPLICATION EXECUTABLES) IN A COMMON DIRECTORY.

2. MODIFY THE "CATALOG" FILES IN \SOFTQUAD\PANOPRO2\CATALOG TO INCLUDE NOTATION FOR THE EAD.DTD. THE FOLLOWING LINES SHOULD BE INSERTED AT THE END:

PUBLIC "-//UC BERKELEY//DTD FINDING AIDS//EN" "findaid.dtd"

PUBLIC "-//Society of American Archivists//DTD ead.dtd (Encoded Archival Description (EAD))//EN" "ead.dtd"

PUBLIC "-//Society of American Archivists//DTD ead.dtd (Encoded Archival Description (EAD) VERSION 1.0)//EN" "ead.dtd"

PUBLIC "-//Society of American Archivists//DTD eadnotat.ent (EAD Notation Declarations)//EN" "eadnotat.ent"

PUBLIC "-//Society of American Archivists//DTD eadbase.ent (EAD Basic Declarations)//EN" "eadbase.ent"

PUBLIC "-//Society of American Archivists//DTD eadchars.ent (EAD Special Characters)//EN" "eadchars.ent"

PUBLIC "-//Society of American Archivists//DTD eadtable.ent (EAD Table Elements)//EN" "eadtable.ent"

3. COPY ALL COMPONENTS OF THE EAD.DTD TO THE DIRECTORY \SOFTQUAD\PANOPRO2\CATALOG - THESE INCLUDE:

EADSGML.DEC
EAD.DTD
EADGRP.DTD
EADBASE.ENT
EADCHARS.ENT
EADNOTAT.ENT
EADTABLE.ENT

HOW TO CUT LARGE FILES INTO SMALLER UNITS

Because of problems in SGML browser, there are limits to the size of tables.
Currently, it seems to be around 1500 lines (including blank lines).
Early in the EAD project, we discovered that many of our more complex finding aids were
"bunching up" at certain points. This was becaus eht tables, as declared at the <C01> level,
were much too long to be parsed. Therefore, we had to start the practice of dividing long
Series artificially by creating parallel groupings of <C01>s.

The following is an example of how to do this:
In the instance OSBMSS.SGM, there is a single series, "Documents" which runs over
1500 lines. Because the "bunching up" appears near the end of the series, it was safe to asume that the series could be neatly divided in half - A-M & N-Z.

So - the <C01> markup from the beginning of the series was copied to the mid-point of the series - before the first entry starting with "N".

Here is the beginning of the series (with the important coding in red):

<dsc type="combined">
<c01><did><unittitle>Documents</unittitle>
</did>

<thead><row><entry>Box</entry><entry>Folder</entry><entry></entry><entry>Date</entry></row></thead>
<c02><drow valign="top"><dentry><unitloc label="box">fpe</unitloc></dentry>
<dentry><unitloc>12</unitloc></dentry>
<dentry><unittitle>
Adams, Joseph Quincy, 1881-1946
ALS to Walter Wilson Greg, Cornell U,
Ithaca, NY
</unittitle></dentry>
<dentry><unitdate>1929 Aug 10</unitdate></dentry></drow>...

Here is the target half-way point in the series:

<c02><drow valign="top"><dentry><unitloc label="box">pe</unitloc></dentry>
<dentry><unitloc>191</unitloc></dentry>
<dentry><unittitle>
Needham, Francis
ALS to James M. Osborn, Worksop,
Nottinghamshire
</unittitle></dentry>
<dentry><unitdate>1937 Jun 6</unitdate></dentry></drow>

And here is the same section with the new coding copied (in red):

</c01>
<c01><did><unittitle>Documents</unittitle>
</did>
<c02><drow valign="top"><dentry><unitloc label="box">pe</unitloc></dentry>
<dentry><unitloc>191</unitloc></dentry>
<dentry><unittitle>
Needham, Francis
ALS to James M. Osborn, Worksop,
Nottinghamshire
</unittitle></dentry>
<dentry><unitdate>1937 Jun 6</unitdate></dentry></drow>

NOTE that the extra tag </C01> had to be inserted to close the previous <C01>

The final touch is to amend the unittitles so that they make sense in the newly divided instance.
The first section should be changed to read: Documents A-M; the second: Documents N-Z.

This general technique can be applied to any long series that does not display cleanly, due to length. For very long files that need to be cut into separate instances (e.g. Gertrude Stein Papersd and Langston Hughes Papers), it is best to examine them to get an idea of how they were prepared. Basically, it's a matter of duplicating the entire instance and moving one or more series out of the old finding aid and into the new one, by themselves.