--> Finding aids -->Preparing Finding Aids for EAD --> 
Preparing XML Files for Webtext

Unless otherwise indicated, all of the following sequences of commands begin at a DOS prompt.
NOTE: All macros used here are found in the configuration file that is evoked using EDIXIT - not EDIX.

1. Mark up the box file

edixit [name].box
alt-v rip [enter]
After file is"ripped", from DOS command line, type either:
xmlit [name] - if the finding aid is a standard box-and-folder listing,

xmlit [name] rec - if the box list contains significant line breaks.

Use this command if your finding aid is a miscellany or similar type of file that lists cataloging "fields" under folder titles (for example, see YCGLMISC.XML), OR, if your finding aid contains lists of names or titles as folder notes (for example, see Series V in STENZELR.XML). What "rec" does is add a line break at the end of every note line, forcing the XML file to be formatted exactly as input. Without that break, your lists and cataloging details will not be as comprehensible to the reader.

[Note: as of July 2002 all new miscellany items, with the exception of YCGL, are cataloged in the Miscellany Database.]

The completed file will be named [name].xml

2. Mark up front matter file.

2.a. Run the XMLFRN marcro:

edixit [name].frn
alt-v xmlfrn [enter]

The cursor will end up on the text string beinecke.XXXXXXXX.

If file has processed correctly, rename and save the file as [name].srn, using the Edix "alt-f" command:

alt-f [name].srn [enter]

NOTE: do not "alt-w" before "alt-f" in this step, or you will overwrite your text-only .frn file with the marked-up version. "Alt-f" saves the marked-up copy separately as [name].srn, and you still have your .frn file to use for the printed version and for permanent storage in URARC.

2.b. Next, replace the eight Xs in the "beinecke.XXXXXXXX" string with the file name:


(Note: the file name is also the EAD id for the EAD file)

In WKENT.XML -- beinecke.wkent
In PRYCEJON.XML -- beinecke.prycejon

This is the string that will be searched in Orbis from the "View catalog record" link at the top of the finding aid file in Webtext. In order for the link to work, you must also insert the following string in the catalog record for the collection, in a NEW (see note) 035 field, subfield 9:

035 __ |9 (YUL)ead.beinecke.[ead id]

035 __ |9 (YUL)ead.beinecke.wkent
035 __ |9 (YUL)ead.beinecke.prycejon

If there is no catalog record for the collection (as is the case with Osborn files and shelves, for instance), neither of these steps need be taken. Just leave the beinecke.XXXXXXXX string as is in the finding aid. The "view catalog record" link from the finding aid will simply produce a search in Orbis with 0 results.


In catalog records created before July 2002, you will see an existing 035 |9, which looks like this: "XXX9999YL" (for example: 035 __ |9 FJE3667YL ). This code is the seven-character bib id from the Notis system (Orbis1, or "old Orbis") which preceded the Endeavor Voyager LMS (Orbis2), plus "YL" which was added during migration to Voyager. DO NOT OVERWRITE OR DELETE THIS CODE; instead, add a separate 035 field for the ead id code, and enter as described above.

The Notis bib ids were migrated into Orbis2/Voyager because they were frequently used as identifiers for individual records, and it was important that they remain part of the record and be keyword-searchable. Previous to the Orbis2/Voyager implementation, this bib id, or "Orbis key," was the code inserted into the finding aid to provide the link to the catalog record. In finding aids for collections with catalog records created before July 2002, this code continued to be used as the link well beyond that date, and you will see XXX9999-type strings in most finding aids marked up before the autumn of 2003. At some point in the future, these may be converted to "beinecke.XXXXXXXX" links, for consistency.

2.c. Add paragraph tags to any additional paragraphs within <admininfo> sections:

The mark-up macro "xmlfrn" assumes one paragraph in each of the following sections: PROVENANCE, OWNERSHIP & LITERARY RIGHTS, CITE AS, RESTRICTIONS ON ACCESS, PROCESSING NOTES, LOCATION OF ASSOCIATED MATERIAL, and MICROFORM/DIGITAL VERSIONS. If your finding aid has more than one paragraph in any of these sections, you must add paragraph tags manually if you wish the paragraphs to display separately:

Type (or copy and paste) the following between the paragraphs: </p><p>

(Note: the macro will add multiple paragraph tags automatically in the bioghist and scopecontent sections.)

3. Link the two marked up files and run translation macros - adding appendix if needed.

3.a.: To link:
edixit [name].xml
Insert blank line at top of the file then move your cursor to the beginning of the file
alt-r [name].srn
At this point, you may have to mark up and add an appendix. See instructions below. You must mark up and add the appendix before running the translation macros in step 3.b., below.

3.b.: To run translation macros:

edixit [name].xml
alt-v transit [enter]

Note: take care not to run the translation macros more than once, since the unicode equivalents of some characters (&, #, and ©) contain & and #, and if you translate them twice, redundant unicode notation will appear in your finding aid. (For example, instead of "Ownership & Literary Rights," you will see "Ownership &#x0026; Literary Rights.")

4. Add cross-references if necessary.

Check to see if there are cross references in the file by searching for See: or See also:
If these are not found in the file, you do not have to run the macro for cross-references.
If they are found, continue with these steps:

To add cross-references:

alt-v refs [enter]
This will end up creating a new file in F2 called: [name].chk. If this file contains data (other than formatting commands), save it and note that manual corrections need to be made. If [name].chk is empty, no more corrections need to be made.

If a file only has references to Box numbers (not specific folders), you may use the macro "boxrefs" instead of "refs".

To add cross-references manually from [name].chk files, print out [name].chk, then within the XML file, search for occurences of data listed in [name].chk and make appropriate links to folder numbers - (usually Oversize or related material).

See below for an example of how to add a cross-reference link.

If you need to strip all folder ids and targets (leaving simple "ref" tags), you may use the macro "remids". (Note: simple "ref" tags without targets are dead links in the online version of the file. Please remove these before sending to XMLload.)

Notes about cross-references:

Do not run this macro if you don't have see or see also references in your file, or if the references do not contain box and folder numbers (in current practice all x-refs MUST contain a box and folder number, but some older files may contain references lacking them). The target is generated from the folder number, so if there isn't one in the cross reference, the result will be a dead link in the file.

If your finding aid is so large that it will be split into two or more files, you must go through and delete the "ref" tags for any x-refs to folders outside the current file, because these links will not work.

5. Add Series IDs and Related Links

5.a. Add links to the appropriate Series headings and to related collections following the patterns on the Enhancement Instructions and Examples page.

You must add the series ID to each series header as described on the Enhancement Instructions and Examples page. Even if it is not a numbered "Series" (e.g. Restricted Fragile or Oversize), you must add an ID at every <c01> level and to the header for each Appendix. Even short finding aids with no links from the Description to the box-and-folder list require a series ID. Without this ID, the links from the navigation window in Webtext will not work. Links to Appendices will simply not appear in the navigation window without the ID.

5.b.: Add Links to Multiple Bioghist sections:

If there is more than one bioghist section, "althead" attributes must be changed for each occurrence. Each head within bioghist is output with a default value of althead="first". Each subsequent open head tag within bioghist (i.e. <bioghist><head althead="first"> - not to be confused with the closing tag: </head>) must have the althead attribute changed according to its order in the finding aid, using standard ordinals: first, second, third, etc. Existing style sheets can accommodate up to 10 bioghist statements. If you have more, consult the head of the MSS Unit.

6. Save UNIX version of file in EditPad.

Open EditPad
Open file
Remove solid block character at very end of file (an EDIX remnant), then SAVE file

NOTE: Style sheet notations for certain files should be changed at this step.
For most files, there is an instruction at the top of the files that reads: xml-stylesheet type="text/xsl" href="stylenative.xsl"
For Osborn files and drawers listings, change the instruction to read: xml-stylesheet type="text/xsl"
For miscellanies, change the instruction to read: xml-stylesheet type="text/xsl" href="style.xsl".
(Note: miscellany files created from Edix finding aids, such as YCGL MSS MISC and the earlier files of GENM MSS MISC and YCAL MSS MISC, use "stylenativemisc.xsl". Those created from the DBText Miscellany Database use "style.xsl".)

7. Proof the file in Internet Explorer.

Open file in IE using the "Open" function under the File menu.
Errors will be reported if the file cannot be opened. Open the XML file in an ASCII editor (EDIX or EDIT MASTER) to manually correct errors. Repeat this process, if necessary, and reload the file until it displays.

Once the file is displayed, click the right mouse button for the mini-menu and select "Validate XML". This will check the XML file for encoding errors that do not effect display. Errors will be reported one at a time. Open the XML file in an ASCII editor (EDIX or EDIT MASTER) to manually correct errors. Repeat this process, if necessary, and reload the file and select "Validate XML" until the message "Validation successful" is reported.
NOTE: Often, when reloading an XML file in IE following a validation attempt, IE will close. You will have to restart IE.

This stage of trouble-shooting may take some finesse. To proof these files, you must have your style sheet file (e.g. stylenative.xsl) in the same directory as your XML files.

The database administrator may have to assist with more complex errors.
NOTE: Also available on the IE mini menu (right button) is the option "View XSL output" which can be helpful in checking the HTML code that IE generates.

8. Move the finished file to J:\MSSUNIT\WEBTEXT\XMLLOAD

At this point, make sure that you've also copied your urfiles (*.box, *.frn, *.reg, *.app) to J:\MSSUNIT\URARC, especially if you've made any corrections.

9. For new files: Add 856 fields to ORBIS catalog record.

Use the following pattern:

If the file name is "wkent" . . .

 856:42:  |3 Finding aid  |u http://webtext.library.yale.edu/xml2html/beinecke.wkent.nav.html

See also: Orbis Cataloging Manual: Mixed Materials Format

Note: for new files you will also add an 035 |9 to the catalog record, using the following pattern:

035 __ |9 (YUL)ead.beinecke.wkent

(See instruction #2, above.)

10. For new files: Add entries to the Public Services file list (j:\mssunit\webtext\auxfiles\publist.htm), using existing entries as models.

Use a generic text editor (but not EDIX) to edit the file - HTML editors tend to add unnecessary coding. Copy the link just above where you will place your link, and paste it in the correct place.


<p><a href="http://webtext.library.yale.edu/xml2html/beinecke.jobaker.nav.html" class="fndLink">Henry Hurford Janes-Josephine Baker Collection JWJ MSS 2</a></p>

11. Alert the database administrator via email about any new, renamed, or significantly changed finding aids.

Instructions for database administrator.


"Rip"-ing multiple files:

alt-r j:\mssunit\ripm.mac [enter]
alt-b 12 [enter]
in window 4, enter list of names of BOX files to be ripped

replace cursor at first position of window 4

change number in line 12 of macro to equal one more than total number of files to be ripped

(e.g., 4 files = change number to 5; 99 files = change number to 100; 1 file = change number to 2)
alt-v config [enter]
alt-v ripm [enter]
alt-x (OK to lose changes) After file is"ripped", from DOS command line, type either:
xmlit [name]
if the finding aid is a standard box-and-folder listing, or:
xmlit [name] rec
if the finding aid is a miscellany or a similar type that lists cataloging details as notes under folder titles.

Completed files will be named [name].sgm

Creating a list of multiple files:

This can be done easily by:

dir *.[ext] [enter]

then edit out the extraneous data - files size, etc. - leaving only a list of "[name.ext]"

which can be done by -




Replace [ext] with the proper extension - (box or xml)


Cross-reference targets
The following example indicates how the finished coding for a cross-reference should look
The text target="f140" has been added manually to complete the link. The text id="f140" was added previously by the "refs" macro. Note that the following example displays a case in which the targeted folder number is different than the folder mentioned. This happens when a folder that is being targeted falls within a range of folders. You must determine the nearest folder number that you can use as a target.

<note><p><ref target="f140">
See: Series II, Correspondence,
William Garland Rogers, Box 10, folder 142

<c03><container type="box"></container>
<container type="folder" id="f140">140-145</container>
<unittitle>Rogers, William Garland</unittitle>

Comments: Nicole Bouché,
Copyright 1997. Beinecke Rare Book and Manuscript Library
All rights reserved.