EAD - XML Finding Aids

EAD - XML Finding Aids - Technical Information (by Sigrun Ragnarsdottir--sr75@cornell.edu)

FAQ

How can I view the xml source code?
There are two ways, either right click in the main frame and select "View source" or go from the navigation bar to the XML link. You should see the xml document, but the navigation bar and the frame will disappear.
Why can't I use Netscape?
In order to view the XML document, you need an XML capable browser--one that can parse the XML document.
When will IE with XML support be available for Machintosh?
MS reports that an XML capable browser will be available for Machintosh later this year.

Scanning and OCR

We used OmniPage Pro 9.0 for scanning and OCR from Scantek III scanner and saved the output in Microsoft Word. However, we had to add handwritten notes by hand. Most of the text went through quite nicely. Some of the text had to be rewritten with digital access in mind. For instance, we added related links and shortened the biographical history. Some of the unittitle were reformatted. We tried not to immitate the paper copy and rather think from scratch about the content of the digital copy with access in mind.

Tagging

We have been developing an EAD/SGML template for our finding aids over the last several months. We've recently converted it to XML. Now we are in the process of writing a up-to-date version of the XSL layout as well as creating a workflow for future finding aids and converting the one already in digital format.

We've experimented with a number of solutions for tagging the raw electronic files:

XML Notepad
The problem with XML Notepad is that it changes all tagnames to uppercase. David wrote a Perl script to reverse this for validation (since element names in XML are case sensitive). The good thing about XML Nodepad is that it leaves the XML document well formatted and well-formed.
TextPad
TextPad is enhanced Notepad. The good thing about TextPad is that many documents can be open simultaniously so, unlike Notepad, navigation is easier. The bad thing is that it doesn't form the text--it can get a little messy. It doesn't make sure the pages are well-formed nor well formatted.
XML Spy
XML Spy can edit both XSL and XML document. It leaves the document well formatted and well-formed. It provides a tree structure of the document which makes it easier to navigate.

Parsing and validating

We used James Clark's SP to parse the XML document. IE5 validates the document when it loads it using the new MSxml parser that ships with IE5. When it doesn't validate it doesn't display the data in the browser.

Web pages (green = htm, yellow = xml and blue = xsl)

General navigation

Figure 1. General navigation of .htm pages

The idea is to have a "default" page linked to a "browser" page, which checks whether the user has IE4.0 or higher and directs the user to the appropriate page. We use Javascript to do the checking.

<html>

<head><title>Browser verify</title></head>

<body onload="browser();">
<script language="JavaScript">
function browser() {
browser_name = navigator.appName;
browser_version = parseFloat(navigator.appVersion);
goodBrowser = false;

if( browser_name == "Microsoft Internet Explorer" && browser_version >= 4.0 ){ goodBrowser = true; }
if( goodBrowser )
{
location.replace("Fuertes.htm");
}
else
{
location.replace("Fuertes-html.htm");
}
}

</script>
</body>
</html>

Code 1. Browser

Fuertes-HTML (an HTML version of the XML document) was created using Javascript to load both Fuertes.xml and Fuertes.xsl. This presents HTML to the browser, which can then copy the output ('view source'). This HTML output was then tweaked using Microsoft's Front Page.

<HTML>
<SCRIPT>
function display()
{
data.transformNodeToObject(ss.XMLDocument, resultTree.XMLDocument);
result.document.write(resultTree.xml);
}

function refresh()
{
result.document.body.innerHTML = "";
display();
}

</SCRIPT>

<SCRIPT FOR="window" EVENT="onload">
data.async = false;
data.load("fuertes.xml");
ss.async = false;
ss.load("fuertes.xsl");
display();
</SCRIPT>

<XML id="data"></XML>
<XML id="ss"></XML>
<XML id="resultTree"></XML>

<FRAMESET>
<FRAME ID="result">
</FRAMESET>
</HTML>

Code 2. Transformation

The Fuertes page

Figure 2. What happens when Fuertes is loaded

Fuertes.htm loads source to two frames, Navigation (left) and Main (right). It is defined so that every hyperlink in navigation points to main.

The source for Navigation is Fuertes-nav.htm, which contains JavaScript that loads Fuertes.xml and Fuertes-nav.xsl. This is useful since we only want to have one xml document. The script is the same as for the transformation except for different xsl document.

The source for Main is Fuertes.xml, which in return loads Fuertes.xsl to display itself.

Back to current implementation page