October 2003
CONTENTS OF THIS NEWSLETTER
BROUGHAM ROMAN CEMETERY AND XML
Mark Bell
Welcome to what I hope will the first of many newsletters. I started this newsletter because I felt there were a number of us out there in archaeology who were using XML and wanted to share ideas. I hope this won't just be a technical forum - but a way of showing what can be done with XML.
This newsletter is smaller that I expected but the next newsletter will go out in a month or so. This of course depends on the feedback I get and the willingness of you all to contribute items. Items can be as short or as long as you wish. Items listing useful links or requesting information, or describing your research are particularly welcome. Please send items to me at the address above.
Currently this newsletter is mailed out as plain text. I'm investigating software to send HTML email to those who request it. This will allow for some more fancy formatting and live links. Eventually copies will be posted on my website at www.archweb.co.uk.
Apparently a session on XML is planned for the forthcoming CAA2004 conference in Italy, further news next time.
BROUGHAM ROMAN CEMETERY AND XML
Mark Bell
Excavations at the Roman cemetery at Brougham will be published very shortly. It is a remarkable story, not only of the excavation conditions, but of the long struggle to get the site published and the lives of the inhabitants of third century Brougham that have been revealed. The cemetery lies to the east of Brougham castle, once in the county of Westmoreland, now in Cumbria. In 1966 and 1967 part of the cemetery was excavated during the straightening of the A66. Excavations were directed by Dorothy Charlesworth on behalf of the Ministry of Public Buildings and Works. It is the largest cemetery ever to have been dug in the Roman north. Excavation conditions can generously be described as 'difficult', as the excavation was carried out in advance and sometimes during a major road-building project. Finds included horse bones in the pyre debris and the biers were decorated with elaborate bone veneers. The finds show cultural links with the Danube area. The third of a century between the excavation and publication have seen a great increase in knowledge of Roman funerary practice. This has enabled the interpretation of the results to be done in much more detail than would have been possible if publication had occurred straight after the excavation had finished. Because of the importance of this collection for Roman archaeology it was decided to include a CD ROM with the publication. The catalogue will be an invaluable research tool. This is where I was brought into the project. There were two parts to my task. The first was to tidy the Access database used by the finds specialists in their study of Brougham and to document it so it could be deposited with the ADS (Archaeological Data Service). The second task was to produce the CD ROM to go with the publication. Early on due to limits on time and resources it was decided to concentrate on the Microsoft Windows platform. Originally the catalogue was in an Access 2000 database. Once tidied up it was converted to several popular formats - Access 2000, Access 97, Excel 97 and ASCII text. These would be available on the CD for uses to import into their own packages. For users who did not have Access, Excel or equivalent packages a searchable CD ROM had to be created. Hence the idea of supplying data as XML - it would have two purposes - this format would hopefully have some life beyond the current database formats on the CD and also it would let users search the data using a web browser.
TECHNICAL DETAILS
The first problem was to convert a normalized database structure into an XML document. The database was structured like this - a master grave table, with child tables of cremated bone, cremated animal bone, pottery and small finds. All these links were possible one to many relationships. Conversion was made very easy by the use of XML Spy (www.xmlspy.com). XML Spy uses the Microsoft data shaping tools to convert data from Access to XML. Microsoft Data Shaping uses a series of extensions to SQL to create a hierarchical record set from relational data which can easily be converted into an XML document. Data shaping is built into ADO (ActiveX Data Objects) and comes with every Microsoft operating system. The second stage was to create a web page with a search form. This was straightforward HTML with JavaScript code to load a form with choices into a series of drop down boxes, then to pass the variables into the search code. The final stage was to create more JavaScript code to select the results from the XML document. When the search button was pressed on the form the following actions occurred: 1. The XML document was loaded into memory 2. The search variables were put into an XPath search string 3. A new XML document was created from the XPath transformation 4. The XML document was transformed into HTML using XSLT 5. The document is displayed on the page with a cascading style sheet (CSS) to neatly format it. CONCLUSION This was a successful attempt to use XML to provide easy access to a dataset. The downside is that the code is proprietary and only of use on Internet Explorer 5.5 and IE 6. Time and money constraints prevented the development of cross browser code that would work on Mozilla and Netscape 7. Also would it be worth developing code for browsers that are now very much in the minority. Standard tools (open source) to access XML via a browser would be very welcome.
GLOSSARY
CSS - cascading style sheet, a standard for describing how XML and HTML documents should be presented and already used in major web browsers.
DOM - Document Object Model - a means of addressing elements in an XML document from a script.
XPath - a set of rules for finding parts of an XML document.
XSLT - the transformation language used by Extensible Stylesheet Language (XSL) to transform input documents to output documents.
End of Newsletter
**************************************************************
This newsletter is © Mark Bell and the individual authors, 2003. Please contact the editor before reproducing material from this newsletter.