Repository Implementation - Page Book TEI Content Model

Input Master

TEI P4 XML file, with local modifications; follow the guidelines and DTD available at http://text.lib.virginia.edu/bin/cgi-dl/dlps/doco/text/kb/markup_guide/

TEI file contains full header and links to page images, but no text transcription

Datastreams

static TEI XML file with full headers and links to page images, but no text transcription

uvaPageBook disseminators, behaviors, and mechanisms (Page images with no transcriptions)

Content Model	Disseminator	BDef	Behaviors	BMech
uvaPageBook	uvaDefault	uvaDefault	getPreview - returns plain text from DescMeta	uvaDefaultPageBook
			getLabel - returns plain text label from DescMeta
			getDescription - returns plain text description from DescMeta
			getFullView(pid)
			getDefaultContent - returns raw xml
	uvaMeta	uvaMeta	getDescMeta - returns raw xml	uvaMeta
			getAdminMeta - returns raw xml
			getDC - returns Dublin Core elements in raw xml
			viewDescMeta(style)
			viewAdminMeta(style)
	uvaPageBook	uvaPageBook	getPageTurner(pid, behav, refb, id, page, pageNum, snum, sblk, query)	uvaPageBook
			getLabel
			getLabelSTX
			getTableViewSTX
			getChunk(id, page, pageNum) - return xml chunk specified by id
			getChunkSTX(id, page, pageNum)
			getDynamicView (pid, behav, refb, id, page, pageNum, query) - returns cocoon app of dynamic view
			getXML - returns raw xml
			getTEIHeader - returns raw XML
			viewTEIHeader (style)
	uvaOKIPageBook	uvaOKI	getAssetDefs(childPid*, parentPid)	uvaOKIPageBook
	uvaRelationalTEI	uvaRelation	getChildDescMeta(childPid, parentPid)	uvaRelationTEI
			getChildDC(childPid, parentPid)
			viewChildDescMeta(childPid, parentPid, style)

Recommendation: Three TEI models are required to handle the great variation between fully transcribed texts without page images, fully transcribed texts with page images, and page image-only texts without transcriptions.

The existing mechanisms for transcription display can be used on variations of TEI encoding provided we are willing to accept that some elements may be rendered (or not) with the different encoding versions. To have a single TEI content model we must either update the markup to bring all to same encoding standard (over and above parsing against the same DTD), or make the mechanisms more flexible to handle the variations. There are costs in terms of people resources and time to do either.

Page images present a slightly different problem. Content models are dependent on how page-turners are implemented; so multiple models may be required to support a page-turner based on sequential file naming versus one based on sequencing through the TEI file. If resources can be allocated to make all page turned texts consistent, then the need for multiple disseminators, and therefore models, will be obviated.

February 18, 2004; revised April 27, May 19, July 23, August 27, September 17, 2004, and August 15, 2005