Now you've learned about how to create TEI documents, but we haven't said anything about what you can do with themߪ

There is a very wide range of things you can do with TEI documents, involving in-depth analysis, data mining, processing to discover patterns of various sorts.

We're going to focus on "publishing", both in the narrow sense of "making them readable online" and also in the broader sense of "exploiting the encoding publicly". But most of the more advanced things you can do with TEI documents use technologies similar to the ones we're talking about here.

The simplest approach to publishing TEI documents is that which we did yesterday: just apply a CSS stylesheet. same process as applying a CSS stylesheet to HTML documents the browser reads the TEI file, which points to a stylesheet the browser reads the stylesheet, and applies its styling to the TEI elements as specified the browser displays the formatted text can control everything including fonts, colors, backgrounds, layouts (where the chunks of text are placed on the page), etc. modern standards-compliant browsers can all do this

This is great, but there are some limitations: can't (at the moment) make links — can make it underlined and blue can't search, except by using your browser's find command can't do any sort of higher-level stuff (of the sort that we'll see in a minute)

The Extensible Stylesheet Language allows you to transform XML documents in many ways: into other XML documents, such as XHTML, TEI, XSLFO, DocBook, etc. into other formats: TeX, RTF, pretty much anything if you can figure out how

Transformation into other XML documents can mean several things: taking the entire TEI document and converting its markup into HTML, so that you now have an HTML-encoded document taking the entire TEI document and transforming bits of it into HTML: for instance, taking just the section headings and making an HTML-encoded TOC; or taking a long TEI document and transforming it into separate HTML files, one for each chapter, accompanied by a TOC; etc. transforming one kind of TEI markup into another: for instance, if you mark up your documents using a customized schema, but you want to exchange data with other projects, you might convert your markup to TEI Lite for easier interchange.

This transformation can be done as a process that you run in advance, and then use the output. For instance, you might have a set of TEI files which you transform to HTML, and then mount the HTML on your web site. When you make an update to the TEI files, you run the transformation again, and remount the resulting HTML.

Tools you need for this kind of transformation: an XSLT processor (some are built into Oxygen); there are several, they have different virtues which we won't go into here an XSLT stylesheet The processor reads the stylesheet, and reads your XML file, and it applies the stylesheet to the file and outputs a result.

Then the result can be used as appropriate: styled with CSS and viewed in a browser if HTML; viewed in a browser or PDF reader if PDF; etc.

You can also run these transformations on the fly, as part of your publication system: your TEI files live on a server when a user requests a file (e.g. by clicking on a URL), the transformation software performs the transformation on the fly and delivers the resulting HTML. the transformation might vary depending on the request: for instance, a user clicking on the sort by date link would get different output — from the same underlying TEI file — that she would get by clicking on the sort by author link

Tools you use for this kind of transformation: e.g. Cocoon web glue for your web application development needs; DIY Framework?; Struts?; RIFE.

This is great, but there are some limitations: still not much searching what searching there is will be slow; you're using a tool not designed for handling searches efficiently not good for managing large aggregations of files efficiently, or for managing them as a group, dealing with information that cuts across the entire aggregation

These kinds of tools are designed to manage large groups of XML files, and to provide certain kinds of advanced functionality: fast, efficient searching transformations involving groups of files: not just transforming each file separately, but doing transformations that involve taking parts of different files and creating new results files: for instance, a sorted list of the first lines from all the poems in a collection.

How do databases fit into a larger XML publication framework? What do they do?

they create and store indexed information: that is, information from the source XML files that has been preprocessed to make it more accessible and easier to manipulate. For instance, they might store tables of all the document metadata (author, title, genre, date, etc.) so that it can be searched and sorted more quickly

An XML index or a word index is not unlike a back-of-the-book index. Imagine you have a library where the books are arranged on the shelves randomly, or in an order that is relatively useless to the patron (e.g., aquisition order, size order). To find a particular book for which she knows the author & title but not the shelf number, a patron simply has to start at shelf slot #1 and proceed down the aisle quickly looking at each book to see if it's the one. If we have an index, we have a list of all the books including author, title, and shelf number, sorted by author's name. Now she can just look up the book by author name, then title, get the shelf number, and walk straight to the right spot and find her book is missing.

For a computer word indexer it's the same thing: the indexer software reads the file, for each word it tucks that word into an alphabetically sorted list, and associates the word's position with it.

XML databases exist as separate modules that can be used as the basis for XML publishing systems, for instance: eXist DBXML Xindice (Apache) BaseX Qizx/db

But there also exist XML publishing systems which include a database component and also other components which handle other aspects of the process: TEI Publisher (uses eXist): show a bit? Philologic (includes its own database, but can also work with MySQL): show WWP site commercial products like Tamino