LIS 450 DP Final Project Spring 02

Final Project Description

The main portion of this project is a DTD for office documentation. It is designed to be usable for training manuals, online tutorials, office handbooks, and other types of office reference materials. The DTD used for all valid document instances included in the final project is version 6.0. Metadata for a document instance can be encoded using the Dublin Core. Metadata can be included within a document instance or in a separate file. The Dublin Core DTD is included.

In addition to the documentation DTD, I explored three different XML applications we touched upon this semester: XSLT, XSL-FO, and the Docbook XML DTD. For part one of the project I created a group of XSL stylesheets that would transform valid document instances of into HTML documents for display on the web. The second part of the project involved the creation of a group of XSL stylesheets that would generate well-formed XSL-FO (Formatting Objects) documents when applied to valid document instances of the DTD. The final part involved the creation of a fairly simple XSL stylesheet that when applied to valid instances of the DTD would result in a valid document instance of the Docbook DTD. The impetus behind this was to use the Docbook XSL distribution available on sourceforge and compare the results of the FO and HTML transformations using this distribution to the results produced by my own stylesheets and DTD.

The two valid document instances used as examples throughout the project are:

Instance One
Instance Two, Dublin Core metadata file for instance two.

Please see the README file for an explanation of the directory structure of the project and for a more detailed explanation of problems with the output of the various transformations

HTML Transformations

Part one of the project is a group of stylesheets that transform valid document document instances into HTML 4.01. The group can be found at this location, or viewed in a single file.

Project DTD to HTML

Issues:

Needed to use a CDATA section to embed javascript in an XSL stylesheet as since the beginning javascript notation and ending notation for a javascript resembles an XML comment .
HTML produced is not valid 4.01, since a property to suppress the output of the dc and rdf namespace declarations to HTML documents needs to be found, additionally the attribute "type" element in invalid, but since some browsers don't recognize the valid language attribute, type has been substituted.
Need to suppress the appearance of the FO namespace xmlns attribute in HTML documents output from the stylesheets.

FO Transformations

Part two of the project is a group of stylesheets that transform valid document document instances into well-formed instances of XSL-FO. The group can be found at this location, or viewed in a single file.

Once the the well-formed FO documents were created, FOP, a XSL formatting objects printer was used to transform the FO documents into the PDF format.

Results of the Transformation

Project DTD to FO

Project DTD to PDF

Issues:

The FOP printer does not implement all of the properties specified in the FO specification so it is important to review the supported properties at this page.
The biggest difficulty faced in creating the template for document layout using FO was creating a suitable template to process the <table> element of the DTD. In specific the issue of calculating the correct width for each column and individual cells that required a special width within the general table. This is necessary because the fop printer does not support the automatic layout of tables like a typically web browser. As a crude workaround I created a template that would assign a fixed amount of space to each column depending on the number specified in the columns attribute available for use with the <table> element. A way to do this based on the length of the string residing in the table cell would likely to be the most efficient.
FOP has been upgraded recently, I plan to rewrite some of the stylesheets and test how the FO property support has been improved.

Docbook XSL Exploration

The third portion of the Assignment was an exploration of using the Docbook XSL Stylesheet distribution maintained. An XSL stylesheet that transforms valid instances of the Project DTD to valid Docbook 4.1.2 was created. The resulting documents from this transformation were then transformed to HTML and FO using the stylesheets of the distribution.

Transformations

Project DTD to Docbook

Docbook to HTML

Issues:

The conversion of my internal document links did not work. These came up as question marks.
The stylesheets were unable to deal with the placement of an anchor before each title element, even though this placement was valid according to the docbook v4.1.2 dtd

Docbook to FO

Issues:

The conversion of my internal document links did not work. The processor gave errors to the effect that the stylesheets were unable to process the anchors and internal references to those anchors within the document.
The stylesheets were unable to deal with the placement of an anchor before each title element, even though this placement was valid according to the docbook v4.1.2 dtd
The tables were not processed correctly, in some instances, they were partially processed in other cases just the text of the table cells was sent to the FO. This was again problematic because the tables posed no problems to validation

Docbook FO to PDF using FOP

Issues:

In general, it was hard to find consistent documentation for learning the FO language. This is probably the result of the specification being so new. Additionally it was difficult to use FOP, at times, because a number of FO properties are not supported by the printer at the moment making the process of using tutorials frustrating because the examples did not work when sent to FOP.

Last Update: 7/28/03