The basic premise of using Word in combination with XML is that authors can be trusted to create highly structured content in Word, providing they get the appropriate assistance and training. If you believe that authors cannot do so, then this strategy is not for your organisation. In addition, authors would rather use a tool they are familiar with than learn something new, and organisations would rather use tools they already have than buy something new.

If you can create structured content in Word, then there are a variety of tools that can convert it into XML in an automated way. One such tool is UpCast (http://www.infinity-loop.de/), which the previous speaker has described. Once in XML, XSLT programs can be used to convert the content into HTML for web publication, and XSL:FO for low-quality print publication. Most typesetting applications, such as QuarkXPress, FrameMaker, and 3B2 can import either Word or XML content and typeset it to an acceptable standard for high-quality print publication.

Typically however, there have been a number of problems with this scenario.

These three issues: content creation, conversion, and publication; have normally been considered as separate tasks, to be carried out using different tools, and by different people. Yet Another Word Converter (YAWC) is a 3-in-1 tool, integrating all 3 steps fairly seamlessly into the Microsoft Word editing environment. This approach greatly reduces the initial setup time and cost for an XML-based publishing environment. It does not address the more complex issues of workflow, version and access control, management reporting, searching, and the host of other features of a full CMS, but it offers a quick, cheap, and relatively painless introduction to the rigours of a formal information management process, and allows authors to gain experience and expertise in markup, and see the immediate results of their work as published documents, without the need for intermediaries such as webmasters or even typesetters.

The YAWC default authoring interface offers a few simple features.

Additional customisations can be easily made using VBA in the normal way, but the basic features go a long way towards encouraging and assisting authors in creating very well marked up content. It is still possible to create a document that does not conform to the target XML DTD required, because there is no formal XML validation of the document. The authoring environment is not really XML-aware, unlike such tools as WorX SE or S4/Text. However, we have found that the quality of the markup is surprisingly good in practice.

YAWC contains a built-in Word to XML converter. For authors, converting a document is a single step. There is no need to start a different application or save the document into a different format. For customisers, the process of mapping the Word content into XML structures is about as simple as it can be. The 80/20 principle applies: 80% of the conversion can be handled by means of a simple, direct mapping between Word styles and XML elements; the remaining 20% of more difficult constructs can be handled using an XSLT script to re-organise the relevant information into the required XML structure.

Although Word, like HTML, is effectively a linear structure with no hierarchy, the conversion process deduces the hierarchy based on the Heading styles uses, and makes this explicit in the XML output. Tables are converted into the HTML table model, retaining cell spanning and alignment information. Nested lists are supported, although they must be carefully marked up to convert correctly, and custom document properties and form fields are also converted into XML.

Some Word artifacts are deliberately not converted, in order to keep things simple. Colours, fonts, point sizes, headers, footers, etc. are all ignored.

Two default output target DTDs are supported by default, HTML and DocBook. These are the most widely used DTDs, and can be used as the basis for developing custom output DTDs if required.

As an aside, many organisations tend to engage in a long and tedious process of document modelling as part of the move to a more structured information management strategy. I would suggest that ignoring or postponing this step, and simply choosing to start with DocBook, allows you to get going in a very short time, which is worth much more than the possible slight loss of precision from using a generic document model.

Publishing content is sometimes just a simple matter of converting Word into HTML. More often however, multiple presentation outputs are required, such as PDF, HTML, and even WML or text. YAWC supports the generation of multiple outputs in a single operation. The publishing process involves a number of steps.

Any number of XSLT post-processing steps are supported, so the conversion process doesn't just deliver the target XML you want, but also any further downstream outputs you need. For example, you can generate DocBook XML, and publish it in HTML and PDF format using Norm Walshs DocBook XSL stylesheets, all in a single step.

Some additional features of the HTML publication stage are worth a mention.

These benefits are realisable because the information is first converted to XML before generating HTML. The XSLT stylesheets packaged with YAWC do all the work of ensuring WAI compliance and placing DC metadata in the HTML head element. Although many Word to HTML converters exist, they generally do not support these particular features. Because they do not use XML as the intermediary, they focus narrowly on web publication, which is OK for small corporate websites, but not for larger information publishers.

CMSs are expensive, difficult to implement, and involve taking a gamble about which vendors will still be around in 5 years. They do not always deliver the benefits we are led to expect. A low-risk and low-cost alternative is to use Word for content creation and XML for content publication, in order to learn about the information management process, and develop valuable experience in managing and publishing XML information. After all, the real goal for most organisations is to be able to publish efficiently and cost-effectively.

YAWC, a plug-in for Word, is designed to address the need to publish quickly and cheaply, and is already used by a variety of publishers, including newspapers, government departments, research institutes, and academic publishers.