This collection of pages is built from XML docbook documents using the toolchain from Norman Walsh toolchain and a customization of his docbook stylesheet and website stylesheet for HTML (A third column was added e.g.)

Automatic transformation is done using the saxon xslt processor, controlled by makefiles.

Automatic upload is done using the sitecopy tool.

Link control is done through Xenu. All these tools are free software.

Lately the navigation is no longer generated directly. Instead, CSS3 commands are used to arrange the div sections properly. The stylesheet does now only generate div sections with proper class annotations. The rest is done by the browser.

Why?

There are a number of compelling reasons to maintain these pages in XML rather than HTML.

  1. Traditional printed documents are desirable for several of these pages.

    Using a more structurally rich format to represent the documents makes it possible to produce print as well as online versions.

  2. All of the intra-document linking is automatic.

    The XML system provides validation for the inra-document links, so there is no danger of broken links. (Links to external pages still have to be validated by an external program.)

  3. Migrating these pages to XML and/or future versions of HTML will only require a new or updated style sheet. No pages will have to be recoded. And the pages will conform to whatever output format is required, with no possibility of random human error.

But why not just use Wordpress for a blog? This would come with built in comment features and many other things. The answer is, that I prefer simple static pages and I simply do not need the things offered by a blog system.

How?

There are a number of steps necessary.

  1. In case of papers in Microsoft word format we used a tool from logictran Logictran Homepage to convert them to XML. More specific: to the docbook schema in XML. The conversion was easy and required little manual interference. .

    Logictran converts embedded pictures but the quality is better if the originals can be used.

  2. After validating the docbook sources (we use the SP parser from James Clark) .

    The XML system provides validation for the inra-document links, so there is no danger of broken links. (Links to external pages still have to be validated by an external program.)

  3. I tried several XML editors (XMLMind, Oxygen, Xemacs with SGML mode, VI etc. Oxygen allowed customization for the website dtd. But in many cases I just use VI or some other simple text editor.