2 Outline of Strategy

The Theory

The basic idea is as follows:

• To write a single L^ATEX source;
• To use `ordinary' L^ATEX to produce a DVI file;
• To use the utility DVIPDFM to convert the DVI file to PDF format;
• To use L^ATEX2HTML to convert the L^ATEX source to HTML.

In terms of packages, this translates as:

• Using the html package for basic hyperlink functionality; the html package comes with L^ATEX2HTML;
• Using the hyperref package for the PDF document functionality (e.g. bookmarks, PDF hyperlinks, document information);
• Using the grahicx package to import any external graphics (this package working in conjunction with the DVIPDFM and L^ATEX2HTML -- hopefully!)

The Practice

As we shall see, though this is straightforward in principle, things are complicated by numerous incompatibilities between the packages and by bugs in the various programs. These issues are explored in detail in Section 3.

A large part of the effort is therefore in understanding, empirically, what works and what doesn't, and in implementing workarounds for bugs.

We will explore these

A Note on PDFLaTeX

An alternative strategy is to use PDFLaTeX rather than DVIPDFM. In principle PFDLaTeX can produce PDF and DVI output from a L^ATEX source. My preference has been for DVIPDFM for the following, essentially personal, reasons:

• `Ordinary' L^ATEX is the more mature product -- and it is already set up and running on my PC;
• DVIPDFM is extremely well written, reliable, easy to use and well documented; this is not to say that that is not true of PDFLaTeX, of course;
• Many of the techniques and issues are the same in the two approaches anyway.

Thus whilst the PDFLaTeX approach is perfectly feasible, it will not be discussed further in this document.

© Charles Clayton 2000. Please use this information at your own risk.