Summary

• Avoid using any display-specific format as the principal form of archiving electronic information, e.g.,a scanned bitmap image, a PostScript or PDF file (visual rendering) or a digitized recording (audio).
• Avoid use of explicit visual layout in the electronic encoding. For instance, avoid use of \vskip in La)TeX documents.
• Use distinct markup to encode semantically distinct objects even if they have the same visual layout.
• Use an encoding system that is extensible by the author; this will ensure that the maximum amount of semantic information is captured at the encoding stage. This minimizes the amount of guesswork that has to be done later.

Electronic document encodings have not always followed these rules, since the markup was viewed purely as a means of producing the visual rendering. Our work points out that the same encoding can be put to multiple uses; it is therefore important to apply principles of good software design and reuse to document encodings as well.

To draw an analogy, we do not currently throw away the program source code once we have successfully compiled it into a running executable; equivalently, it is important to retain the high-level document encodings that produce the final display form in which information is disseminated.