troff_to_html - a TROFF to HTML Convertor

Paul E. Dunne

Department of Computer Science

University of Liverpool

Liverpool L69 3BX

Great Britain

ped@uk.ac.liv.csc

Description

The following describes a very basic text conversion facility for translating a troff source file (using the -ms macro set) to a HTML file suitable for vieweing using mosaic or netscape.

The translator is invoked (locally) via the command

~ped/bin/troff_to_html < input_file > output_file

A (uuencoded compressed) binary (HP7000 Series) is here

Source program (written in Ada) is here

where input_file is a troff source file (that may include directives for the pre-processors pic, and eqn), built using the -ms macro set. output_file will be a HTML source file.

A C++ version (compiling under c89) has been produced by Dave McGaw and may be found here. Please note that this requires a dictionary file for special characters located here and an include (.h) file that is here.

-ms macros Recognised and Translated

• .NH Numbered headings (parameters to .NH are ignored, e.g. .NH 2 processes as .NH)
• .SH Un-numbered headings
• .LP New paragraph and serves as break for .IP (see below)
• .PP New paragraph and serves as break for .IP (see below)
• .IP Indented paragraphs (paragraph labels are ignored); these are processed as un-ordered lists, the list being terminated when .LP or .PP directive encountered.
• .DS/.DE The text between these is left unaltered, hence .DS/.DE translate to HTML <pre> and </pre>.
• .EQ/.EN An (extremely crude) attempt is made to process the eqn text between these.
• .PS./.PE Translated to an anchor to a given file `picture<n>.gif' (where <n>) starts from 1. This file should be created (by the user) to hold the relevant picture.
• .TS/.TE Simple table handling.

Extensions August 1996: Two artificial directives have been added, to allow printing of $ symbols: These are:

• .NE (No Eqn) treat $ in text as character not as opening of eqn expression. (default)
• .EO (Eqn On) Treat $ in text as start of eqn expression.

Ada Version

No limit on number of rows/column. New version (July 1996) uses HTML <table> feature.

C++ Version

Also incorporates HTML <table> (errors in previous version corrected, July 1996). Multiple row formats allowed (max 70 rows/20 columns).

NoteNo font changes or eqn inside table entries. Multi-column spans generally don't look very good. It is assumed that the table specification is valid (otherwise conversion program will crash).

troff Directives Recognised and Translated.

• .ce Printed as a 3rd level heading. Parameters to .ce are ignored.
• .ft Change font. Roman (R), Italic (I), Bold (B), Courier (C) are all recognised.
• .ul Print next line of text in italic font.
• .so Will create an anchor point to the file indicated by the .so command. Link in text will be indicated as `anchor<n>'.
• .br/.sp Causes a line break. Parameters to .sp are ignored.
• Text in the form $...$ is regarded as introducing an eqn expression. If the first character after $ is the character ^ then the text is left in its literal form.

Note: All lines opening with a full-stop are ignored unless the line is a directive recognised by the translator.