\G comments should appear immediately above or below the definition of the word it belongs to. The definition line should contain no more than the definition, a stack comment and a \ comment after which the wordset and pronounciation. An isolated block of \G comments is placed at the beginning of the glossary file. A typical glossary session may look like: NEWGLOS MAKEGLOS SOURCE1.STR MAKEGLOS SOURCE2.STR WRITEGLOS GLOS.GLO BUILD-HLINE addr --- Build header line for glossary entry. GLOS-COMMENT? --- flag Determine if line at HERE is glossary comment, if so. allot it, else store into oldline. INSERT-HEADER addr --- Insert the header into the list at the alphabetically correct place. MAKE-GLOSENTRY fid --- fid flag Read lines from the file fid until \G line encountered. Collect all adjacent \G lines and find header line. then insert entry into list flag=0 if no entry found. MAKEGLOS "name" This command reads a source file and builds glossary info for it in memory. NEWGLOS This command starts a fresh glossary. PROCESS-HEADER Process the header information stored in OLDLINE SCAN-WORD ---- addr len Scan a word on oldline through pointer charptr WRITE-GLOSENTRY addr fid --- write the glossary entry at address addr to file fid. WRITEGLOS "name" This command writes the glossary info from memory to a file. The glossary info may be collected from more source files. \G \G is an alias for \, so it is a comment till end-of-line, but it has a special meaning for the Glossary Generator.