Notes on using the T3 -> TeX conversion programs The T3 -> TeX conversion program is normally started using a batch file. The batch file uses default auxiliary files so that only the file being converted needs to be entered. The batch file always uses .T3 as the file extension for the input and .TEX for the output file names. The auxiliary files control the default character mappings and conversion processes. Alternatively, the conversion programs can be run directly and different auxiliary files can be specified. This is useful if you are experimenting with changes to some of the auxiliary files. Sample command line for the T3 to TeX conversion program: t3totexx doc.t3 convert.cnv out.tex t3totex.dat Explanation of command line parameters and flags: doc.t3 the T3 source file being converted - this file is generated from T3 by DOS exporting of a document. convert.cnv the file that is used during the DOS import - it defines tokens that appear in doc.t3. out.tex the name of the TeX file to be generated t3totex.dat the name of the datafile that is used for character mappings, character translations and special control sequences. One more file is necessary for TeX to process the converted output file. The file "t3totex.tex" contains TeX macro definitions for macros used in the file "out.tex". The macro definition file "t3totex.tex" must be available to TeX for proper results. The file "t3totex.dat" is a standard ASCII file and can be edited by the brave user. It consists of: CHARACTER SET NUMBERS These are correspondences between T3 font names and character-set numbers (0,...,12). Any font that uses the set number 0 will be treated as roman. Any font that uses the set number 2 will be treated as italics. CONTROL SEQUENCES Sequence 1: used for diagnostic purposes, can be any string of digits from 0 to 9, or the letter n. The only digits recognized at this time are 0, 1 and 9. Digit 0: causes the contents of the buffer to be printed to the screen each time it is sent to the output file. Digit 1: prints to the screen an analysis of the items being processed. For example, w4: a four letter word w3}: a three letter word that ends a font group c: a column c_: a subscript c^*: a superscript that was caught earlier [5: start of a special list of 5 items to build superscripts, subscripts, a numerator, denominator, or radical ]: end of a special list em0: math mode was not ended after item PUNCT: punctuation BINOP: binary operator LEFT: a left delimiter DASHm: a dash in mathematics (a minus sign) Printing diagnostics slows down the program. Sequence 2: contains the name of a TeX file (without the .tex extension) to be included at the top of the output file. This will normally be the file t3totex.tex which contains macros used by t3totex. Sequence 3: a line to be included at the end of the output file. This is \vfill\eject\end for plain TeX. No embedded spaces are allowed in this line at the moment. CHARACTER TRANSLATIONS The line 1 96 10 \alpha\ means that character 96 of the italics font (set number 1) is a mathematics character (type 10) and will be translated as "\alpha ". You can put comments after each line provided you leave a space. Internal spaces in the translations are indicated by "\ ", as is the final space, if any. Here are the types of characters in t3totex.dat. MATH 0 letter 10 mathchar 1 left 11 mathleft 2 right 12 mathright 3 notminus 13 binop 4 binrel 14 binrel 5 punct 15 mathpunct 6 apo 16 math 7 digit 17 up 8 dash 18 down 9 notmath Characters of type notmath are those that are forbidden to appear in mathematics mode in TeX: mostly they are ordinary characters with accents (o-umlaut, a-acute). Watch out if you use this type when modifying t3totex.dat because letters that are next to such characters will not be permitted to be in mathematics mode: the reason is so that in the name "Goedel", for example, with an o-umlaut instead of oe, the "G" will not be interpreted as mathematics. On-line superscripts and subscripts have type up and down. Most of the Greek font is implemented; this font is for writing Greek, not mathematics. Some of the accents aren't too good. This is all done in t3totex.dat, and can be modified. The Cyrillic font is transliterated by t3totex.dat. Font 18 is a special internal font known only to the t3totex program. It is an intermediate step in translating built-up symbols: for example, if the key sequence iint is recognized, it will be replaced by character 11 of font-set 18. The actual translation to TeX takes place in accordance with t3totex.dat, which can be modified by the user. The file t3totex.tex is a standard ASCII file containing the TeX macros used by t3totex. It is not used directly by t3totex, but must be available when the TeX file created by t3totex is processed. The default magnification is magstep1. This can be changed by editing t3totex.tex. This file also contains the definition of \miss, which determines what is printed when the program comes across a character it can't interpret (usually a piece of a built-up construction that has not been recognized). The supplied definition is a small black box. When the program comes across a column that it can't make sense of in terms of subscripts, superscripts, or some built-up construction, it puts it out as a column using the macros \tcol, \dB, \nB, and \mB that are defined in t3totex.tex. You can see how these are used by examining a generated TeX file. The tcol constructions should normally be eliminated from the TeX file one way or another. You can find them by searching for the string "tcol". A lot of times they will be indicated on the printed page by the word "tcol" in the right hand margin; but this won't happen if the column is too deep in the TeX code. The subscript/superscript algorithm interprets anything it can as a subscript or a superscript: if there is no character on the main line, everything above the main line is treated as a superscript, and everything below as a subscript. This often gives recognizable representations even in odd cases; for example, a matrix may come out with all its entries being subscripts and superscripts on a left parenthesis. The lines on fractions and radicals are expected to be built up from the character S1 114, which is the shift-minus on the math keyboard. Line formats and named texts, other than footnotes, are ignored. Line formats are included as comments in the TeX file. The combination of an honest superscript on top of an on-line subscript (like S1 50) will not be treated correctly. Similarly for an honest subscript below an on-line superscript. Subscripts below a prime, if they are in the normal positions, will be recognized. Built-up vertical lines are translated into \left| or \right|. This can give some odd effects if the line is meant to stand alone, and is used between some genuine left/right delimiters.