TeX for Indology

[Gary Tubb]

kichenas at math.umn.edu asked
> Is there a (La)TeX style file which enables
> the input of Romanized text? All the tools for this are
> present in TeX (macrons, dots over and under letters,...)
> and all that is needed is the definition of an 
> environment which would enable the user to input
> \begin{tamil}
> kaNTanen kaRpinuk k-aNiyai
> \end{tamil}
> and produce the output corresponding to
> {\em ka\d n{}\d t{}a\b n{}e\b n kaRpi\b n{}uk k-a\d n{}iyai}

   I agree with Nath Rao's explanation that for Roman script a simple
filter may be best.  I would add that because much of my typing is in
English with chunks of Sanskrit, including proper names that need
uppercase letters but not italics, I prefer not to use a scheme that
relies on uppercase letters or that begins a non-English group.  
   What I find quickest and easiest is simply to type a semicolon before
any letter that needs a diacritic of any kind.  I use a semicolon
because it is the easiest non-letter to type (and faster than using a
control-key code), it does not normally occur before a letter, and it is
relatively unobtrusive visually.  Then any context-sensitive filter or
group of macros can be used to convert ";Sa;nkara;sarm;a ;Dhu;n;dhir;aja
Mu;njat;irtha" into "\'Sa\.nkara\'sarm\=a \d{D}hu\d{n}\d{d}hir\=aja
Mu\~njat\={\i}rtha" and so on, throughout the entire document.
   I haven't used it for Tamil, but for Sanskrit there is only one
frequent ambiguity, which I avoid by typing ;x for the retroflex s and
reserving ;s for the palatal s.  For the less frequent difficulties
(word-final velar n and the long r vowel) I simply type in the TeX
commands myself, rather than spend more time devising shortcuts.
   As an example of how to do this, the following six-line ex script
can be used within any vi-type text editor to process a file in the
course of editing it.  If you use UNIX the same script can be used with
very slight changes to run either ex or sed on a file directly.  The
commands determine which type of nasal to use by looking at the
following consonant, even if an optional-hyphen (\-) command intervenes.

  " Ex script for inserting TeX commands for Sanskrit diacritics.
  " Type text with semicolon before each character that needs diacritic.
  " Use ;x instead of ;s for the retroflex sibilant (underdot s).
  " Long r vowel and word-final velar n must be entered manually.
  " Run from inside vi by entering ":so <scriptname>" in command mode.
  " For sed script delete % in each line and change " to # in comments.
  %s/;\([nN][\\-]*[kg]\)/\\\.\1/g       "overdot for velar nasal
  %s/;\([nN][\\-]*[cj]\)/\\\~\1/g       "tilde for palatal nasal
  %s/;\([rlmhtdnxRLMHTDNX]\)/\\d{\1}/g  "all the underdots
  %s/;i/\\={\\i}/g                      "macron for undotted long i
  %s/;\([auAIU]\)/\\=\1/g               "macrons for other long vowels
  %s/;\([sS]\)/\\'\1/g                  "mark for palatal s 

Something similar could be done for Tamil.  If there are too many
ambiguities you could also use colons or commas in addition to
semicolons (but probably not a period, which occurs before letters in
filenames and elsewhere).

     Gary Tubb,
     Columbia University.