- <http://groups.google.com/groups?selm=Bv0A9M.27B%40rice.edu&output=gplain>.
-
-This input method is based upon the Telugu Rice Transliteration
-Standard (RTS) specification and its Rice Inverse
-Transliterator (RIT) supplement.
-
-The original RTS specification was written by Ananda Kishore and
-Rama Rao Kanneganti in 1992 and can presently be accessed in the
-\"soc.culture.indian.telugu\" newsgroup archives (see
-<http://groups.google.com/groups?selm=Bv0A9M.27B\%40rice.edu&output=gplain>).
-
-The RIT supplement adds alternative combinations for
-transliteration but, in general, does not distract from the
-original specification (see
-<http://www.teluguworld.org/RIT/rit3.0/manual.html>). Whenever a
-supplemental combination conflicts with the original RTS, the RTS
-version has precedence and the supplemental combination is
-disregarded (such as 'ea' from RIT 3.0).
-
-Finally, this input method deviates slightly from the RTS in the
-following ways:
-
- (1) The combinations \"\@n\", \"\@2\", \"~c\", and \"~j\"
- yield \"�\" because their corresponding glyphs do not
- yet exist in Telugu's Unicode chart.
-
- (2) If it appears at the end of a word, the combination
- \"m\" yields \"ం\". The user can type \"m&\" to
- bypass this behavior and force \"m\" to yield \"మ్\".
+ <http://groups.google.com/groups?selm=Bv0A9M.27B@rice.edu>.
+
+This input method is based on the Telugu Rice Transliteration Standard (RTS)
+specification[1] and its Rice Inverse Transliterator (RIT) supplement[2].
+
+The original RTS specification was written by Ananda Kishore and Rama Rao
+Kanneganti in 1992 and can presently be accessed in the archives[1] of the
+'soc.culture.indian.telugu' USENET newsgroup.
+
+The RIT supplement[2] enriches RTS with alternative combinations. However,
+in cases where RIT and RTS define conflicting mappings for the same
+combination, such as 'ea', only the RTS mapping is honored.
+
+Finally, this input method deviates from the RTS in the following ways:
+
+* The combination '\@n' yields '�' because its corresponding glyph does not
+ yet exist in the Telugu unicode chart.
+
+* The combination 'm' yields 'ం' if it appears at the end of a word. The
+ user can type 'm&' to bypass this behavior and force 'm' to yield 'మ్'.
+
+* The sunna prevention operator '&' can be used to force a more literal
+ transliteration of consonant compounds such as 'jn' by writing 'j&n'.
+
+[1]: http://groups.google.com/groups?selm=Bv0A9M.27B@rice.edu
+[2]: http://www.teluguworld.org/RIT/rit3.0/manual.html