[Y-Indology] Alphabetical arrangement of Sanskrit words

[Guest guest]

Dear Dr Vasudevan

in the Indology forum you asked recently:

 

> Sir,

> I would like to know the details of the software for

> arranging sanskrit words

> alphabetically for preparing index of Sanskrit verses. Kindly

> inform me the full details.

> Expecting your early response.

 

Here is a short note on the method I use to sort Sanskrit words in my

dictionary

(at http://sanskrit.inria.fr").

 

I am using TeX/devnag only as an external I/O format, XML/HTML/

Unicode is another one.

All internal computations are realized over specific data structures.

For instance,

the type word is represented by lists of letters (Sanskrit phonemes,

represented as

integers). For instance:

 

# code "yogananda";

- : Word.word = [42; 12; 19; 1; 36; 1; 36; 34; 1]

# decode [42; 12; 19; 1; 36; 1; 36; 34; 1];

- : string = "yogananda"

 

The basic data structures of the Sanskrit software are defined in the

ZEN Ocaml toolkit,

a GPL free software effort available at http://sanskrit.inria.fr/huet/

ZEN/index.html

where you have fully documented sources.

 

Sorting of the dictionary entries is done on the word data structure.

 

The dictionary order of entries, checked within each level (words,

subwords, subsubwords),

is effected by comparing normalized words, where normalisation

eliminates non-original

anusvaara and visarga.

 

Thus:

# code_string "yogananda" = code_string "yogana.mda";

- : bool = True

# code_string "du.hsaha";

- : list int = [34; 5; 48; 48; 1; 49; 1]

# decode [34; 5; 48; 48; 1; 49; 1];

- : string = "dussaha"

 

Velthuis notation is only used at parsing/unparsing time, like

Unicode strings for devanagari or

diacritics representations for the HTML pages,

but all computations are done on the more abstract word structure.

 

I hope this answers your questions.

Gerard Huet