Fwd: Indices of Sanskrit Texts

[Dominik Wujastyk]

---------- Forwarded message ----------
From: Harry Spier <hspier.muktabodha at gmail.com>
Date: 27 August 2010 02:19
Subject: Re: Indices of Sanskrit Texts
Re: Searchable Indices of Sanskrit Texts

Dear list members,

The digital library of the  Muktabodha Indological Research Institutes
www.muktabodha.org contains around  two hundred searchable medieval
religious texts (mostly Tantric and Agamic).  It contains a very powerful
search engine that allows you to search for either simple words, expressions
or complex patterns of words in both Kyoto-Harvard and Velthuis
transliterations.  The search engine brings up all the lines in the texts
where  the search pattern was found and clicking on a line brings up the
full text at the highlighted line.

For example, recently Madhav Deshpande asked on the list a question about
anusvAra and bindu.  By searching in the Muktabodha digital library with the
search pattern

*<((nusvAr).*(bind))|((bind).*(nusvAr))>
*
This will give you about 25 references in 7 different tantric texts to where
anusvAra and bindu are mentioned in the same sentence in tantric texts.  You
can then click on any references you are interested in and the e-text will
come up at the correct location.

This is a complex search pattern but you can do much simpler searchs and
with practice the "regular expression" syntax becomes easy.
The *< >*in  *<((nusvAr).*(bind))|((bind).*(nusvAr))> *means you are using
Harvard-Kyoto  I.e. everything between < and > is Harvard-Kyoto.  If you
wish to use Velthuis then you put the search pattern between *{ *and *}*.


*(nusvAr).*(bind) *means all sentences containing both nusvAr and bind no
matter how separate in the same sentence(I'm truncating the words so you get
them whatever the sandhi
*(bind).*(nusvAr) *means all sentences containing both bind and nusvAr no
matter how separate in the same sentence.

The *| *means search for both patterns and the *( )*'s are just a simple way
of separating items in the search pattern.
These are whats known as a "regular expression" search pattern.

*.** means there may be letters between the items searched for.

The beauty of "regular expression" search patterns is that you can search
for variations.  For example above I'm searching not only for where bindu
follows anusvAra but also where anusvAra follows bindu.

I've always thought that a "regular expression" search engine could be an
extremely valuable tool in Indological research.

To use another example, a few years back on the Yahoo Indology list Michael
Witzel indicated that  the high frequency of "ha" at pada final in the
Uttarakhanda of the Ramayana showed that that khanda was relatively late in
response to someones question.  It was very easy to find all lines with "ha"
and also all pada final "ha" with a regular expression search engine to get
a relative count of its position in the pada.  (in that particular case it
was the search engine of a programmers editor that was used).

Regards,
Harry Spier
Muktabodha