K493750 at CZHRZU1A.EARN
Mon Jun 7 17:20:25 UTC 1993
My name has been mentioned repeatedly in the recent discussion on the
dissemination of Sanskrit e-texts so that I feel called upon to make some
comments. I support Dominik's initiative whole-heartedly; perhaps an
initiative with personal commitment is going to be more successful than the
institutionalized attempts to centralize and pool Sanskrit texts; at least
I have never received any feedback to the several questionaires which I
filled in and sent back.
To the problem of the level of perfection which provides Dominik's starting
point I would like to add the problem of the competitiveness of academic life;
after all one would like to get some kind of acknowledgement for the effort.
This is not my personal problem, but it may be among the causes which hinder
the free sharing of resources. Another problem area may be the copyright about
which I need not bother as long as I am using the electronic copy of a printed
book only privately.
Concerning the level of perfection, I have no qualms about typing errors
(there may be plenty in this letter), but I tend to be embarrassed when it
comes to inaccuracies of tagging or mistakes in the analysis (e.g. of
compounds). Another aspect of the competitiveness? I am not pleading for
lowering the standards but for allowing for more teamwork in getting closer
to the ideal.
When planning the Tuebingen Puraa.na Project we were advised to follow the
strategy assumed also by the TLG. The Brahmapuraa.na was typed twice by
two Sanskritists; however, we "cheated" in typing two different printed
editions of the text. After automatic collation of the two versions (something
which TUSTEP is equipped to do) we had to look at all the differences and
decide whether it was a mistake or a variant. Variants had to be proofread
conventionally (and I am sure we missed many a typing error in the variants,
but we gained considerably for the breadth of our textual basis). I recall
(the fact, not the example) of one case in the transliteration of the
Vi.s.nupuraa.na where the same mistake was repeated four times in the four
versions of one chapter produced by three Sanskritists: It was not a typing
error but a combination of words which was entered as compound where a compound
was syntactically not possible.
"TUSTEP--format" is somewhat of a misnomer, as has been pointed out. The
Tuebingen System of Textprocessing Programs (TUSTEP) is the tool used to
handle the input. Dominik calls it the "Schreiner format" -- I feel honoured,
but perhaps something like the "Tuebingen--Zuerich format" ("TZ"?) would
be more appropriate (Renate Soehnen--Thieme has been using it at SOAS, Lars
Fosse supports it ...). What is specific to this format seems to be the fact
that sandhi--changes are marked and that nominal compounds are dissolved.
And since Dominik emphasizes the importance of having verbatim texts without
sandhis marked or compounds dissolved I may repeat that this version of
any of the texts in TZ--inputformat can be automatically generated (I call it
the textformat; also the "pausaformat" of each word, i. e. the form a word
would take at the end of a sentence can be generated; cp. the published
materials for the BrP). Sandhi markers and compound dissolution do not yet
make for a "tagged" texts. I suppose tagging should be and must be problem
oriented; there is probably no such thing as a "fully tagged" text as long
as anyone can conceive of a new problem or a new method to be applied to a
(to the same) text.
I hope Dominik's initiative will have tangible results.
More information about the INDOLOGY