sana-aitta.db.dtd

<!--
DTD for sana-aitta.db

This part quasicode only, to describe the source.

The sana-aitta.db.xml was generated from the .docx with 
antiword -x db

The structure is (content is marked in BOLD):

1 <para>
2 <emphasis role="bold">
3   LEMMA
4 </emphasis>
5 nothing
6 TRANSLATION TO NORWEGIAN OR FINNISH OR LATIN
7 ;
8 EXAMPLE FROM THE NOVEL.
9 the string &gt;
10 REFERENCE TO ANOTHER LEMMA
11one of the strings " AR " or " KT "
12 a page number
13 ;
14 </para>

The boundary signal for the translation is eihter parentheses (around the translation)
or semicolon (after it).
8 need not be present if there is a reference to another lemma in the article
9, 10 is a reference to another lemma.
11, 12 is source infomation, reference to novel (AR or KT) and to page number.
Then either boundary 13 and loop back to 5, or go to 14.
Only the lemma and the boundaries para and emphasis are compulsatory.

The goal is not to give a full parse of the input data, but to parse
down to a basic tripartition of the string.


çs -> š
çc -> č
çz -> ž
    
-->

<!ELEMENT rootdict (entry+) >
<!ELEMENT entry (lemma, mgr*, refgr?) >
<!ELEMENT lemma (#) >
<!ELEMENT mgr (trgr*, exgr*, syngr*) >
<!ELEMENT trgr (#PCDATA) >
<!ELEMENT exgr (#PCDATA) >
<!ELEMENT syngr (#PCDATA) >